Home

- Department of computing science

1. Split Single Split Single Split 5months 1778 180 14 42416 429 75 229 1536 3months 1092 110 55 43292 43860 213 1546 46 Table 18 shows the outcome of calculation regarding cost and savings in euros Existing cost in euro s is depicted in the first two data columns based on existing usage in kWh and the price of single and split tariffs Following two columns are not results from HEP calculations but are added to this table for clarity These columns show an estimation of total cost per year based on the existing cost and usage from the two previous columns The final two columns show the prediction of savings in euro s if the scheduling calculation chapter 4 6 is applied and a single tariff is used on one hand and a split tariff on the other used the data in Tables 17 and 18 to calculate the percentage difference between the measured periods in kWh consumption shown in Table 19 and the percentage difference in cost and savings between using a single or split tariff shown in Table 20 The percentage difference is determined by the difference of the two values divided by the average of the two values Table 19 Percentage difference of usage calculations between periods Between Existing Existing Estimated Estimated periods usage usage usage savings kWh kWh day kWh day kWh year Tariff Both Both Both Both 3 5 months 50 72 2044 20 7
2. Wl Energy Optimization PROPOSED USAGE SINGLE TARIFF Standby Refrigerator Standby value Standby value Standby value Standby value Standby value Standby value Standby value Standby value Standby value Standby value Start schedule kWh usage ix 0 0013 Usage savings Estimated average usage 5 158 kWh day Estimated kWh savings 10 23 kWhjyear Estimated savings 2 29 Efyear PROPOSED USAGE SPLIT TARRIF Standby Refrigerator Standby value Standby value Standby value Standby value Standby value Standby value Standby value Standby value Standby value Standby value kWh usage ix 0 0013 Start schedule Usage savings Existing average usage 5 186 kWh day Ld 1 Hi I Estimated average usage 5 158 kWh day ran Aidi 56 55 58 565 dd d 12 08 14 00 16 0 16 200 25 508 E Estimated kWh savings 10 23 kwh year Time of day Estimated savings 15 36 year Proposed usage with a split tariff 5 month period 64 amp 3 Tooo X Desn wy Time of day B Average energy usage in EWh s x 0 001 ith a single tariff 3 month period Existing usage w D SM EGET n 3 A t n n Tooo X Desn Ymy A 3 Time of day B Average energy usage in EWh s x 0 001 ith a split tariff 3 month period 65 Existing usage w PROPOSED USAGE SINGLE TARIFF kWh usage ix 0 0014 o 8 BS ESE SE E i i A AA A
3. In my evaluation concluded that peak periods could be reduced if appliances where scheduled according to provided suggestions In regards to flattening the peaks calculated the average absolute deviation from the mean power based on the graphs produced This resulted in flattening of peaks if a split tariff was used in the calculations but not if a single tariff was used Since this was based on the graphs produced in the HEP which show the average usage for 24 hours over a selected period found that could possibly flatten the peaks but this method of using average consumption data is not effective enough e Can we provide a means to see if different types of tariffs are adding to monetary savings In the design and implementation of HEP I included a venue for the end user to provide a single or a split tariff The end user can see the difference in cost when applying both tariffs When evaluated this solution found that based on existing usage a split tariff would not be cost efficient if the appliances where not scheduled On the other hand there was a potential for a 14896 saving or higher if the appliances where scheduled and a split tariff was used believe can now proceed to answer the research question positively The application HEP has been shown to be an effective solution to optimizing electricity It is designed to work with existing appliances by using a Home Automation System such as Plugwise and it does not require an
4. and the session running at 21 00 will be added to the Strict Appliance s existing usage when the graph is displayed This occurs in all data sets for the clothes dryer 50 washing machine and dish washer several times and can explain the difference in results between the single and split tariffs Based on this information since the split tariff schedules all start at 23 00 and there were no occurrences of Flexible Appliances running a session at that time can conclude that the proposed split tariff does provide a flattening of peaks On the other hand am not able to formulate a conclusion regarding the proposed single tariff since the data set used has several occurrences of Flexible Appliances running a second or third session at the same time as the proposed schedule is defined 6 1 3 Calculation results When examining the outcomes of the calculations performed by HEP I see that there is not a significant difference in existing usage based on 1 3 and 5 month analysis 3 months analysis yields the highest kWh usage per day or 2 0496 higher than 5 months and 1 6496 higher than 1 month while the difference between 1 and 5 months is only 0 4096 Similarly the estimated usage in kWh per day after applying the schedules for single and split tariffs is highest for 3 months analysis and lowest for 5 months analysis The percentage difference is very similar 3 months is 2 0996 higher than 5 months and 1 5996 higher than 1 month while the
5. appliances used and gives the end user information about potential savings in euro and kWh 12 3 Background The EU has enforced measures in order to improve energy efficiency in new appliances There are measures taken in regards to labeling incentives and regulations 3 1 Labeling and regulations regarding energy efficiency in the EU Directive 2009 125 EC is the most resent eco design framework for the minimum energy performance standards MEPS which appliances in the EU must meet in order to be eligible for sale 24 These standards are determined for each type of appliance and other products using energy where each type is given a minimum energy efficiency index EEI within each energy efficiency class These minima are revised and lowered which results in phasing out of lower energy efficiency classes and effectively banning product sales in the EU of appliances that do not comply with these standards The case of incandescent lamps is an example of a successful phase out which started in September 2009 and finished at the end of September 2012 25 00 ENERG 50 eHepruWn EVEOYELa Va R ENERC x fa Wi L f pia YZae XYZ Y Z Acor nnumj kg YZae acer ENERGY STAR Figure 3 a New EU energy label b CE marking and c EU Energy Star label The European energy labeling scheme was established by directive 92 75 EC in 1992 In May 2010 a new directive 2010 30 EU 26 was adopted and became effecti
6. 1 2 has been completed 2 HEP displays existing energy use in a graph and in written format as well as the existing estimated cost 2a End user wishes to see information about estimated cost 2 End user clicks on Appliance List button 3 End user completes use case 5 3 2 4 End user clicks the Existing energy use button 5 HEP displays existing energy use in a graph and in written format as well as the existing estimated cost 28 5 3 4 View proposed usage The following use case shows the actions which need to be taken when the end user wishes to see the outcome of the analysis or the proposed usage Proposed usage displays the appliances chosen along with their proposed standby values the proposed scheduling times of Flexible Appliances and estimated usage and saving information in kWh and euro In addition it shows average kWh per day in a graphical format NETAS View proposed usage End user clicks the Proposed energy use button End user elects to view the proposed energy usage and has completed use case 5 32 The proposed energy usage is displayed 1 End user clicks the Proposed energy use button 2 HEP displays proposed energy use in a graph and in written format Alternative Flows 1a Rules have not been assigned to appliances 2 An alert box with the message You have to assign rules to all appliances strict flexible is displayed 3 End user clicks the OK button on the alert
7. 5 Observed appliance values standby values and classifications Close examination of recorded data for each appliance resulted in the findings of actual standby value ASV the range of measured values lowest to highest consumption reading larger than zero whether there is a zero reading on an appliance without turning it off These findings are shown in Table 3 as ASV Range and Zero reading respectively 34 In order to determine the actual standby value for each appliance the recorded consumption data was examined with further tests being carried out when needed The methodology for each appliance is as follows Appliances determined to have standby values e Coffee maker with bean grinder appliance 2 has a built in programmable standby function which was active between 20 00 and 07 00 each night and returned a consistent reading of 4 e Fan on a timer appliance 10 activated the fan four hours each day and returned a constant reading of 1 for the remaining eighteen hours e Stereo system appliance 12 and CD player appliance 13 both had a light indicating that they where plugged in but an on off button needed to be pushed either on the appliance itself or on a remote control in order to activate them The initial reading without pushing the on off button returned constant readings of 2 for the stereo system and 6 for the CD player e Dishwasher appliance 4 had a blinking light which turned on as soon as a program was sele
8. 5 minutes before the hour it is evident that the HEP determines that the appliance needs 180 minutes to run instead of 70 If the timeslots were divided into 15 minute intervals the maximum reading the previous appliance would get is 90 minutes This would result in much tighter schedules for high consuming appliances A great benefit for the application would be to add the ability to use real time pricing of energy since this is one of the promised features of the smart grid 10 This means that a component would have to be added that could collect price information from electricity providers in real time determine which provider to use and give suggestions on schedules based on dynamic pricing This would involve a prediction algorithm to determine what the price will be in the next twenty four hours based on current price information and trends found that the current graphical display of consumption did not produce true flattening of peaks It could be more effective if we produced a graph with one day of consumption for the purpose of flattening of peaks This does not however interfere with the proposed scheduling and standby power detection since also display these in words and provide calculations showing potential savings and usage 55 Bibliography 1 Energy 20 20 20 targets European Union Online European Union June 28 2013 Cited July 1 2013 http europa eu pol ener index_en htm 2 Ellis M Gadgets And Gigawatts
9. ARMED ETT 56 Appendix Images explaining the Graphical User Interface of HEP ccccccccocccnnncnnnncnnnnnnnancnnnnnnnanonnnnos 60 Appendix II Screenshots from HEP of optimization evaluatioOn ooocccconocnnnonaconnnnnannnnnnaronnnnanonnnnnnoss 63 List of Figures Figure 1 Final electricity consumption by sectors in the EU 27 oocccccnccccnncnnnnaccnonacononanononananonacnnnarononanonona 6 Figure 2 Breakdown of the EU 27 residential electricity consumption 2007 and 2009 7 Figure 3 a New EU energy label b CE marking and c EU Energy Star label eee 13 SUE Ch Fy Pu 0 EEE UIDI EM LUCAM EPI LEE UTE DEDE C E EMI DU IUE 15 FE NNN 17 Figure 6 System components and data flow X id t2 number value eee 22 Figure 7 HEP classes showing main variables and methods ccccceeccceeecccesececeneceeeeeeeeeecsseeeceseeceeeneeees 25 Figure 8 The home lab including the Plugwise ZigBee mesh NetWwork ccccccssecccsseeeeeeeeseeeceseneseeeneeees 31 Figure 9 Average electricity consumption for one day using 5 months of data ccoocccccnccccnnccnnnacinonanoso 42 Figure 10 Average electricity consumption for one day using 3 months of data oocccccnccccnnccncnaccnonanaso 43 Figure 11 Average electricity consumption for one day using 1 month of data eese 44 List of Tables Table 1 Summary of device
10. M 15 55 13 55 14 50 16 500 18 00 25 55 33 05 55 55 Time of day B Average potential energy usage in kWh s x 0 001 g kWh usage ix 0 0013 e amp E 8 E amp E 8 amp 5 8 10 00 17 00 14 580 16 58 16 00 20 00 Er d 55 58 Time of day B Average potential energy usage in kWh s x 0 001 Proposed usage with a split tariff 3 month period 66 12 10 11 11 Too 7o X s rsu wiy Time of day M Td a 2 a a un LT E qi a m un I E qi E qu qi m m qi E AEST ERR RRR ARR RR a Too 7o X sheen My z i Existing usage with single tariff 1 month period 12 10 11 11 on FY Energy Optir 16 258 18 050 20 00 Er d 55 58 12 000 14 580 10 00 68 58 56 58 55 28 dr d 54 58 Time of day B Average energy usage in EWh s x 0 001 th split tariff 1 month period 67 ing usage wi Exist repere al M 18 05 20 00 rare in 55 58 18 58 20 00 rd 55 58 a a i i i i pay A AAA 16 50 16 54 14 0 PE EE Time of day Time of day i RB i mm mim em e i eo co c n i i s to ij n i ent tin sae i E E BA B Average potential energy usage in kWh s x 0 001 B Average potential energy usage in kWh s x 0 001 i inf nt h r ts nd an Jmf ao t t at bjj at tt jJn i D BR EE mmm dmm mmm fa ensis ir i i d vas cn et a a iN Gh d i Eos
11. Paris International Energy Agency 2009 97B 92 64 05953 5 3 Number of private households by household composistion number of children and age of youngest child eurostat Online European Commission European Union 1995 2012 June 26 2013 Cited July 1 2013 http appsso eurostat ec europa eu nui show do 4 Smart meter Wikipedia Online Wikipedia Foundation Inc October 17 2013 Cited October 18 2013 http en wikipedia org wiki Smart_meter 5 Smart Appliances Whirlpool Online Whirlpool 2013 Cited March 29 2013 http www whirlpool com smart appliances 6 Paolo Bertoldi Bettina Hirl Nicola Lanca Energy Efficiency Status Report 2012 European Commission and Joint Research Centre Luxembourg Publications Office of the European Union 2012 JRC Scientific and Policy Report ISBN 978 92 79 25604 2 7 P Bertoldi B Atanasiu Electricity Consumption and Efficiency Trends in European Union European Commission and Joint Research Centre Institute for Energy Luxembourg Publications Office of the European Union 2009 Status Report 2009 ISBN 978 92 79 13614 6 8 Smart grid Wikipedia Online Wikipedia Foundation Inc October 18 2013 Cited October 20 2013 http en wikipedia org wiki Smart_grid 9 The Status of Europe s Smart Grid Deployment Efforts Greentech Media Online Greentech Media Inc 2013 Cited October 18 2013 http www greentechmedia com articles read european smart gr
12. all appliances Remove standby power consumption from existing data readings for all appliances Add together total SA power consumptions Determine power consumption time to run and usage per session for each FA Assign slot priorities to FA based on average power consumption per session Determine schedules for FA a pi A uh ec s Calculate estimated savings in power consumption kWh and cost euro 19 4 5 Standby calculation Close analysis of the standby values versus the running values resulted in the following 3 rules 1 The value has to occur at least twice in a row and be followed by a zero reading no power consumption in at least one session 2 If the appliance is never turned off never has a zero reading the value has to occur at least two times in a row 3 The ratio of the value highest reading per session cannot exceed 7 A closer examination of the calculation is shown in Algorithm 1 where the numbers 1 2 and 3 correspond to the numbers of the rules above i hour 1 to 24 j day included in analysis days are selected by the end user x i consumption value at given hour and day j high highest value found low lowest value found y j temporary variable holding lowest values found for each day in analysis standby value standby value found in analysis Algorithm 1 f 1 j 1 repeat high x i j low x i j if high x i 1 j then high x i 1 j if low gt x i 1 j then low x
13. and 148 10 more in 5 months analysis Based on the above information it can be concluded that there is not a significant difference in the amount of data used in the analysis however if the data included does not detect a standby value of an appliance the difference can be as high as 19 In addition it can be concluded that without scheduling the appliances there is slightly less potential at least 1 30 less for cost savings if you use a split tariff but with scheduling the appliances there is a potential of 14896 or more savings by using a split tariff rather than a single tariff 6 1 4 Requirements verification Finally the functional requirements were verified by means of stating design decisions implementation classes and use cases All requirements defined have been verified as having been satisfied which allows me to say that the HEP provides the functionality it was designed to provide 52 7 Conclusion and further work have researched existing solutions such as Smart Appliances and labeling also researched other solutions documented which resulted in a selection of three Smart Home solutions All of these solutions aim for the same outcome to optimize electricity usage in a residence Two of the three solutions allowed the use of existing appliances while one required Smart Appliances One solution required a Smart meter and put a restriction on the type of appliances while another required the use of sensors in additio
14. and Sustainable Energy Reviews 2012 6851 6860 28 What is CE Marking CE Mark CE Marking Online Wellkang Tech Consulting 1996 2013 Cited March 25 2013 http www ce marking org what is ce marking html 29 CE marking Enertprise and Industry Online European Commission Cited March 28 2013 http ec europa eu enterprise policies single market goods cemarking 57 30 Introducing EU ENERGY STAR EU ENERGY STAR Lableing Energy Efficient Office Equipment Online European Commission Directorate General for Energy Cited March 28 2013 http www eu energystar org en index html 31 Communication from the commisssion on the implementation of the ENERGY STAR programme in the European Union in the period 2006 2010 EUR Lex Access to European Union Law Online 2013 Cited March 29 2013 http eur lex europa eu LexUriServ LexUriServ do uri COM 2011 0337 FIN EN PDF 32 International Conference on Energy Efficiency in Domestic Appliances and Lighting At the doorstep of the new EU Label Joint Research Centre Online May 24 26 2011 Cited April 4 2013 http re jrc ec europa eu energyefficiency EEDAL EEDAL11 Proceedings 21 EEDAL 2011 At the door step of the new EU Label Anton Eckl GfK Standards amp Labels2 pdf 33 Live from LG s CES 2011 press event engadget Online Engadget January 5 2011 http www engadget com 2011 01 05 live from lgs ces 2011 press event 34 Smart home appliances are a big d
15. each appliance or None if a standby value was not detected using the same appliances and the same data files in all cases By measuring these results was able to prove the extent of correctness as well as incorrectness of the 30 algorithm designed to detect standby values by calculating precision recall F measure specificity and accuracy 6 1 2 Optimization In order to evaluate the results of the optimization three time periods were processed 1 3 and 5 months using the same consumption data The possible flattening of peak energy usage resulting from processing appliances in the HEP was examined The results are quantified by the use of average absolute deviation from the mean power show comparative results between existing usage and suggested usage and compare the results in order to identify whether the time periods used in the analysis make a difference In addition the result of applying single or split tariff was examined by comparing the outcome of the scheduling as well as the monetary values resulting from the same appliance measurements 6 2 Experimental setup The application HEP was designed and developed using real life consumption data collected in a residence occupied by a family of 3 in the Netherlands Plugwise chapter 4 2 home automation system was used to measure appliances and collect consumption data In addition in the case of not being able to use the Plugwise home automation system to measure consumption r
16. from 1 3 and 5 months of data ooccccnnccnnnnccncnaninonaninnnacnnnnccnonanoso 46 Table 17 Outcomes of usage calculations performed by HEP ooocccccnnccnnncnocnnnnnaconnonanonnnnnncnnnnnncnnnonanonnos 46 Table 18 Outcomes of cost calculations performed by HEP occcconccnccnnccnnncnocnnnnnnccnnonanonnnnnaconnnnnornnonanonnos 46 Table 19 Percentage difference of usage calculations between periodS oocccccncccncnoncnonacnnonccnnnaninonanoso 47 Table 20 Percentage difference of cost calculations between single and split tariffs 47 Table Re ENES 48 Table 22 Summary of existing solutions and HEP ccccceeseccccesscceceesececeeececeeescceeeesecessueceeeegecestenseeetes 53 1 Introduction One of the biggest problems facing Europe today is the challenge caused by increased energy consumption and the increase of greenhouse gas emissions In response the EU has defined the EU 20 20 20 target which is aiming for a 2096 reduction in Europe s annual energy consumption a cut of at least 2096 in EU greenhouse gas emissions and that 20 of EU energy consumption comes from renewable energy sources by 2020 1 As energy costs are increasing more and more consumers are becoming actively interested in reducing their energy consumption New technologies available today have been claimed to have the potential to save at least 4096 of residential electricity consumption in most types of appliances 2
17. i 1 j 353 1 repeat if high x i 1 j then high x i 1 j if low x i 1 j then low x i 1 j if x i j 0 and x i 1 j O and x i 1 j x i 2 j and if y j lt x i 1 j then y j x i 1 j else if x i j O and x i 1 j 2 1 and x i 2 j O and x i 2 j x i 3 j and if y j lt x i 1 j then y j x i 2 j H1 until The end of 24 hours 2 Find lowest value if never powered off if low 0 then 1 repeat if x i j low and x i 1 j low then yj low until The end of 24 hours 3 The ratio of the value highest reading per session cannot exceed 7 if high O and y j high gt 0 07 then y j 0 j H1 until All days in analysis have been read standby value most frequent value in y 20 All appliances go through the standby calculation because when the data was examined it became apparent that appliances with standby power consumption would give a misleading reading in regards to the time period these appliances run in one session 4 6 Scheduling the appliances Appliances which are identified as FA by the end user are given a schedule time In order to optimize energy consumption it is important to try to assign high consumption appliances to different timeslots that do not overlap In the case of the end user with a split tariff contract assigning timeslots to the flexible appliances in off peak timeslots is of value The schedule times are determi
18. of hours needed to run one session of all Flexible Appliances based on FA X h Depending on the tariff it locates a continuous time period where the consumption data of the SA appliances is lowest if the tariff is a single tariff and assigns the times FA X Schedule In the case of a split tariff it takes the time period where the tariff is off peak and assigns the times FA X Schedule If the off peak time period is too short it calculates whether the overlapping time slots should be added before after or combined based on lowest consumption data from the Strict Appliances Then it adds each Flexible Appliance s consumption data per hour FA X kWh h and sends it to the Usage Savings component along with FA X Schedule the total kWh consumption per hour for all Strict Appliances SA Total kWh h and the tariff information Tariff A Java component which calculates existing usage and cost based on total existing usage Appliance Total kWh h received from the Consumption component and tariff information Tariff received from the Scheduler component In addition it calculates the following based on the stripped information received from the Scheduler component Proposed average power consumption usage per day kWh day Proposed kWh savings a year KWh year Whole cost of existing usage based on single tariff Cost of existing usage per day based on single tariff day Whole cost of proposed usage based on split tariff Cost
19. ri s i p v d d kd d Gt D jn Po eee Po eee LT n T 1 RR HR EES Too 7o X sheen wy G29 298999998 R R GR 4 TE Fan uma H Ay a Ix T Ay a Fr D ERMEER 55 5555 5 207 kWhjday 5 184 kwh day 8 442 kWh year 1 89 Efyear HE i 3933 i i Proposed usage with a single tariff 1 month period E E PA avings Existing average usage Estimated average usage Estimated kWh savings Estimated savings PROPOSED USAGE SPLIT TARRIF avings Existing average usage Estimated average usage Estimated savings Usage s Usage s MM Energy Optimization WF Energy Optimization Proposed usage with a split tariff 1 month period 68
20. still meeting its operational goal as a heuristic to determine its priority at any time An intelligent smart home gateway is the focal point of their system In addition the system uses Insteon 23 products to monitor and control the appliances The gateway receives consumption data from appliances via Insteon real time electricity prices generation data from renewable energy sources via Insteon as well as holding the scheduling policy In order to compute the slack for each appliance temperature and humidity sensors are placed inside or near each appliance Their method is able to flatten household demand over each day more exactly the absolute average deviation from the mean power was decreased by 2396 on the day they measured in a real home setting Solution Ill Smart Home Simulation Prymek et al 13 introduce a simulation model using smart appliances based on priority where high priority appliances are satisfied before the lower priority ones They satisfy the end user behavior by instructing them to set rules for the appliances 8 rules ranging from unnecessary time unlimited to health risk avoidance In addition they pre classify appliances according to their control mechanism user expectancies and power consumption profiles This simulation software can be used both with virtual appliances as well as with real appliances The purpose is to give the end user a way to simulate results of different scenarios depending on the end us
21. users that is also for residential consumer devices such as real time consumption and historical consumption In the survey over 9096 wanted a better overview of their electricity consumption as well as a better understanding of the impact individual devices could have on their energy bill and behavior By providing a means for the end user to visually monitor individual appliances as well as the potential for savings by changed behavior there is a potential for reduction in electricity consumption for the end user http www ict nobel eu 10 2 2 Smart Home solutions A trend of smart homes which can aid the inhabitants in energy consumption reduction has emerged This is achieved by ways of monitoring and controlling devices by rescheduling the operation times according to different criteria Several good solutions have been presented three of which are summarized below Solution SmartCap Due to the challenges involved in demand side management Barker et al 12 designed SmartCap which is a system that automatically monitors and controls household electricity usage by focusing on quantifying the benefits of scheduling transparent background electrical devices These background devices are air conditioner refrigerator freezer dehumidifier and heat recovery ventilation system HRV Their method is an online approach that uses each appliance s current slack the time period an appliance can be disconnected from electricity while
22. 18 15 16 17 39 In the development of this application the Home automation equipment from Plugwise 39 was used because it is easy to use and install as well as being successfully deployed in other experimental setup 10 Figure 4 A Plugwise network 27 The basic network modules Figure 4 consist of e Circle This is the actual plug which plugs in between the socket and the appliance plug It measures the energy consumption of the connected appliance stores the data and transmits it to the Source software using a wireless ZigBee mesh network In addition it switches the appliance wirelessly on or off according to switching schemes and or standby rules set in the Source e Circle This plug has the role of the coordinator in the network It keeps track of all the other Circles and communicates to the Stick In addition it contains a clock which 15 the Circles use to synchronize their time stamps Once it has been configured it can serve the role of a regular Circle e Stick This is a USB dongle which communicates wirelessly with the Circles via a ZigBee mesh network It receives power consumption data from the Circles and provides it to the Source as well as transmits the switching schemes and standby rules from the Source to the Circles e Source This is the software used to control the whole network It allows you to view your energy consumption as well as create schemes and standby rules to automatically switch appliance
23. 4 In Table 19 the first data column depicts the percentage difference between the three periods of existing usage in kWh where 1 month has the lowest amount and 5 months have the highest amount of kWh used The second column shows the difference based on existing usage per day Here 5 months have the lowest usage and 3 months the highest per day In the final two columns the scheduling calculations chapter 4 6 have been applied The third data column shows the difference in usage between the periods where 5 months have the lowest estimated usage and 3 months the highest In the final column which depicts estimated savings in kWh per year 5 months have the highest amount of estimated savings and 1 month has the lowest Table 20 Percentage difference of cost calculations between single and split tariffs Between Existing cost Estimated existing Estimated savings tariffs 3 cost year year 5months 1 31 1 3196 148 1096 3months 1 31 1 3096 151 5696 1 4096 1 4096 157 6596 The first two data columns in Table 20 are showing the percentage difference between using single or split tariff based on existing usage In all periods the split tariff shows a higher cost than single tariff The third and final data column depicts the percentage difference in predicted savings between the two 47 tariffs after the scheduling calculation Chapter 4 6 has been applied where again the split tariff shows higher savings than single t
24. Appendix II 61 Wl Energy Optimization gt Home Energy Planner Appliances Refrigerator C flexible Coffee maker C flexible Dishwasher strict flexible BELEIEEHE C flexible Select start date A O flexible I Tariff Single tariff TV strict flexible plan single tal i i 0 2241 Router iQ strict flexible i Vad Split tariff Fan on a timer strict flexible Offpeak tariff Peak tariff lo 2125 5 0 2387 S Washing machine strict a flexible SMTWTFS start end Clothes dryer strict flexible OMAMNO 07 00 2300 Split tariff 62 Appendix Il Screenshots from HEP of optimization evaluation Below are provided screenshots of the actual data used in the evaluations obtained from HEP MM Energy Optimization EXISTING USAGE SINGLE TARIFF Whole 793 402 kWh Average 5 186 kWh day 12 6 11 11 kWh usage ix 0 001 5 00 54 58 56 00 55 55 15 50 13 00 14 585 Time of day MM Energy Optimization EXISTING USAGE SPLIT TARRIF Whole 793 402 kWh Average 5 186 kWh day Period 12 6 11 11 Appliances used Refrigerator Coffee maker Dishwasher kWh usage ix 0 0011 Central heating Set top box TV tu 04 55 56 58 55 55 15 50 12 00 14 58 16 58 18 00 20 00 re 25 55 Time of day B Average energy usage in EWh s x 0 001 Existing usage with a split tariff 5 month period 63
25. DV gt Q Nov wy Ww QG QS gY YN Av Hours SSN Proposed Single O X c gt O NY A a E E E A B o Hours Y kWh x0 001 YODA AC SO LA 49 9 S9 NY x 23 Q Nov Nw Ww G GARN vy Hours ESE sh Y Figure 9 Average electricity consumption for one day using 5 months of data 42 gt Existing kWh x0 001 NS vA Qf AO Q9 NAN LDL DS NY VD gt Q Nov wy Ww QG QS gY YN Av Hours SSN Proposed Single O X c gt O NY AP a E E E A B o Hours Y kWh x0 001 Sv Am Q5 SO LA 49 9 S NY x 23 Q Nov mw Ww QG TGN vy Hours ESE E ES Figure 10 Average electricity consumption for one day using 3 months of data 43 gt Existing kWh x0 001 NS Vv 2 Qf AO 0 ND 0D DS NY xl gt VON VT Ww QG TGN Ay Y Hours FP wy sh ES Proposed Single a O X c x oY NY 4 XP 9 uo od ASQ B o Hours Y kWh x0 001 YODA AC SO LA 49 9 S9 NY x 23 Q Nov Nw Ww G GARN vy Hours ESE sh Y Figure 11 Average electricity consumption for one day using 1 month of data 44 Average absolute deviation from the mean power was used to quantify the flattening over an interval of one day s average power consumption usage in the same manner as Barker et al in their SmartCap project 12 The formula used was 47 2 labs mean T 1 1 where X X X24 is the dataset containing the consumption data existing proposed with single tariff and
26. Summary of devices and methods in relevant solutions Solution Solution Il Solution Ill Existing appliances X X o Smart appliances External measuring device ooo X Sensors sis X Appliance type restriction J No End user classification of appliancetype X Automatic control 0 X Show potential savings euro and or kWh X X Show potential difference in pricing plans Standby power detection removal J o While Solution allows for the use of existing appliances with an external measuring device it requires the use of sensors diminishes the control of the end user by keeping the scheduling fully automatic and puts a restriction of the type of appliances Solution II allows for all types of appliances and gives the end user the control of classifying them according to their own requirements but it requires Smart appliances Solution IIl uses existing appliances with an external measuring device but puts a restriction on the type of appliances and requires a Smart meter The HEP solution presented in this paper differs by adding standby power detection and removal from the total consumption of each appliance as well as providing differences in potential savings by using a single tariff plan or TOU tariff plan which none of the other solutions provide It also allows the use of existing appliances with an external measuring device without putting a restriction on the type of
27. University of Groningen Faculty of Mathematics and Science Software Engineering and Distributive Systems Monitoring and reducing the consumption of home electric appliances Master s thesis Student Sunna Bj rg Sigurj nsd ttir Primary supervisor Prof Dr ir Marco Aiello Secondary supervisor Prof Dr ir Paris Avgeriou ember 2013 Table of Contents Fo orn S 4 Eee c T 4 in T T TU TRE 5 FACTAE SI CUAL ON MT 6 LL SE ON CNG OU UU Utm 7 1 2 Problem statement and research QUESTION cccccessececceseccecesececseseceecenecceeeusecessunecessenceesauecetsuneces 8 1 3 Thesis contribution anden N rre 9 NNN 10 2 1 End user DENAVIOT arcade ini 10 22 SIMA 66 EE EEE E N 11 2 3 Discussion and Home Energy Planer contribution ooccccccoccnncnnocnnnnnncnnnonanonnnonacnnonononnnonaronnonanonnos 12 S 200 600 EEE NE 13 3 1 Labeling and regulations regarding energy efficiency in the EU cc ooccncccnnccnncnnncnnnnnacnnnnnaconnonanonoss 13 SPAM A e EE EEE TE 14 Fo NeW Smart a INC tel 14 3 4 Old appliances A Home Automation solution ssec 15 TETEN TNNT 17 1 1 System contextand requirements ota dee pepe dia 17 FEN 18 12 DENN NN 19 TTS EN NNN Need 19 TPM 20 4 6 Scheduling the appliances rrnrrrnnnnnnrrnrnnnrrnrnnnernnnnnernnnnnnrnnrnnnennrnnnsvnnnnnennnnnnesnnnnunsne
28. WTA 4000 handleiding 5 75067 html 58 43 Bosch waa 24170 Berlina 1200 gebruikers handleiding com Online Gebruikershandleiding com http www gebruikershandleiding com Bosch waa 24170 berlina 1200 handleiding 5 305380 html 44 Source Software Manual Plugwise Online http www plugwise com sites default files plugwise user manual en O pdf 45 Enkel of dubbeltarief only available in Dutch essent Online 2011 http www essent nl content particulier producten alle energietarieven enkeltarief of dubbeltarief i ndex html 46 Precision and recall Wikipedia The free encyclopedia Online Wikipedia Foundation Inc May 13 2013 Cited May 22 2013 http en wikipedia org wiki Precision and recall 47 Absolute deviation Wikipedia The Free Encyclopedia Online Wikipedia Foundation Inc April 22 2013 Cited May 22 2012 http en wikipedia org wiki Absolute deviation 48 Commission Regulation EU No 1015 2010 implementing Directive 2009 125 EC of the European Parliament and of the Council with regard to ecodesign requirements for household washing machines EUR Lex Access to European Union law Online 2013 Cited March 28 2013 http eur lex europa eu LexUriServ LexUriServ do uri 0J L 2010 293 0021 0030 EN PDF 49 Dehems Digital Birmingham Online Birmingham City Council 2010 Cited May 2 2013 http www digitalbirmingham co uk projects dehems 50 Optimizing Energy Costs for Offices Conne
29. ach figure shows three graphs representing existing usage proposed usage with a single tariff and proposed usage with a split tariff during one day Each graph represents in blue the average usage vertical axis per hour horizontal axis based on the data calculated over the respective periods To further explain the value in the first hour 0 1 Hours of existing usage represents the total consumption of all measured appliances which were running in the first hour each day divided by the number of days in the selected period The first hour of the single and the split tariff graphs show the predicted outcome of average usage after the scheduling calculation Chapter 4 6 has been applied In order to detect the flattening of peaks the mean for each of the three datasets is displayed in red in each graph This allows for a clear comparison of peak magnitudes and duration of peak periods As an example when viewing Figure 9 one can see that the existing usage has one peak with duration of 14 hours 9 00 23 00 and a maximum magnitude of 0 393 kWh 20 00 21 00 The proposed single tariff usage has two peaks with duration of 2 1 00 3 00 and 4 hours 20 00 24 00 and a maximum magnitude of 0 902 kWh 20 00 21 00 and finally the proposed split tariff has three peaks with duration of 2 hours each 4 00 6 00 20 00 22 00 23 00 1 00 and a maximum magnitude of 0 778 kWh 23 00 24 00 41 gt Existing kWh x0 001 NS vA Qf AO Q9 ND 0D OS NY
30. application such as HEP Home Energy Planner will be referred to as HEP from this point in the remainder of the document 9 2 Related work When faced with electricity trends and the design of the new Smart Grid several solutions to energy savings have been presented T A Nguyen and M Aiello summarize and compare several studies on building energy and comfort management BECM systems 19 where simulation results show up to 58 savings on energy for lighting and 10 4096 for HVAC system In the commercial sector a study conducted by Georgievski et al 10 showed an average economic savings of about 35 by presenting an approach based on measuring consumption on individual devices as well as measuring the production on generating units In addition end users where able to determine requirements which lead to policies for each device and energy contracts were closed for short term time intervals dynamically from different providers This approach assumes that the Smart Grid is available and that it is able to offer dynamic prices from different energy providers In the following two sections we will take a closer look at solutions presented where some deal directly with the Smart Grid assuming that the end user is connected to the Smart Grid has a Smart Meter and Smart Appliances and others focus on the behavior of the end user and what role they have in energy and consumption savings 2 1 End user behavior We can t be using that much I
31. ariff 6 5 Requirement verification A set of functional requirements were determined in order to capture the desired functionalities of HEP Table 20 depicts the verification of each of the requirements Table 21 Requirement verification ID Functional Requirement Verification Home Energy Planner HEP px ceste The HEP shall find standby values of end user Satisfied by standby algorithm defined in defined appliances when applicable chapter 4 5 d o The HEP shall provide a schedule with proposed Satisfied by scheduling process defined in running times of end user defined appliances chapter 4 6 within a 24 hour period m uae o a a r appliances chosen AppliancesToUse in chapter 5 2 Tb apptanesTouse metopers2 O on single tariff and or split tariff AppliancesToUse in chapter 5 2 hal He Satisfied by implementation of class chosen AppliancesToUse in chapter 5 2 TE velaneestemdapersz appliances chosen AppliancesToUse in chapter 5 2 appliances chosen AppliancesToUse in chapter 5 2 hell The HEP shall show results lexically and graphically LayoutCombined in chapter 5 2 The HEP shall be easy to use in terms of Satisfied by implementing alerts when end supplying the end user with information on user has not entered all information input needed in order to complete analysis needed see Use case 5 3 2 The end user shall have control over which Satisfied by appliance consumption data appliances are used in analysi
32. ariffs show a decrease in peak time periods when compared with existing usage which allows us to conclude that HEP does reduce peak time periods Calculating the average absolute deviation from the mean results in flattening of peaks for split tariff compared with existing usage but an increase of peaks for single tariff compared with existing usage Based on this could conclude that HEP does only optimize consumption usage when a split tariff is being used but in the following section will examine further the reasons behind this 6 1 2 1 Reasoning for high magnitude peaks When examining the graphs of proposed single and split tariffs it was noted that they all yield one peak with a great magnitude The reason for this peak is that the graphs are displaying 24 hours of consumption usage based on an average of the whole usage during the time periods This means that in both cases am displaying one session of each flexible appliance at the same hour every day it is scheduled compared to different times during the day for each day of existing usage To explain this further both tariffs schedule the clothes dryer as the first appliance in their schedule since this is the appliance which has the highest consumption The clothes dryer s total consumption data ranges from 2 5 2 9 kWh and takes 2 hours to run each session This means that the first 2 hours in a schedule will enclose 2500 2900 kWh x0 001 each time the clothes dryer ran for one session w
33. box 4 HEP displays the previous view with the empty rule field as well as other previously correctly selected fields 29 6 Evaluation It is important that the HEP gives the end user correct and accurate results since the main objective of the system is to aid the end user in optimizing energy usage Quantitative evaluations were performed as well as a verification of the requirements listed in chapter 4 1 6 1 Objective and metrics used In order to evaluate the HEP the two major functions of HEP were examined 1 Standby power detection of appliances by the HEP 2 The extent of optimization resulting in applying schedules suggested by the HEP What is the result of the optimization scheduling How much data is needed 6 1 2 Standby power detection Standby power detection is based on the 3 design choices rules 1 3 defined in chapter 4 5 By applying Algorithm 1 chapter 4 5 which resulted from these design choices to the appliances that were measured was able to determine the standby power on five the coffee machine dishwasher washing machine stereo system and fan on a timer out of six appliances which measured standby power The appliance the CD player which was not identified had very little difference in active usage power consumption and standby power consumption If removed rule number 3 in order to detect this appliance two more appliances were wrongly identified to have standby power the central heating system and DVD p
34. but not all consumers are willing or in the position of acquiring such appliances whilst their old appliances are still serviceable There were 211 922 500 households 3 recorded in the EU 27 alone in 2012 A vast majority of these households can be assumed to be equipped with the majority of needed appliances With the coming of Smart meters a meter that records electrical consumption in intervals of an hour or less as well as communicating that information back to the utility for monitoring and billing purposes 4 and Smart appliances a household appliance that is able to communicate with other smart appliances as well as the electrical grid in order to reduce power consumption and increase the quality of life 5 it is evident that these will not replace old equipment in all of these households today We need a way for consumers to manage their consumption whilst using their existing appliances In addition we are surrounded by more and more electrical appliances and gadgets in such a way that we are unaware of the consumption around us Many households can count between 20 and 30 separate electronic devices spread within their walls 2 These appliances become ubiquitous to us in the sense that we do not even realize that we are consuming electricity as many appliances today are left on a standby mode such as the case of several types of multimedia appliances surrounding the TV which might not be turned off manually each time the TV is turned
35. ch load can vary within 2 7 0 2 kWh and assigned randomly generated numbers divided into the documented time slots 6 2 3 Data collection during experiment A Plugwise mesh network was set up as seen in Figure 8 Energy consumption data was collected over a period of 5 months continuously in an existing residence of 3 persons in the Netherlands on the following appliances 1 Refrigerator 2 Coffee maker with bean grinder 3 Oven 4 Dishwasher 5 Central heating system 6 Set top box 7 DVD player 8 TV In addition to the continuous reading by the Circle s each appliance s state was recorded in order to detect actual standby power readings In order to get a greater variety the following appliances were measured while plugged in over a period of 1 week In addition each appliance s state was also recorded i e Film recorder charging Film recorder fully charged 9 Router 10 Fan on a timer 11 Hand vacuum cleaner charger 12 Stereo system 13 CD player 14 Wireless doorbell unit 15 Bread toaster 16 Film recorder charger 17 Radio The router and the fan on a timer appliances 9 and 10 respectively were observed as appliances running on a pattern as explained in chapter 6 2 1 which was applied to their readings in order to produce 5 months of data 33 The following appliances power consumption was documented by way of the runtimes of each session used and program used over the same 5 month perio
36. cted but an on button needed to be pushed in order to activate the session In addition when the session was completed the light remained blinking until the program knob was turned to off The readings before and after activation of the session returned 2 the test was completed for three hours both before and after running a session e Washing machine appliance 18 was observed to perform in the same manner as the dishwasher with a blinking light indicating that a program had been selected an on button to activate the session and a blinking light remaining lit until the program knob was turned to off This resulted in the assumption that it had a standby value which was determined to be 2 the same as the dishwasher Appliances determined not to have standby values e Refrigerator appliance 1 and central heating system appliance 5 are appliances which are constantly running and cannot be turned off by use of the Plugwise standby killer alone since they would need an additional mechanism to be turned on again automatically Therefore they are both noted as not having a standby value in this evaluation The refrigerator did not return a low value larger than zero which could indicate a standby value but the central heating system returned the value 7 constantly for periods of one to four hours which could indicate a standby value Since we did not desire to detect this value rule 3 The ratio of the value highest reading per session cannot exc
37. cted to the Smart Grid I Georgievski V Degeler G A Pagani T A Nguyen A Lazovik and M Aiello 4 Groningen IEEE Transactions on Smart Grid 2013 Vol 3 2273 2285 59 Appendix I Images explaining the Graphical User Interface of HEP When the application is started the end user adds the appliances to use in the analysis by selecting the Add appliance button on the start up page FW Energy Optimization Home Energy Planner Existing energy use Proposed energy use Appliance list Sawe tariff Start up The end user is able to choose which appliances to include in the analysis and uploads the csv file Wl Energy Optimization EN Refrigerator EN Refrigerator short EN Router EN Set top box EN Stereo EN Toaster si TV Copy ig TV E TV old sg Doorbell s Washing machine 4 DVD player El Fan on a timer EN Film recorder Save tariff Select appliance to upload file 60 When the appliances have been uploaded they are listed on the left Immediately following the name of each appliance are two radio buttons where each appliance is classified either as strict or flexible The table displays one week consumption for one appliance The appliance chosen can by changed by clicking on the name in the appliance list In addition the end user chooses the date to start the analysis on below the table which in turn determines the start of the week displayed in the table above The final input needed
38. ctrical technologies customer participation and behaviors One of the largest is the Pecan Street Project 11 which is a smart grid research project founded in the city of Austin Texas in 2008 The residents of the Mueller community in Austin are early adopters residing in 1000 homes Some of these homes are green built houses and others are traditional built houses The houses are equipped with energy management systems and most of the advanced smart grid technologies are incorporated The Pecan Street Project is used to evaluate different smart grid standards in terms of interoperability analyze different dynamic pricing models as well as study the effects of incorporating PHEV s solar panels energy storage and smart appliances The residents are rewarded by financial assistance when purchasing equipment such as solar PV cells solar water heaters and efficient air conditioners as well as providing rebates for home efficiency measures 1 2 Problem statement and research question The emergence of the smart grid smart meters and smart appliances provide consumers of electricity with new means to observe their energy consumption more closely and to have the option to change their consumption habits to improve on their savings With increased awareness of the possibilities of decreasing electricity consumption and the possibility of saving money on their electricity bill residents of existing homes should be looking for solutions With an exist
39. ctricity companies this is done on line via connection to the electricity company and therefore there is no need to be connected to a Smart Grid or having a Smart Meter so that the end user can choose a cost effective time to run the appliance In addition they are equipped with a Smart Delay feature which schedules energy intensive tasks at cost effective times such as running the defrosting of the freezer or running the dishwasher 36 In 2012 LG came out with NFC tagged appliances LG Smart ThinQ line consists of washers dryers refrigerators and an electric oven range with grill 37 These appliances are equipped with a full range of Smart ThinQ technologies where they can be monitored from a Smart TV or smartphone via their Wi Fi device to device connectivity In addition LG offers a Home Energy Management System HeMS which aids the end user to manage their smart appliances lighting and HVAC in a more power efficient manner 38 The HeMS collects the energy usage via a smart meter which can be viewed on a PC or smartphone and suggests ways to reduce overall energy consumption This solution only works for end users that have a smart meter and are connected to a Smart Grid enabled power company 3 4 Old appliances A Home Automation solution There are several companies offering solutions in home automation allowing the use of existing appliances by means of reading the energy usage information via the socket of individual appliances
40. d as for appliances 1 8 above as explained in chapter 6 2 2 18 Washing machine 19 Clothes dryer 6 2 4 Data processing The Plugwise user manual states that the consumption data can be exported to a CSV file 44 but that facility is only available in the Source Pro which did not have access to Therefore manually inserted the measured data from Source to the CSV files The structure of the file is assumed to be the same as the table views in Source Table 6 where the first column represents weekdays and columns 2 25 represent 24 hours in one day Table 6 Data representation in table format as shown in Source v 32 8 47 38 4s 25 a2 2 45 27 22 6 24 50 26 59 25 23 Sa 24 35 25 w sa sr ss 40 30 aa 33 se 28 as 25 60 25 an ss se ss aa aa 7 ss 55 ae 4 HWEXEJEJESEIELTEJESEZESEA3EJEMTEJEJEJLIESEJEAETEZETES AMEGEJESEQEMEXEJEJEJEXESESEAESLIEOEEZEAESEJESEES The consumption data from Source was manually inserted into CSV files named after each appliance type using the following structure Table 7 Structure of the CSV files The following is repeated for each week used in analysis Row Column Data type timeslots date weekday usage readings per hour shown in orange above The data for the router the fan on a timer the washing machine and the clothes dryer were inserted in the same manner using the measured timed and calculated values as described in chapter 5 1 6 2
41. d in 2006 and is an appliance specific label focusing on office equipment identifying equipment that meets certain energy efficiency standards This has proved to be effective in moving the market to a greater efficiency 31 between the years 2008 and first half of 2009 the percentage increased from 4596 to 6696 in EU ENERGY STAR registered models In the second half of 2009 a new more demanding specification was implemented which resulted in a decrease down to 3096 of EU Energy Star registered models but by the end of 2010 it had increased again to 5096 3 2 Incentives in the EU Several countries within the EU are offering subsidies and or tax benefits to end users who purchase energy efficient appliances In January 2007 the Italian Government introduced a tax subsidy program in order to promote the sales of highly efficient cold appliances End users who bought an A A refrigerator or freezer could deduct 2096 of the appliance cost up to a maximum of 200 euro from their income tax As a result there was a growth in the sales of A class cold appliances by a factor of 2 5 in 2007 compared to the previous year The share of A appliances in sales reached 45 596 in 2008 and 6296 in 2010 6 In 2009 and 2010 Austria ran a scrapping bonus program on cold appliances during two periods September December 2009 and September November 2010 where the end user could apply for reimbursement when replacing an old appliance with a new A appliance up
42. difference between 1 and 5 months is only 0 5096 The explanation for this is different amounts of existing usage in different time periods By applying the scheduling of Flexible Appliances single and split tariffs result in a decrease by 0 44 0 4996 and 0 5496 in kWh per day 1 3 and 5 months respectively This yields estimated savings of 8 442 9 498 and 10 23 kWh per year The difference of 19 1596 between 1 month analysis and 5 months analysis is caused mainly by one appliance the coffee maker not being detected as having a standby value in the 1 month analysis because when calculating the estimated savings in kWh the only difference between existing usage and usage after applying the schedules is that standby power values have been removed from the appliances For the same reason standby values for the coffee maker are not removed the 1 month analysis 8 442 results in the lowest amount of savings in kWh per year but not the 3 months 9 498 which one could assume since 3 months yield the highest estimated usage in kWh per day When examining the outcome of the cost of existing usage between single or split tariffs the percentage difference is very similar for all three time periods the split tariff is 1 3196 higher than the single tariff in 3 and 5 months analysis and 1 4096 higher in 1 month analysis The reason for this is that the single tariff is the same or 0 2241 per kWh during all hours of the week but the split tariff is sli
43. duler which suggests times to run appliances in order to flatten peaks and reduce peak periods explain the design and concepts implemented in the development of HEP do extensive evaluation on the work completed in HEP in regards to flattening of peaks reducing peak periods and standby power detection which proves to be effective in HEP The content of this thesis is organized in the following manner In Chapter 2 related work is researched look at published work in regards to end user behavior as well as inventions regarding Smart Home solutions Chapter 3 presents the background where research existing solutions in labeling and incentives new smart appliances and solutions provided for existing appliances Chapter 4 describes the design and methodology of HEP This includes the system context algorithm for standby power detection scheduling of appliances and how the data was collected Chapter 5 is dedicated to the implementation of HEP where we describe the main components of the system the logical view as well as describing the most used runtime views In Chapter 6 the evaluation of the HEP is presented This includes the experimental setup the objective results from all the experiments and concludes with a discussion of the results Finally Chapter 7 presents further work and conclusion summarize my findings during research and evaluation of HEP and propose further development in the design and development of an optimization
44. eal at CES 2013 TechHive Online IDG Consumers amp SMB 1998 2013 Cited March 24 2013 http www pcworld com article 2024393 smart home appliances are a big deal at ces 2013 html 35 NXP builds smart washing machine with NFC and fabric detection Pocket lint Online Pocket lint Itd March 12 2012 Cited March 24 2013 http www pocket lint com news 114566 nxp smart washing machine nfc rfid 36 Whirlpool 6th Sense Live Appliances Have Arrived Abt Online Abt Exclusive March 8 2013 Cited March 29 2013 http blog abt com 2013 03 whirlpool 6th sense live appliances 37 Discover LG SMART THINQ Appliances LG Life s Good Online LG 2013 Cited March 29 2013 http www lg com us discover smartthing thing jsp 38 2012 CES sneak peek LG SMART THINQ technology for home appliances LG Life s Good Online LG http www lgnewsroom com ces2012 view php product code 96 amp product type 96 amp post index 188 6 39 Plugwise Online Plugwise http www plugwise com idplugtype f 40 home ZigBee Alliance Online ZigBee Alliance 2013 Cited May 4 2013 http zigbee org Home aspx 41 How to set up your Plugwise network Plugwise Online Plugwise Cited May 4 2013 http www plugwise com sites default files how to set up your plugwise network en 20100903 web pdf 42 Bosch WTA 4000 gebruikers handleiding com Online Gebruikershandleiding com http www gebruikershandleiding com Bosch
45. eal life usage of appliances was timed and recorded 1A db Circles Ec 4 Edi Ta anl El JE P Pa 3 V a sf P T v ov Pas Figure 8 The home lab including the Plugwise ZigBee mesh network Figure 8 shows the actual floor plan of the home lab with the setup of the appliances measured which are identified as follows Circle Refrigerator 1 Coffee maker with bean grinder 2 Hand vacuum cleaner additional appliances 3 Fan on a timer additional appliances 4 Oven 5 Dishwasher 31 6 Central heating system 7 Set top box additional appliances 8 DVD player 9 TV 10 Router 11 Stereo system additional appliances 6 2 1 Measured appliances During the design and development of HEP used one Circle and seven Circles In order to get a larger variety of appliance readings as well as more accurate total appliance consumption in 24 hours the data collection began by identify appliances which had consistent usage and therefore the runtime without using a Circle could be documented Fan on a timer Circle 3 was running on a timer and therefore showed the same reading every day four hours running with 15 16 kWh x0 001 for three hours one hour with a 10 kWh x0 001 reading and the remaining twenty hours had a reading of 1 kWh x0 001 The set top box Circle 7 had a consistent reading of 8 kWh x0 001 when plugged in no standby setting available In addition it was c
46. eed 796 was applied to the standby detection algorithm in chapter 4 5 e Oven appliance 3 bread toaster appliance 15 and clothes dryer appliance 19 are all appliances which are turned on show a different reading every time which is quite high and are either turned off oven or turn themselves off with a O reading when a session is completed bread toaster and clothes dryer e Set top box appliance 6 and router appliance 9 are appliances which need to be turned on and off with a button and have the same constant reading when turned on and O when turned off An exception did occur in some instances in the first hour when turned on and or the last hour when turned off of a session where the value was lower or higher It was determined that 35 this was the cause of turning the appliance on off in different timeframes within an hour since when these appliances were tested for exactly three hours starting at 1 minute past the first hour and ending at 59 minutes past the third hour this variation did not occur DVD player appliance 7 showed a reading of 4 or 5 when turned on whether there was a disc playing or not An exception did occur in some instances in the first hour when turned on and or the last hour when turned off of a session where the value was lower or higher It was determined that this was the cause of turning the appliance on off in different timeframes within an hour since when these appliances were tested for
47. een the ability of the test case applied to identify negative results tp tn A EE LI So tn fp fn 39 Accuracy is a parameter for the test cases It gives the proportion of both true results measured tp and tn Table 11 below presents the results from the calculations Table 11 Precision recall F measure specificity and accuracy of standby value test cases Precision Recall F measure Specificity Accuracy Test case 1 Test case 2 Test case 3 Test case 4a Test case 4b Test case 5 Test case 1 shows the results when the algorithm is used in its completeness and test case 5 shows the results when a standby number is identified by simply finding the highest occurrence of the lowest value found for each appliance without applying the algorithm The other test cases are designed around the rules in chapter 4 5 by omitting each rule from the algorithm in each test case This is reflected in the different outcomes depending on what is being measured As an example we recall that rule 1 says that the value has to occur at least twice in row and be followed by a zero reading this ensures that the value is at the end of one running session Often appliances had a very low reading as the first and last hour of reading in a session due to the appliance being turned on or off when only few minutes where left of the hour These values can easily be identified as the standby value if this rule is omitted and therefore g
48. er 1 00 400 6 4 2 Calculation results existing and proposed usage and savings The results of electricity usage and savings given by HEP were then recorded Table 17 shows the results of the calculations performed by HEP displayed for the same three periods used above 1 3 and 5 months Table 17 Outcomes of usage calculations performed by HEP Existing Existing Estimated Estimated usage usage usage savings kWh kWh day kWh day kWh year Tariff Both Both Botht Both 5months 793 402 5 186 5 158 103 gt 3months 486 944 5 293 15267 9 498 Single and Split Table 17 shows the outcome of the calculations regarding electricity usage and savings in kWh The results of the calculations are the same for both single and split tariffs in all cases and are therefore labeled Both in the Tariff row The first two data columns show total existing usage in kWh s and the average per day based on the total existing usage divided by number of days in the dataset The following column is the outcome from the calculations performed after the scheduling calculation chapter 4 6 has been applied that is the estimated kWh usage per day The final column shows the predicted savings in kWh per year based on the outcome of the scheduling calculation Table 18 Outcomes of cost calculations performed by HEP Existing cost Estimated existing Estimated savings 3 cost year year Tariff Single
49. er 2013 Vol 56 244 257 20 J Dobbyn G Thomas Seeing the light the impact of micro generation on the way we use energy Qualitative research findings London The Hub Research Consultants on behalf of the Sustainable Consumption Roundtable 2005 21 Low energy dwellings the contribution of behaviours to actual performance Zachary M Gill Michael J Tierney lan M Pegg amp Neil Allan 38 5 s l Building Research amp Information 2010 491 508 22 A Survey Towards Understanding Residential Prosumers in Smart Grid Neighbourhoods P Concalves Da Silva S Karnouskos D Llic Berline 3rd IEEE PES Innovative Smart Grid Technologies Europe 2012 978 1 4673 2597 4 23 Insteon Online Insteon 2013 Cited April 4 2013 http www insteon com 24 Eco design of Energy Related Products Energy Efficiency Online European Commission Cited January 7 2013 http ec europa eu energy efficiency ecodesign eco design en htm 25 Commission Regulation EC No 244 2009 implementing Directive 2005 32 EC EUR Lex Online 2013 Cited January 7 2013 http eur lex europa eu LexUriServ LexUriServ do uri 0J L 2009 076 0003 0016 en PDF 26 Household appliances Labeling Energy Efficiency Online European Commission Cited January 13 2013 http ec europa eu energy efficiency labelling labelling_en htm 27 Sustainability at home Policy measures for energy efficient appliances Kelly G 16 s l Renewable
50. ergy optimization functions This puts an emphasis on alternate solutions where the end user is able to measure and schedule existing appliances e What types of energy measurement solutions are available for existing appliances today The European Union and individual countries are actively implementing regulations and directives where the main goal is to reduce energy consumption in all sectors which has led to a great improvement in energy efficient appliances Several companies are providing Home 53 Automation Systems which allow the end user to measure the energy consumption at individual appliance level and provide the end user with the means to be aware of their own consumption have designed implemented and evaluated the Home Energy Planner the main purpose of which is to aid the end user in creating schedules and detecting standby power on existing appliances measured with Plugwise This enables me to answer the remaining sub questions which relate to HEP e Can we create effective standby power detection on the appliances measured designed and implement an algorithm which detects standby values on the measured appliances In my evaluation of the algorithm was able to prove that it produces 95 accuracy 10096 specificity and an F measure of 0 91 This is a highly effective standby power detection algorithm which can be used on the appliances measured e Can we reduce peak periods and flatten peaks using the optimization application
51. ers expectations as well as results from using different types of schedulers Solution III Residential Energy Management scheme Han et al 14 propose a wireless sensor network based Residential Energy Management scheme for non urgent appliances The non urgent appliances are e g washing machine clothes drier dishwasher pool pump and plug in hybrid electric vehicle PHEV Their scheme is based on time of use TOU tariff Zigbee transmission technologies and uses a smart meter to measure the appliances When the end user wishes to turn on a non urgent appliance their algorithm checks if power is available in the renewable energy generations and storage battery If it is available it runs the appliance if not it calculates the price of running the appliance now and at a later time and gives the end user the postponement option together with the different prices Based on this information the end user makes their decision of 11 running the appliance now or at a later time This means that the end user is deciding on when to turn on the appliance based on one running cycle 2 3 Discussion and Home Energy Planer contribution When comparing the Smart home solutions presented in the previous section we see that all solutions aim to reduce electricity consumption peaks and thereby optimizing the electricity consumption Table 1 provides an overview of the methods and technologies used in all three solutions for comparison purposes Table 1
52. exactly three hours starting at 1 minute past the first hour and ending at 59 minutes past the third hour this variation did not occur TV appliance 8 showed a reading of 44 or 45 when turned on and a reading of O when turned off both with a button on the appliance and with a remote control An exception did occur in some instances in the first hour when turned on and or the last hour when turned off of a session where the value was lower or higher lt was determined that this was the cause of turning the appliance on off in different timeframes within an hour since when these appliances were tested for exactly three hours starting at 1 minute past the first hour and ending at 59 minutes past the third hour this variation did not occur Radio appliance 17 showed a constant reading of 3 at all times when turned on and O when turned off Hand vacuum cleaner charger appliance 11 and film recorder charger appliance 16 are both chargers which showed a constant reading when the appliance was positioned in the charger unit and O when it was taken out of the charger unit An exception did occur in some instances in the first hour appliance in charger unit and or the last hour appliance removed from charger unit of a session where the value was lower or higher It was determined that this was the cause of inserting and removing the appliance in different timeframes within an hour since when these appliances were tested for exactly three
53. f information is sent to the Scheduler component Tariff Existing consumption for each appliance Consumption X is received from the Consumption component and displayed in table format Consumption A Java component which reads the consumption data from the uploaded files chosen in the Appliance GUI component It adds together the total existing kWh h consumption for all appliances and sends it to the Usage Savings component Appliance Total kWh h It sends the existing consumption for each appliance Consumption X to the Appliance GUI component and Standby Detection component In addition it sends the appliance classification Classification X to the Standby Detection component Standby Detection A Java component which receives existing consumption data on all appliances from the Consumption component It detects the standby values of all appliances and sends them to the Result GUI component Appl X Standby In addition it removes the detected standby value from the existing consumption data for each appliance After removing the standby consumption from all appliances it sends each Strict Appliance stripped consumption data SA X to the Strict Appliances component and each Flexible Appliance s stripped consumption data FA X to the Flexible Appliances component Strict Appliances A Java component which receives each Strict Appliance s stripped consumption data SA X from the Standby Detection component and adds the total stripped cons
54. from the end user is the tariff information The screenshot below shows how the end user has inserted the tariff for a single tariff as well as for the split tariff By not having the check box Split tariff selected the information for the split tariff is dimmed and no included in the calculation MM Energy Optimization Home Energy Planner Appliances Clothes dryer Refrigerator i stri C flexible S Coffee maker i i C flexible Dishwasher E strict flexible ulel JAE Central heating 9 strict j flexible min Select start date Set top box strict j flexible 12 6 z Tariff plan Single tariff lo 2241 TV strict C flexible Router strict flexible is iih NE F Split tariff Fan on a ti 5 strict a flexibl z an on a timer j stri 3 flexible offpeak tariif Peak tariff E kWh ES E kWh 0 2387 Washing machine 0 stri flexible SMTWTEFS stat end Clothes dryer strict a flexible 07 00 23 00 Proposed ener se seve ttt Single tariff The final screen shot below shows how the end user has selected the check box Split tariff and thus the information needed for a split tariff is visible and editable and the single tariff text box is dimmed and disabled Finally when the end user has given all information needed and selects the buttons Existing energy use or Proposed energy use the calculations are performed and the results are shown as can be seen in the screen shots in
55. ghtly more expensive or 0 2387 peak Monday through Friday from 7 00 23 00 and slightly less expensive or 0 2125 off peak Monday through Friday from 23 00 7 00 and on Saturdays and Sundays The refrigerator and central heating system were the only appliances which ran 24 hours The appliances which were highest in consumption readings see Table 8 the oven the dishwasher the washing machine and the clothes dryer all ran their sessions on Mondays Fridays during the peak period with the exception of the dishwasher ending two sessions between 23 00 0 00 both sessions are only included in the 5 months analysis All other appliances ran in the time period from 8 00 23 00 with the exception of the TV the set top box and the router which would often run until 1 00 Thus one can see that the majority of the electricity consumption took place between 8 00 and 23 00 during all days of the 51 week Since the tariff is higher during this period on five days out of seven it is clear that the existing cost using a split tariff will result in a higher total price On the other hand when examine the difference in estimated savings per year after applying the scheduling and using a single or split tariff the outcome is the opposite The percentage difference results in a higher outcome for the split tariff in all three time periods the split tariff saves 157 65 more than the single tariff in 1 month analysis 151 56 more in 3 months analysis
56. he previous view with the empty day fields as well as other previously correctly inserted fields 1a Nothing has been selected in the start dropdown box 2 An alert box with the message You have to select start time for a split tariff is displayed 3 End user clicks the OK button on the alert box 4 HEP displays the previous view with the empty start fields as well as other previously correctly inserted fields 1a Nothing has been selected in the end dropdown box 2 An alert box with the message You have to select end time for a split tariff is displayed 3 End user clicks the OK button on the alert box 4 HEP displays the previous view with the empty end fields as well as other previously correctly inserted fields 5 3 3 View existing usage When the end user wishes to view existing consumption usage based on chosen appliances and time period the following steps in this use case need to be taken This will give the information regarding usage in kWh in both written and graphical format along with a list of appliances chosen If estimated cost information is also desired use case 5 3 2 needs to be completed in addition This use case can be performed at any time after an appliance has been chosen or added Name View existing usage Success Scenario Alternative Flows 1 End user clicks the Existing energy use button 2 HEP displays existing energy use in a graph and in written format 2a Use case 5
57. hich was on average every two to three days This consumption is added to the existing usage of the Strict Appliances This resulted in the highest peak in the range 902 969 for the single tariff and 778 847 for the split tariff The positioning of these highest peaks is consistent with the scheduling times of the clothes dryer 20 00 21 00 and 19 00 for single tariff and 23 00 for split tariff Similarly when examine the reason for the different results in average absolute deviation from the mean where the split tariff results in flattening of peaks but the single tariff increases them need to take a closer look at why they differ Both tariffs yield the same appliance order in their schedules clothes dryer washing machine dishwasher and fan on a timer but different time schedules 20 00 9 00 21 00 10 00 19 00 8 00 on the single tariff and 23 00 12 00 on the split tariff The schedule is created by moving one session of a Flexible Appliance to the assigned time slot If there are additional sessions of the Flexible Appliance in one day they are added to the Strict Appliance s existing usage when the graph is displayed As an example one day which is included in all appliances has the clothes dryer running one session at 18 00 and another session at 21 00 This means that the session running at 18 00 will be schedules in the first time slot of each schedule which is 20 00 21 00 and 19 00 for the single tariff and 23 00 for the split tariff
58. hours starting at 1 minute past the first hour and ending at 59 minutes past the third hour this variation did not occur Wireless doorbell unit appliance 14 is an appliance which is always plugged in and has no on off function This appliance showed a reading of 2 and 3 at all times 36 Table 8 Appliance readings and classifications based on outcomes of data measurements ID Appliance ASV Range Zero reading On Off Type 2 Coffee maker with bean grinder 4 1 103 N 3 Oven None 465 1468 N 4 Dishwasher 2 1946 N 5 Centralheating system None 791 N 6 Settopbox None 114 N 7 DVDplayer Nome 16 N 8 IV Nome 16 N Y SA 9 Router Noe 110 N Y S 10 Fanonatimer 1 117 N N FA 11 Hand vacuum cleaner charger None 23 Y N SA 12 Stereosystem 1 2 246 N Y SA 13 CDplayer 1 6 611 N Y SA 14 Wireless doorbellunit Noe 23 N N SA 15 Bread toaster Noe 2851 N Y SA 16 Film recorder charger Noe 49 Y N FA 17 Radio Nome 3 N Y SA 18 Washingmachine 2 11076 N Y FA 19 Clothes dryer None 5210 2324 N Y FA In addition to the above findings Table 7 shows two classifications an on off appliance and finally ON ON whether an appliance type is classified as Strict Appliance SA or Flexible Appliance FA An on off appliance classification is based on whether
59. ids a 2012 status update 10 Optimizing Energy Costs for Offices Connected to the Smart Grid Georgievski V Degeler G A Pagani T A Nguyen A Lazovik M Aiello s l IEEE Transactions on Smart Grid 2013 Vol 3 4 2273 2285 11 The Pecan Street Project Pecan Street Reasearch Institute Online Pecan Street Inc 2010 Cited October 3 2013 http www pecanstreet org 12 Smart Cap Flattening Peak Electricity Demend in Smart Homes S Barker A Mishra D Irwin P Shenoy J Albrecht Lugano IEEE International Conference on Pervasive Computing and Communication 2012 978 1 4673 0258 13 Priority Based Smart Household Power Control Model M Prymek A Hor k s l IEEE Electrical Power and Energy Conference 2012 978 1 4673 2080 56 14 Novel WSN Based Residential Energy Management Scheme in Smart Grid P Han J Wang Y Han Q Zhao s l IEEE International Conference on Information Schience and Technology 2012 978 1 4577 0345 4 15 EnergyHub Online EnergyHub Inc 2009 2013 http www energyhub com 16 P3 International Online P3 International Corporation 2013 http www p3international com 17 thinkeco Online ThinkEco 2013 http www thinkecoinc com 18 E inzicht Essent Online Essent 2013 http www essent nl mijn e e inzicht index2 html 19 Energy Intelligent Buildings based on User Activity A Survey T A Nguyen M Aiello Groningen Energy and Buildings Elsevi
60. ing home am referring to people who own their home and already own all the electrical appliances they need or are renting their home which does include all the electrical appliances needed These people might be in the position of not wanting to buy for example a new refrigerator or washing machine in order to decrease their electricity consumption but still wish for possible solutions to optimizing their existing home s electricity usage Additionally not all homes are equipped with smart meters today Is there a way to make existing appliances smarter by use of existing technology In other words can we provide home automation to an existing home without forcing the use of smart appliances and smart meters And finally when being faced with consumption higher than that of air conditioners and ventilation as shown in chapter 1 1 the following question arises how do we make residents more aware of their wastage of electricity in regards to standby power In order to reach the goal of optimizing electricity usage even out the electricity usage and only use the electricity needed in an existing residence the following research question must be answered Can optimization of electricity and automation of appliances provide a solution which aids a resident in reducing electricity consumption in an existing residence with existing appliances today In order to answer this question the following sub questions have to be answered e What is state of
61. it is an appliance which is turned off and on by way of a button or automatically turned off after one running session The type classifications SA and FA are selected in order to simulate the input of the end user In the case of the appliance not having a standby value None is recorded in the ASV column All numerical results are shown in kWh x 0 001 and finally the zero reading and on off appliance is denoted with Y yes and N no 6 2 6 Tariff information Requirements FR5 and FR7 state that the system shall display existing cost and proposed cost savings respectively In addition requirement FR4 states that the HEP shall provide analysis information based on single tariff and or split tariff In order to simulate the monetary values and savings for a single tariff real tariffs were obtained from my personal energy invoice from Essent 45 where a single tariff is 0 2241 kWh A quotation via email was obtained as well from Essent regarding the split tariff which is 0 2387 kWh during peak periods and 0 2125 kWh during off peak periods Peak periods are from 07 00 to 23 00 Monday through Friday and off peak periods are at all other times 37 6 3 Standby power detection evaluation Evaluation of the standby power detection was performed by means of measuring the validity of the three rules determined during the design of the HEP 1 The value has to occur at least twice in row and followed by a zero reading no power consumptio
62. iving a low outcome in specificity 0 23 and F measure 0 50 Another example is when we omit rule 3 which states that the ratio of the value highest reading per session cannot exceed 7 we can easily identify standby values on appliances incorrectly Table 10 shows that 9 appliances where identified incorrectly in test case 4a which results in a very low specificity outcome 0 08 40 6 4 Optimization evaluation An optimization evaluation was performed using the appliances which measured power consumption over a period of 5 months numbering is according to chapter 6 2 Experimental setup This prototype of the HEP allows for 10 appliances to be used in one analysis and therefore made a decision to remove two of the twelve appliances based on the classification SA or FA power consumption and usage As a result of these observations it was decided to leave out the oven appliance 3 and the DVD player appliance 7 which resulted in the reduced list of appliances in Table 12 Table 12 Appliances used in optimization evaluation ID Appliance ASV Type NEW 2 EJ 4 Dishwasher 2 FA 5 Central heating system None SA 6 8 9 10 Fanonatimr 1 FA Washing machine 6 4 1 Graphical results flattening of peaks The appliances were processed through the HEP using data collected over a period of 5 months 3 months and 1 month with the results shown in graphical form in Figures 9 11 on the following pages E
63. k can become an energy supplier as well that is consumers can become prosumers Any user of the smart grid with an energy source will be able to sell excess energy to the grid The energy usage and supply will be measured by smart meters 4 where the EU directive to its member states is to have 80 percent smart meters installed by 2020 9 A smart meter is an electrical meter that records consumption of electric energy and communicates that information back to a device for monitoring and billing purposes The smart grid promises to provide us with dynamic pricing in the future while today the trend in most countries is to provide consumers with a choice of several electricity providers as well as different electricity tariffs depending on the time of use 10 Today most new appliances produced have the potential to use up to 4096 less electricity 2 which is not enough to put them into the category of smart appliances One aspect of smart appliances is that they are produced to be able to measure their own consumption and make use of that information to save electricity and money as well as being programmable Smart appliances are in their infancy but are starting to emerge on the public market On the other hand there are several home automation solutions available which make use of wireless technologies as well as existing home wiring to connect and automate appliances There are several research projects devoted to the smart grid new emerging ele
64. l usage of individual appliances 17 In order to get an overview of what the HEP s functions are a set of functional requirements where defined Table 2 HEP Functional Requirements ID Functional Requirement Home Energy Planner HEP The HEP shall find standby values of end user defined appliances when applicable The HEP shall provide a schedule with proposed running times of end user defined appliances within a 24 hour period The HEP shall provide the existing usage of appliances chosen The HEP shall provide analysis information based on single tariff and or split tariff The HEP shall be easy to use in terms of supplying the end user with information on input EDS FR 11 The end user shall have control over which appliances are used in scheduling NEN FR3 FRA i FR5 The HEP shall provide existing cost of appliances chosen o lt FRE FR8 NEM FR 12 The end user shall have control over which time period is used in analysis FR 13 The end user shall have control over tariff cost used in analysis 4 2 Data collection A home lab was set up in order to collect the data which was used in the design of HEP Plugwise home automation system was used to collect the electricity consumption from individual appliances This system is composed of a Circle seven Circles a Stick and the Source application as described in chapter 3 4 The Source was installed on a laptop running Windows 7 o
65. layer determined that it was more accurate for the analysis calculations as well as usefulness of the HEP for the end user to determine standby power on fewer appliances than the actual count rather than giving false conclusions on appliances that should not be turned off examined the correctness of these design choices by means of recording the appliances found with and without standby values according to the following test cases 1 How well does the standby detection perform with the algorithm complete How well does the standby detection perform without rule 1 The value has to occur at least twice in row and followed by a zero reading no power consumption in at least one session 3 How well does the standby detection perform without rule 2 If the appliance is never turned off never has a zero reading the value has to occur at least two times in a row 4 a How well does the standby detection perform without rule 3 The ratio of the value highest reading per session cannot exceed 796 b How well does the standby detection perform if the ratio is 896 5 How well does the standby detection perform when none of the rules from chapter 4 5 test cases 2 4b are applied in other words by means of finding the highest occurrence of the lowest value found for each appliance in one day excluding a zero reading when other values are found in that particular day Each test case was recorded by means of the standby value detected for
66. load The washing machine manual states that the delicate program runs for 50 minutes and consumes 0 3 kWh a 40 C program runs for 110 minutes and consumes 0 55 kWh and a 90 C program runs for 110 minutes and consumes 1 90 kWh which was consistent with my documented timeframes The manufacturer s data is based on each load of wash being a fully loaded machine except only half full for the delicate wash there will always be a variation on the weight of the load and therefore assume that each load s power consumption could vary 0 2 kWh The randomly generated total consumption of each load was divided randomly into the documented timeslots Finally a standby consumption of 0 002 kWh is assumed since the washing machine does not power off when the program has completed but 32 there is a light blinking which indicates that the program has finished The standby consumption value is the same as the standby consumption value measured from the dishwasher which also does not power off when completed but has a blinking light indicating that the program is finished The clothes dryer always ran for 80 minutes for each load and once completed it does not consume any standby power In the manual it is stated that a 5kg load of clothes which comes from a washing machine using a 1200 spin cycle consumes 2 7 kWh Even though each load of wash was for a fully loaded machine there will always be some variation in weight Therefore it was assumed that ea
67. n in at least one session 2 If the appliance is never turned off never has a zero reading the value has to occur at least three times in a row 3 The ratio of the value highest reading per session cannot exceed 7 The test cases numbered 1 5 in chapter 6 1 2 were performed by means of removing the appropriate functions in the implementation of the HEP In order to detect standby power all 19 appliances listed in chapter 6 2 3 were processed through the standby power detection in HEP One week of usage for each appliance was used where the week of usage was chosen based on the appliance having at least one session were the standby power was measured or recorded for at least two time slots This is done due to the design decision that the standby power algorithm only detects standby power on appliances which have this criterion as described in chapter 5 6 Table 9 Results of experimental measurements of 19 appliances shown in kWh x 0 001 Appliance ASV 1 2 3 4a Refrigerator None None 5 None 69 _ Coffeemaker 4 4 2 4 4 No Dishwasher 2 2 Nome 2 2 2 7 Set topbox None None None None 8 NN ft 2 7 5 10 1 10 NN 06 6 8 8 Ma o2 EE AN Washing machine 9 The results in Table 9 were then grouped together as appliances with standby power consumption which Ja 0 pa 00 ES 2 3 a 5 6 E 8 9 19 were correctly identified true positive appliance
68. n Figure 7 showing the main classes variables and methods In addition Table 5 describes the rationale of the main classes Figure 7 HEP classes showing main variables and methods Table 5 Rationale of the main classes Class Rationale LayoutCombined The main class of the application and the GUI see Appendix I AppliancesToUse Adds up existing power consumption usage of all chosen appliances Adds up proposed power consumption usage of all chosen appliances Provides all usage and cost calculations e Whole existing power consumption usage kWh Existing average power consumption usage per day kWh day Proposed average power consumption usage per day kWh day Proposed kWh savings a year KWh year Whole cost of existing usage based on single tariff Cost of existing usage per day based on single tariff day Whole cost of proposed usage based on split tariff Cost of proposed usage per day based on split tariff day e Estimated savings per year year Calculates proposed scheduling times of FA Reads the power consumption data from the CSV files chosen by the end user 25 MEE Verifies that the uploaded files are of correct format MyTableModel Model for the tables showing power consumption data given in the CSV files using the dates chosen by the end user StandbyCalculations Finds standby values associated with each appliance where applicable ArrangeAppliances Finds one session of an ap
69. n to the external power measuring device The main differences in HEP and these three existing solutions are that HEP has standby power detection you can see the difference in usage and saving of single and split tariffs and you can see savings in kWh and Euros if you exclude an appliance such as if you stop using clothes dryer A summary of the existing solutions and HEP can be seen in Table 22 below Table 22 Summary of existing solutions and HEP Solution Solution Il Solution Ill Existing appliances OX P X X Smart appliances Sensors 0000000000000 End user classification of appliancetype x XxX Automaticcontro O X Show potential savings euro and or kwh X X X Show potential difference in pricing plans pp X Standby powerdetection removal YOOX Reduction of electricity consumption peaks X X X X In order to answer the research question Can optimization of electricity and automation of appliances provide a solution which aids a resident in reducing electricity consumption in an existing residence with existing appliances today a set of sub questions were researched in order to gain a better understanding and provide a solution e What is the state of the art in Smart appliances Smart appliances are just emerging on the public market as of 2012 2013 So far they are limited in types and functions where some require a connection with a smart meter in order to utilize the en
70. ned by the following process 1 All appliance readings are stripped of their standby values according to standby rules in previous section 2 SAare added together as strict usage FA are analyzed and each appliance is given the following a Hours needed to run one session highest occurrence b Electricity consumed in one session 4 FAare given slot priority values based on average electricity consumption in one session with the highest being assigned slot priority one and ascending 5 Total hours needed to run all FA one session each 6 Find concurrent timeslots for the total times needed to run one session of each FA a Single tariff finds lowest total strict usage b Split tariff finds lowest total strict usage including off peak time period 7 Assign timeslots to flexible appliances based on slot priority values 4 7 Usage and savings The HEP provides the end user with estimated average usage in kWh s per day estimated savings in kWh s and Euros per year These calculations are based on the results from the previous sections as well as the tariffs provided by the end user 21 5 Implementation This chapter describes the implementation of HEP by first describing the hardware and development tools used Then continue by explaining the main components and the data flow between them the main classes and their rational and finally the most important runtime views of the system 5 1 Development tools The HEP was developed on an As
71. nnnnesennnnesennnusseeenne 21 O PANNEN T AA 21 SN 22 GA Jol E Ane o EEE m 22 54 SEN COMPONENTS RR EEE cs 22 PA VEW EN Emm 25 A E 26 ENN 26 5 3 2 Save tariff information acaba 26 5 SV TN UP 28 5 34 VIEW proposed us 29 PENN 30 Oe Objective and metrics EN 30 6 12 Standb DOW Clr detec NN 30 dl OMAN A A ts 31 PA A LE ULLA E ME ELE 31 HM xeasiedesmedusdbicmerusihee dolis ueste testamur alice mad A 32 6 2 2 Timed appliances ecd er E cerdo EYE a Red 32 6 2 3 Data collection during EXPELIMENL cccccescccccsseccccessececeesececeeeecceseusecesseeecesseeeceesegaesetsenecetseges 33 TDN 34 6 2 5 Observed appliance values standby values and classifications eese 34 626 ET NONNE 37 6 3 Standby power detection evaluation saadan 38 Gu Optimo SANN NS 41 6 4 1 Graphical results flattening of peaks sees 41 6 4 2 Calculation results existing and proposed usage and SaviNgS ccccooccnnccnocnnnnnnncnnonanonnnnanonononos 46 6 5 RedquiremehtVerifiCadt ION ais GS 48 Ko DISCUSSION mee n IDE E 49 SLI Standby d tectloD s oe e ru etae pub dad 49 6 1 2 Flattening of peaks and reduction of peak periodS ocoocccnccnocnnnonnnonnnnnaconnnnnccnnonanonnnnanonnnonos 49 6 1 2 1 Reasoning for high magnitude peaks ota 50 6 13 CalCUlatiOn 1 CSUIS ccc svclesssics tends n O onderenass 51 6 14 Requirements VENCIO ii id 52 Conclusi n and T Ene VO 53 A
72. of 0 95 can conclude that it provides over 9096 validity 6 1 2 Flattening of peaks and reduction of peak periods In order to examine the flattening of the peaks constructed the graphs which resulted in HEP from the evaluation performed on 1 3 and 5 moths of data In all three time periods constructed one graph for existing usage one for proposed usage with a single tariff and one for proposed usage with a split tariff Then proceeded to calculate the average absolute deviation from the mean power on the resulting readings The magnitude of the deviation quantifies the variation in power consumption and allows me to quantify the flattening of peak periods also examined the reduction of peak periods which is defined as the periods where the consumption data goes above the mean e Existing versus proposed split tariff When comparing the existing usage with proposed split tariff usage the split tariff usage yields a greater reduction of peaks in all three time periods by 24 9696 1896 and 17 2696 1 3 and 5 months respectively differences in average absolute deviation from the mean power These results are consistent with the reduction of peak periods of the graphs in Figures 9 11 In the 3 and 5 months data analysis we can see that the existing graphs show the time period 9 00 23 00 as a 14 hour peak compared to three 2 hour peaks in the graphs showing the proposed split tariff This yields a reduction from 14 hours in peak periods to 6 hour
73. of proposed usage per day based on split tariff day Estimated savings per year year It sends the schedule for all stripped FA appliances along with total stripped SA appliances times and usage information Appl X Schedule to the Result GUI component In addition it sends all the results from the calculations of usage Usage result the calculations of savings Savings results and total existing usage Appliance Total kWh h to the same component A Java Swing component where the results from the analysis and existing information is displayed From the Usage Saving component it receives the schedule for all stripped FA appliances along with total stripped SA appliances times and usage information Appl X Schedule all the results from the calculations of usage Usage result the calculations of savings Savings results and total existing usage Appliance Total kWh h In addition it receives the standby values of all appliances Appl X Standby from the Standby Detection component A lexical representation of existing usage and cost proposed usage cost and savings in kWh and is provided based on chosen single or double tariff In addition the times to schedule Flexible Appliances along with standby killer values for all appliances are shown lexically JFreeChart API is used to display existing and proposed usage information graphically 24 5 2 Logical view The logical view of HEP is depicted in graphical format i
74. off The great increase of gadgets within a household has a great impact of the amount of electricity we are consuming They may not consume great amounts of electricity on their own but collectively when many are added up the consumption increases We need to make consumers more aware of their own electricity usage and supply them with options to manage usage and reduce waste 1 1 Current situation id Industry E Transport ii Residential E Agriculture Forestry id Services E Other Figure 1 Final electricity consumption by sectors in the EU 27 6 Figure 1 shows the electricity consumption in the EU 27 in 2010 where the residential sector is responsible for the second highest consumption This shows that there is a need to put a greater emphasis on efficiency in the use of electrical household appliances and equipment In the Energy Efficiency Report 2012 6 it is stated that final electricity usage decreased in the year 2009 but the final residential electricity consumption per capita did not decrease between the years 2008 and 2009 It increased by 1 5896 which shows that the overall decrease in electricity consumption per capita 4 9696 was not achieved from within the residential sector However on the other hand the growth rate reported in the EU 27 is the second lowest in 2010 or 0 4296 over the period 2000 2010 The lowest growth rate was reported in 2007 or 0 8096 which they state can be related to higher temperat
75. onsistently turned on when the TV Circle 9 was turned on in the morning and turned off when the TV was turned off for the last time in the evening Therefore the reading of the set top box is calculated by using the timing measurements from the TV The Hand vacuum cleaner Circle 2 and the Stereo system Circle 11 where extremely sparingly used therefore the readings collected from them are only used when determining the algorithms used in the analysis on standby power consumption 6 2 2 Timed appliances Washing machines and clothes dryers where responsible for 7 296 of the average household power consumption in 2009 6 These types of appliances are high in consumption and are appliances which do not have to run on command but the end user can schedule the runtime within a time period Therefore these two appliances were important to include in the design and analysis of the HEP The washing machine and clothes dryer in the residential setting used for collection of data where not connected to power through conventional sockets and therefore the Circles were not an option In order to simulate the power readings of these two appliances we documented the times and dates the two appliances were powered on and off as well as documenting the program that the washing machine was using delicate 40 C and 90 C We then used the technical information contained in the users manuals dryer 42 washer 43 to determine the total average usage per
76. ormat is wrong ex 1 5601 is displayed 3 End user clicks the OK button on the alert box 4 HEP displays the previous view with the empty Offpeak tariff field as well as other previously correctly inserted fields 1b The format of the input in the Offpeak tariff field is wrong 2 An alert box with the message The Offpeak tariff format is wrong ex 1 5601 is displayed 3 End user clicks the OK button on the alert box 4 HEP displays the previous view with the wrong input in the Offpeak tariff field as well as other previously correctly inserted fields 1a Nothing has been typed into the Peak tariff field 2 An alert box with the message The Peak tariff format is wrong ex 1 5601 is displayed 3 End user clicks the OK button on the alert box 4 HEP displays the previous view with the empty Peak tariff field as well as other previously correctly inserted fields 1b The format of the input in the Peak tariff field is wrong 2 An alert box with the message The Peak tariff format is wrong ex 1 5601 is displayed 3 End user clicks the OK button on the alert box 4 HEP displays the previous view with the wrong input in the Peak tariff field as well as other previously correctly inserted fields 1a Nothing has been selected in the days checkboxes 2 An alert box with the message You have to select at least one day for a split tariff is displayed 3 End user clicks the OK button on the alert box 27 4 HEP displays t
77. perating system A total of 8 appliances were measured in order to perform the required analysis needed for the design Table 3 List of appliances measured Appliance 18 4 3 Appliance classification In using the HEP the end user has complete control It is up to them to classify the appliances used in the application in accordance with their own expectations and needs There are only two classifications which the end user needs to assign to each appliance 1 Strict appliance SA These are appliances which the end user does not want to be given schedules because they need to run all the time or be on demand from the end user These are typically appliances such as refrigerators heating systems entertainment devices etc 2 Flexible appliance FA A flexible appliance can be assigned a schedule and be run at different times within 24 hours These could be appliances such as washing machines dishwashers or phone chargers which are appliances which need to be run but this can be completed within a timeframe as opposed to an exact time such as phone needs to be charged before 7 00 am 4 4 System overview After careful examination of the usage data it became apparent that in order to optimize the electricity it would be most convenient to move one session of an FA to a period where electricity consumption was low In conclusion the following steps were identified which need to be taken in HEP Determine standby power consumption for
78. pliance running total hours and power usage 5 3 Runtime Views describe four runtime views which are identified in the form of use cases There are other views involving the HEP but these are the most important and most used views The Success Scenario row depicts the flow when the use case is correct while the Alternative Flows row depicts the use case based on several possible inputs which lead to a different flow and show the alternative flows along with error messages when applicable 5 3 1 Add appliance The end user is in control of which appliances to use in the analysis and therefore has to select the appliance consumption files In this use case the end user is selecting one appliance to add to the selection which can be done at anytime during a session in HEP NET Add an appliance Success Scenario 1 End user clicks the Add appliance button 2 HEP displays a file upload pop up window 3 End user selects a file to upload and clicks Open button 4 HEP displays the newly selected appliance in the list of appliances Alternative Flows 3a File selected has already been uploaded 4 HEP displays the previously selected appliances in the list of appliances 3b File selected is in wrong format 4 An alert box with the message The file you chose is not the right format Choose another file is displayed 5 End user clicks the OK button on the alert box 6 HEP displays the previously selected appliances in
79. proposed with split tariff over a period of 24 hours x is the value of the consumption dataset at hour and mean T is the mean value of the dataset Table 13 Average absolute deviation from the mean power using 5 months of data Average absolute deviation from the mean power in kWh x0 001 5 months 115 53 Proposed Single 124 23 Proposed Split 97 17 Table 14 Average absolute deviation from the mean power using 3 months of data Average absolute deviation from the mean power in kWh x0 001 3 months Existing 115 75 Proposed Single 115 76 Proposed Split 96 64 Table 15 Average absolute deviation from the mean power using 1 month of data Average absolute deviation from the mean power in kWh x0 001 1 month 122 04 Proposed Single 129 75 Proposed Split 94 87 Tables 13 15 display the average absolute deviation from the mean power for 5 3 and 1 month of data respectively Each table shows the results for existing usage proposed usage with a single tariff and proposed usage with a split tariff using the given dataset The lower the outcome of the calculation is the greater the flattening over the given interval 45 All three time periods resulted in the same schedules of the Flexible Appliances which were as follows Table 16 Scheduling times resulting from 1 3 and 5 months of data Single tariff Split tariff Clothes dryer 20 00 23 00 Washing machine 22 00 Dishwash
80. s The 1 month graph shows a slightly different outcome where the existing graph has two peaks one 4 hour peak and one 7 hour peak and the proposed split tariff has two 2 hour peaks and one 1 hour peak This yields a reduction from 11 hours to 5 hours in peak periods e Existing versus proposed single tariff When comparing the existing usage with proposed single tariff usage the single tariff usage yields a greater increase in average absolute deviation from the mean power in all three time periods by 6 1296 0 0000996 and 7 2696 1 3 and 5 months respectively When I examine the reduction of peak periods in the graphs in Figures 9 11 I see that the existing 3 and 5 months data analysis has a continuous 14 hour peak period compared to two peaks in the proposed single tariff 5 month analysis one 2 hour and one 4 hour peak and 49 three peaks in the proposed single tariff 3 month analysis one 1 hour and two 2 hour peaks The 1 month analysis shows two peaks in existing usage one 4 hour and one 7 hour and two peaks in the proposed single tariff 1 month analysis one 2 hour and one 4 hour peak Even though this means a reduction of peak periods by 5 9 and 8 hours 1 3 and 5 months respectively by using the proposed single tariff the increase in the average absolute deviation is caused by the magnitude of the two 4 hour peaks and one 2 hour peak in the three different time periods of the proposed single tariff All three time periods for both t
81. s being uploaded by end user see Use case 53 1 appliances are used in scheduling appliance type as FA see chapter 4 3 period is used in analysis change start date of appliance data used in analysis user to insert the tariffs see Use case 5 3 2 48 6 6 Discussion The following section provides a discussion with interpretations of the evaluation results provided in previous sections of this chapter 6 1 1 Standby detection calculated precision recall F measure specificity and accuracy in order to evaluate the quality of the algorithm designed for standby detection Six test cases were identified where test case 1 represents the application of the algorithm in its entirety in test case 5 the algorithm is not applied at all but the standby value is based on the lowest reading found The four test cases 2 3 4a and 4b represent different rules not being applied in running of the algorithm Test case 1 scored the highest average value as well as having the highest value in F measure 0 91 and accuracy 0 95 where the highest value is 1 0 and lowest O as seen in Table 6 The F measure is a weighted average of precision and recall and accuracy is the proportion of true positive tp and true negative tn measurements These results provide a conclusion that test case 1 or the designed algorithm works very well and should not be implemented without the use of the rules defined in chapter 4 5 With an F measure of 0 91 and accuracy
82. s and methods in relevant SOlJUTIONS oocccccnccncnocnnonacinonacinonananonacnnonccnonanoss 12 Table 2 EP 18 Table 3 List of appliances measured orcas erro EE 18 Table 4 HEP components and data flow description cccooocccconoccnnonanonnnonacnnnonaronnnnnnonnnnnncnnnnnnrnnnonanonnos 23 Table 5 Rationale of the main classes carnada 25 Table 6 Data representation in table format as shown in SOUFCB ooccconccccnnccnnnacnnonacinonaninonacononacnonarononanono 34 Table ENN 34 Table 8 Appliance readings and classifications based on outcomes of data measurements 37 Table 9 Results of experimental measurements of 19 appliances shown in kWh x 0 001 38 Table 10 Summarized results of standby value valuation cccccccesseccccsseccceesscceeeesececsueceeeeeceeseneeeetas 39 Table 11 Precision recall F measure specificity and accuracy of standby value test cases 40 Table 12 Appliances used in optimization evaluatiON ooccccnncnnnnocnnonannnnncnnnnacononanononananonacanonccnnnanononanoso 41 Table 13 Average absolute deviation from the mean power using 5 months of data 45 Table 14 Average absolute deviation from the mean power using 3 months of data 45 Table 15 Average absolute deviation from the mean power using 1 month of data 45 Table 16 Scheduling times resulting
83. s on or off The communication protocol used by Plugwise is ZigBee 40 ZigBee is optimized to consume as little energy as possible which limits the broadcast range of a single module Plugwise recommends 5 10 meters between modules in the Plugwise network but since each module can pass information to other modules one is able to cover large areas 41 16 4 Design and methodology In the research and design of the HEP a simple workflow was applied as shown in Figure 5 Initial design Data collection Refinement of design Figure 5 Iterative workflow The initial design included defining the system context and requirements Data collection analysis and refinement of the design were performed in several iterations where the design focused on optimizing the electricity consumption in a real home setting 4 1 System context and requirements HEP was designed to provide the end user with a suggestion on how to optimize energy usage on existing appliances using a home automation system When using a home automation system such as Plugwise the end user is provided with the consumption of individual appliances and is able to create schedules which will turn appliances on and off as well as standby power killers HEP provides the end user with information which will reduce the work of the end user and could add to the incentive of the end user by doing more than measuring the energy usage and possibly reducing the actua
84. s which were not detected false negative appliances 38 which were incorrectly detected false positive and finally correct absence of standby values detected true negative Results are shown in Table 10 below Table 10 Summarized results of standby value evaluation Test cases 1 2 3 5 t9 Appliances correctly identified with correct standby value 5 2 3 6 5 3 13 Appliances with standby value missing 1 3 3 0 1 1 ie Appliances incorrectly identified 0 3 0 9 1 9 ive Correct absence of standby value 13 11 13 4 12 6 The metrics used for evaluation are precision recall F measure specificity and accuracy This was done by means of classification using the following five formulas 46 tp Precision tp fp Precision is defined as the number of items correctly identified as belonging to the positive class divided by the total number of items incorrectly labeled as belonging to the class tp Recall dido tp fn The recall is defined as the number of items correctly identified as belonging to the positive class divided by the total number of items which were not labeled but should have been labeled as belonging to the class Precission Recall Fy 2 Precission Recall The F measure Fi is known as the measure which combines precision and recall into a harmonic mean where precision and recall are evenly weighted tn Specificity tn fp Specificity gives the relation betw
85. t s just the 2 of us in this 2 bed flat am out all day and we are on income support just don t know how the bills are so high think there is something wrong with them a female 30s London commenting whilst in broad daylight lights were on in most rooms and a TV and radio were playing in an unoccupied bedroom and all appliances in the sitting room TV DVD stereo computer were on standby 20 This comment recorded in a qualitative study performed on several households in the UK shows that the end user needs to be made aware of how they are consuming electricity in their homes Several studies have been conducted were end user behavior is the focus Gill et al 21 conducted a custom made behavioral survey on inhabitants of a UK EcoHomes site They state that The significance of information feedback and simple explanation of electrical consumption in terms understood by the users cost being the only one referred to by any occupants is crucial to reducing electrical energy waste in dwellings In most homes today the electricity meters are tucked away out of sight and only give the end user a total consumption usage of all electrical appliances in the home Goncalves Da Silva et al 22 conducted a survey in 2011 which was open to the public for a period of 2 months on the NOBEL project website Even though this survey was intended to better understand residential prosumers some of the conclusions are relevant for all end
86. the art in Smart appliances e What types of electricity measurement solutions are available for existing appliances today e Can we create effective standby power detection on the appliances measured e Can we reduce peak periods and flatten peaks using the optimization application e Can we provide a mean to see if different types of electricity tariffs are resulting in savings in cost 1 3 Thesis contribution and organization Many parties are finding ways to enable residents to optimize electricity usage in existing homes 12 13 14 as well as many home automation solutions being available 15 16 17 18 In order to answer the research question and sub questions in the previous chapter do research regarding what solutions are available both which have been implemented such as smart appliances and which are suggested in published works In order to optimize the electricity use in an existing residence with existing appliances use a home automation system which measures the electricity consumption of individual appliances In addition the home automation system allows for automation of the appliances in the form of the end user being able to determine timing schedules to turn appliances on and off For the purpose of optimization an application which aids the end user in creating these schedules was needed which lead to me designing and implementing the Home Energy Planner HEP The HEP includes standby power detection and a sche
87. the list of appliances 5 3 2 Save tariff information The end user inserts the tariff information into HEP There is an option for using a single tariff and or a split tariff This can be done at any time after an appliance has been uploaded The end user is able to see the difference in cost and savings based on type of tariff as long as the chosen tariff type is saved before further information is chosen such as 5 3 3 View existing usage if estimated cost information is wanted and 5 3 4 View proposed usage 26 NETAS Save tariff information End user clicks the Save tariff button Success Scenario 1 End user clicks the Save tariff button 2 HEP displays the tariff information inside the input fields Alternative Flows Single tariff 1a Nothing has been typed into the Single tariff field 2 An alert box with the message The Single tariff format is wrong ex 1 5601 is displayed 3 End user clicks the OK button on the alert box 4 HEP displays the previous view with the empty Single tariff field 1b The format of the input in the Single tariff field is wrong 2 An alert box with the message The Single tariff format is wrong ex 1 5601 is displayed 3 End user clicks the OK button on the alert box 4 HEP displays the previous view with the wrong input in the Single tariff field Split tariff 1a Nothing has been typed into the Offpeak tariff field 2 An alert box with the message The Offpeak tariff f
88. to a maximum of 100 euro During the first period the campaign resulted in 30 sales in Sep Oct 2009 and 34 sales in Nov Dec 2009 which was a considerable increase from the 1596 in the months before Jul Aug 2009 The second period did not see quite as large an increase in the months between the two periods the sales where between 21 and 2596 whereas they rose to 38 40 and 39 Sep Oct and Nov 2010 respectively during the campaign which was concluded to be due to early replacement buyers already being triggered by the first campaign 32 3 3 New Smart Appliances Smart appliances have been introduced by many manufacturers for some time now 33 34 35 but it was not until 2012 13 that they became available to the public Whirlpool and LG are two of the companies which have produced lines of smart appliances that are only available in the US at this time Below is an example of smart appliances which these companies have produced but the information gathered only refers to electricity savings Whirlpool s Smart Appliances with 6 Sense Live Technology 5 came out in spring of 2013 and offers washing machines dryers refrigerators and dishwashers equipped with a technology which gives the 14 end user the ability to manage energy usage and control features remotely via Wi Fi using a Smart phone tablet or computer These appliances are equipped with Whirlpool Rate Revealer a software that checks for TOU prices offered by the ele
89. umption data per hour SA Total kWh h and sends it to the Scheduler component Flexible Appliances A Java component which receives each Flexible Appliance s stripped consumption data FA X from the Standby Detection component For each Flexible Appliance it finds the highest total kWh needed to run each appliance for one session FA X kWh as well as finding the longest time in hours to run each appliance for one session FA X h which is then sent to the Slot Priority component Slot Priority A Java component which assigns slot priority to all Flexible Appliances based on highest kWh needed FA X kWh and the largest amount of hours needed FA X h to run one session of an appliance in ascending order which data was received from the Flexible Appliance component Each Flexible Appliance slot priority number FA X Slot 8 is then sent to the Scheduler component along with FA X kWh and FA X h A Java component which creates the schedule of all Flexible Appliances It receives the total kWh h consumption for all stripped strict appliances from the 23 Usage Savings Strict Appliance component and tariff information Tariff from the Appliance GUI component In addition it receives the highest kWh needed FA X kWh and the largest amount of hours needed FA X h to run one session of each Flexible Appliance along with each Flexible Appliance slot priority number FA X Slot from the Slot Priority component It adds up the total amount
90. ures during that year which implies less energy use for heating Residential electricity consumption Q on 0 c Q po Pus Q A Heatin l 8 Electric systems Water Washer Dish Set top Lighti amp electric i heaters ee Oth Dryer washers boxes E boilers Figure 2 Breakdown of the EU 27 residential electricity consumption 2007 and 2009 7 6 Figure 2 shows a comparison of breakdown of residential electricity consumption in the EU 27 between the years 2007 and 2009 Heating systems cold appliances lighting and water heaters are responsible for almost half of the total consumption The category other includes home appliance standby consumption entertainment devices office equipment vacuum cleaners coffee machines etc In 2007 the home appliance standby consumption was 5 996 or 47 5 TWh yr which means that it was consuming more than air conditioners and ventilation This shows that there is a great need to give more focus to standby consumption Even though appliances and devices are becoming more efficient due to energy efficiency measures taken the general trend is that the residential sector consumption is increasing 1 1 State of the art The smart grid 8 can be seen as the successor of the conventional electrical infrastructure which uses information and communications technology to automate the production and distribution of electricity One innovation of the smart grid is that consumers of the networ
91. us K53S laptop running a Windows7 Premium operating system The programming language chosen was Java and the developer tool was Eclipse Java EE IDE Indigo Service Release 2 GUI components were implemented in Java Swing which is the primary Java GUI widget toolkit and the JFreeChart 1 0 14 API was used for the creation of the graphs 5 2 System components Apphance X assification X Date sumption X Consumption X Classification X Tariff FA X kWh FAXh Appliance Total kWh h SA Total kWh h ATEN FA X kWh FAXh rcr Appl X Standby SA Total kWh h Tariff Appliance Total kWh h Appl X Schedule Usage result Savings result Figure 6 System components and data flow X id number value 22 Figure 6 depicts the HEP components and the main data flow between them In Table 4 a detailed description of each component including the data flow between them is specified Table 4 HEP components and data flow description Component Description Appliance GUI A Java Swing component where the end user chooses the appliances classifies them as strict appliance SA or flexible appliance FA chooses the dates to use in the analysis and inserts single tariff or split tariff JFileChooser a Java swing component is used for the uploading of the consumption files Appliance id Appliance X its classification Classification X and dates Date to use in the analysis is sent to the Consumption component The tarif
92. ve in July 2011 The new directive introduced three new energy efficiency classes A A A in addition to the previous classes A G as well as providing the information with pictograms rather than words The labels as shown in Figure 3 a give the energy efficiency based on these classes which are determined by a given appliance s EEI given in kWh per year In addition the labels give information about noise consumption capacity standby power etc based on the type of appliance This type of labeling is created to give the end user performance i Example given for washing machine For the calculation of the Energy Efficiency Index EEI of a household washing machine model the weighted annual energy consumption of a household washing machine for the standard 60 C cotton program at full and partial load and for the standard 40 C cotton program at partial load is compared to its standard annual energy consumption 48 13 information in order to be able to compare with other products with the objective of encouraging the purchase of more energy efficient models 27 Other well known types of labeling are the European Conformity mark 28 or CE marking 29 and the EU ENERGY STAR label 30 shown in Figure 3 b and 3 c These types of labels are meant to give the end user the knowledge that a product complies with minimum EU performance standards before being placed on the market The EU Energy Star label was implemente
93. y additional equipment The evaluations proved that the application works very well on standby power detection and reduction of peak periods In addition it provides suggestions to the end user on how to schedule their appliances and assign standby killers which reduce electricity consumption Finally the end user is in control of which appliances to use in the analysis which provides the end user with a way of seeing how the usage would change if an individual appliance was not included HEP will benefit from further work involving more extensive analysis in order to produce more efficient schedules for appliances One approach could be to use a database with consumption information from appliance producers The end user would then be able to identify the appliance by make and model and that would allow for better detection regarding the standby values as well as more accurate estimates of average running times of appliances In the case of a particular appliance not being available in the database the end user could classify it by type and purchase year Then the database would be searched for appliances with similar properties and estimate the consumption and running times 54 Another approach could be to re define the timeslots into smaller intervals The HEP timeslots are one hour each which can in some cases be very inefficient Looking at the example given in chapter 7 where the appliance needs 70 minutes to run one session but is started at

- Department of computing science

Contents

Download Pdf Manuals

Related Search

Related Contents