Home

Chapter 8, Thermal Conductivity

image

Contents

1. results short list lt Rootname gt LST one rootname batch Short description Description Data file designation Filename Root name serial sample ID TK04 Results rootname_serial TC Calculated thermal conductivity TK04 Results calculated_tc LET STD LET SAM of std dev CON TK04 Results let_or_sd Solutions No of solutions found TK04 Results solutions Start time Start of approx time interval s TK04 Results time_start Time Length of approx time interval s TK04 Results time_length End time End of optimal time interval s TK04 Results time_end Eval Evaluation method SAM or CON TK04 Results eval_method Hints Comments from DWL file TK04 Results hints Table 8 7 TK04 appended results short list lt Rootname gt LST all rootnames Short description Description Data file designation Filename Root name serial sample id TK04 Results rootname_serial TC Calculated thermal conductivity TK04 Results calculated_tc LET STD LET SAM of std dev CON TK04 Results let_or_sd Solutions Number of solutions found TK04 Results solutions Start time Start of approximate time interval s TK04 Results time_start Time Length of approx time interval s TK04 Results time_length End time End of optimal time interval s TK04 Results time_end Eval Evaluation method SAM or CON TK04 Results eval_method Hints Comment
2. ASTM 1993 kat q 47 In t a l 12 T1 4 k t is the apparent thermal conductivity because the true conductivity k is approached only by a sufficiently large heating duration This method assumes that the measurement curve is linear and ignores the imperfections of the experiment expressed in the constant C In practice the correct choice of a time interval is difficult During the early stage of heating the source temperature is affected by the contact resistance between the source and the surrounding material During the later stage of heating boundary effects of the finite length of the source affect the measurement The position of the optimum interval generally differs from measurement to measurement The two systems presently available on the ship employ different procedures to select the time interval the older Thermcon 85 system relies on operator judgment based on visual examination of the In f vs T plot the newer TK04 system uses an PP Handbook Peter Blum November 1997 8 1 ENVIRONMENTAL algorithm that automatically finds the optimal time interval Erbas 1985 More information is provided about each in the following sections EFFECTS In situ thermal conductivity is a function of in situ temperature and pressure conditions Corrections may be applied to laboratory measurements on cores based on in situ information and theoretical and empirical relationships Data in the ODP database are not correc
3. are constants The error in this procedure is far smaller than the general uncertainty in thermal conductivity measurements The constants are available to the data reduction program and are used for conversion of measured resistance into temperature Electronics Technicians are responsible for entering the constants of a new resistor into the program Do not attempt to re calibrate the thermistors a specialized facility is required ODP procedure with the Thermcon 85 system includes a calibration of each needle probe using standard materials of known thermal conductivity values Table 8 1 These values were established on Legs 127 129 and 131 and on subsequent legs using this same instrument This calibration should be viewed as a relative one that makes ODP shipboard data a little more consistent Table 8 I_ Standard materials used for calibrations and control measurements Standard material Thermal conductivity W m k Black rubber 0 54 Red rubber 0 96 Macor 1 61 The standard measurements must be entered into a separate spreadsheet and the liner coefficients slope intercept determined The coefficients are then entered into the PROBES DAT file using the PROBES program utility The thermal conductivity values returned by the PC TC program are subsequently corrected using these coefficients PP Handbook Peter Blum November 1997 8 3 PERFORMANCE Precision About 5 Systematic evaluation
4. half space needle probe HLQ it is expected that the total amount of produced heat penetrates into the sample The thermal conductivity is thus calculated with twice the heating power used for the full space solution This assumption is justified if the thermal conductivity of the samples is not lower than about W m K at lower values an error arises because some of the produced heat is penetrating the probe half space in which case it is necessary to determine correction factors to compensate for the heat loss Extended evaluation using special parameters adapted to circumstances yields an uncertainty of less than 2 This is clearly smaller than variations caused by sample preparation and inhomogeneities in rocks and sediments and special evaluations are appropriate only for standard materials and fundamental material investigations Random variations of thermal conductivity in natural materials such as sediments and rocks typically give an uncertainty of about 5 Routine evaluation using the TKEVA EXE has an accuracy of about 5 and is therefore appropriate 1 Bring cores to temperature equilibrium about 4 hr Hard rock specimens should be placed in a water bath to equilibrate 2 Soft sediment drill holes into core liner Also drill a small hole in semiconsolidated sediment if necessary Apply thermal joint compound if necessary Insert full space probes carefully into sediment Hard rocks prepare smooth surface on a half core s
5. is recommended to process the data immediately Load the PROCESS program from the TCMENU screen The run just completed will appear as the default run to be processed Accept or change it Select the position to be processed and the drift correction The In vs T graph will be displayed Select the time interval to be processed by moving the cross hairs on the screen For routine processing use the same interval used for secondary probe calibration Adjust if necessary Press enter to calculate conductivity and the fit parameter Warnings will come up if the nonlinear component is considered too large the fit is poor the segment is considered too short etc Press enter twice to write the conductivity of a segment to the Results file Data reduction with the TC PC program written for the Thermcon 85 system is based on a least squares fit of the measured temperatures to the following equation which is a variation of Equation XXX 107 8 4 T q Ak In f At B 6 PP Handbook Peter Blum November 1997 The constant A is the temperature drift rate also including edge effect asymmetry nonzero epoxy conductivity etc during measurement and is expressed in K min The constant B represents other imperfections in the experiment The unknowns in this system are k A and B so when more than three data pairs are acquired the system is overdetermined Using the previous equation for the rate of heating the coeffic
6. is the SAM that ensures Approximation that only results of physical significance are considered The critical choice of time Method SAM interval for calculation of conductivity selected manually by the user with the Thermcon 85 system is accomplished by an algorithm that automatically finds the optimal time interval The solution can be judged in great detail and the data reevaluated with different boundary parameters if warranted The following explanations are modified from the Teka user manual The first evaluation step is an approximation to the solution of a constantly heated line source Kristiansen 1982 T t A Aoln A3 In t A4179 9 The coefficients A are calculated with the least squares method A4 A3 and A4 are related to source geometry and thermal properties A is calculated by A q 4nk 10 where q is the heating power Wm and k W m K is the thermal conductivity If the coefficients A are determined T t can be expressed analytically and the apparent thermal conductivity K can be calculated by differentiating Equation on page 9 with respect to In k t dT din t q 47 A gt A3 1 t In t t Ay t 11 It can be shown that the desired value k is at kg tmax Where tnax 18 the extreme time The requirement for the maximum is dt Kg tmax 9 12 and tray 1S nage ADA A gt 0 13 The logarithm of the extreme time LET becomes PP Handbook Peter Blum Novemb
7. of conventional evaluations is not as good as that of SAM evaluations and the quality cannot be verified graphically The program TKCON EXE is used for the conventional evaluation The structure and application is similar to the TKSAM EXE program The configuration file TKCON INI includes the following standard parameters e minimum duration of interval 30 s e start time 30s e end time 120 s and e standard deviation of fit 0 003 PP Handbook Peter Blum November 1997 8 11 Half Space Measurements PERFORMANCE Precision Accuracy MEASUREMENT Standard Settings for Data Acquisition 8 12 Existing data can be evaluated later with the conventional method i e after the SAM method has failed to yield solutions Automatic Evaluation with TKCON can be set by typing TKMEAS EVA CON or if the option TKMEAS DCL 20 EVA CON is entered Calling TKMEAS without the EVA option invokes evaluation with TKSAM EXE A short list of results is created by TKCON with similar structure as the file created by TKSAM The difference is that instead of LET the standard deviation is reported The evaluation method used SAM CON is indicated in each line of the file A long list of results for each measurement can be produced by typing prior to starting TKMEAS set TKCON ON The long list includes the calculated values of thermal conductivity standard deviation and the start duration and end of each interval For the
8. represents an output string from the program If a position was not used some strings are omitted and some return zero values The file name is a combination of hole ID and run number Table 8 2 TC PC Processed Data file Short description Description Data file designations Leg Leg TC PC Results 1 4 leg Site Site TC PC Results 8 11 site Hole Hole TC PC Results 13 hole Core Core TC PC Results 15 17 core Core type Core type TC PC Results 19 core_type Section Section TC PC Results 21 22 section_or_std Top Interval top cm TC PC Results 24 28 interval_top Bottom Interval bottom cm TC PC Results 30 34 interval_bottom Space Space model TC PC Results 49 full_or_half Run No Run number TC PC Results 51 53 run_number Probe Probe number TC PC Results 55 57 probe_number Position Position number TC PC Results 59 position_number TC uncorr Uncorr thermal conductivity W m K TC PC Results 61 67 calculated_tc PP Handbook Peter Blum November 1997 Table 8 2 TC PC Processed Data file TC corr Corr thermal conductivity W m k TC PC Results 69 75 corrected_tc R2 Standard error R2 TC PC Results 77 87 standard_error Drift Calculated drift C s TC PC Results 89 97 calculated_drift Lower end Lower end point used TC PC Results 99 100 lower_end_point Fi
9. 8 THERMAL CONDUCTIVITY 8 1 Principles PHYSICAL BACKGROUND The coefficient of thermal conductivity k W m K is a measure of the rate q W at which heat flows through a material It is the coefficient of heat transfer across a steady state temperature difference T T4 over a distance x3 x1 or q k AT Ax 1 Thermal conductivity can be measured by transient heating of a material with a known heating power generated from a source of known geometry and measuring the temperature change with time The method assumes isotropic materials Theoretical discussion for measuring thermal conductivity with cylindrical sources is found in Blackwell 1954 Carslaw and Jaeger 1959 De Vries et al 1958 Von Herzen and Maxwell 1959 Kristiansen 1982 and Vacquier 1985 For a full space needle probe the length L can be assumed to be infinite and the problem is reduced to two dimensions Given the resistance R of a looped wire in a needle the generated heat is q 2i R L 2 where R L is the resistance of the needle per unit length At any time after heating has started the temperature T is related to the thermal conductivity k by T q 47k In C 3 where q is the heat input per unit length and unit time and C is a constant A simple way of calculating the thermal conductivity coefficient k is picking T and T at times ft and fy respectively from the temperature versus time measurement curve see also
10. ENU controls the overall data acquisition process e COLLECT communicates with the Thermcon 85 performs drift study collects raw data and writes raw data file monitors bad data conditions warnings not written to data file PP Handbook Peter Blum November 1997 CALIBRATION Power Supply Digital Volt Meter and Heater Current Needle Probe Resistance Needle Probe Secondary Calibration e PROCESS allows selection of probe positions allows for optional correction for temperature drift at drift study termination allows selection of optimal interval reduces the raw data and calculates thermal conductivity writes to a results file and e PROBES used to enter thermistor calibration coefficients for new probes and secondary probe calibration constants into the PROBES DAT file The user normally runs TCMENU Interaction with the COLLECT and PROCESS programs is accomplished via menu selection The calibration data must be entered into the PROBES DAT file when appropriate Calibration must be periodically performed by an ODP Electronics Technician Refer to the Thermcon 85 manual for details The thermistors in each needle probe are calibrated at the factory over a range of temperatures usually 15 to 75 C and fit to an equation of the form Tl alpha beta n R gamma In R 5 where T is the temperature in degrees Kelvin R is the thermistor resistance in ohms and alpha beta and gamma
11. _proc_thermcon tcon_proc_time_first tcon_proc_time_last tcon_raw_drift_status tcon_raw_pos_num tcon_run_minutes tcon_run_number tcon_run_status TCON cycle tcon_id PK1 FK tcon_cycle_num PK2 tcon_raw_heater_current tcon_raw_heater_curr_time tcon_raw_rel_voltage tcon_raw_rel_voltage_time TCON probe cycle tcon_id PK1 FK tcon_cycle_num PK2 FK tcon_probe_num PK3 tcon_raw_time tcon_raw_voltage The standard queries will be defined once the upload routine has been implemented PP Handbook Peter Blum November 1997 8 3 TKO4 System EQUIPMENT ODP purchased the TK04 system in late 1995 and deployed it permanently on the ship on Leg 168 1996 The system was to replace the ailing Thermcon 85 device built at the Woods Hole Oceanographic Institution WHOD and in service on the ship for many years Currently both systems are available to the user on the ship The TK04 was built by the Berlin company Teka based on an apparatus that had been developed at the Technische Universitat Berlin It was used successfully for thousands of measurements on material from the Continental Deep Drilling Program KTB The TK04 consists of e automatic self test heating and measurement unit TK04 e full space VLQ and half space HLQ needle probes e vice and manual hydraulic pump for half space contact measurements on rocks and e Macor standards for both types of needl
12. ame gt starts evaluation using the standard parameters no ERG file is created and e Batch mode evaluating a sequence of data files after typing TKSAM type return instead of a filename all DWL files in the directory will be evaluated The manufacturer s manual should be consulted for details in regard to file path requirements data quality issues etc Graphical Evaluation The program TKGRAPH can be used to visualize and judge the quality of all valid SAM evaluation results for thermal conductivity ERG files are required for plotting Four graphs are presented for each measurement e thermal conductivity vs LET e thermal conductivity vs interval duration e thermal conductivity vs start of interval and e thermal conductivity vs end of interval A series of files can also be viewed Consult the manufacturer s manual for system configuration practical hints guidance for the judgment of results etc Evaluation with Under certain experimental circumstances e g porous material high water Conventional Method content the SAM evaluation may not accept any results because the measurements are too disturbed for the sensitive approximations In these cases results may be obtained using the conventional evaluation method in which thermal conductivity is calculated from the inverse slope of the heating curve in a section of logarithmic linearity In general a heating duration gt 80 s becomes necessary Accuracy
13. e half space flag 1 half TC PC Raw 9 5 half_space_flag Probe m Probe secondary calibr slope TC PC Raw 9 6 probe_m1 Probe mg Probe secondary calibr intercept TC PC Raw 9 7 probe_m0 Lower end Lower end point probe calib s TC PC Raw 9 8 time_at_first_point Upper end Upper end point probe calib s TC PC Raw 9 9 time_at_last_point Comment Position specific comment TC PC Raw 10 comment Parameters repeated for other positions Drift time Drift no of readings length s TC PC Raw one line two values Drift study for first position Drift t T String of time temperature pairs TC PC Raw one line unlimited pairs Drift end Temp rate fit at end of drift study TC PC Raw one line three values Drift study repeated for other positions PP Handbook Peter Blum November 1997 Table 8 3 TC PC Raw Data file free format Drift status Drift status OK OVERRIDE TC PC Raw one line one alpha string Data for positions 1 5 Data Cycle ref volt 11 to 15 current TC PC Raw multiple lines 3 8 values per line Data repeated for each meas cycle Run status Run status NORMAL TC PC Raw one line one alpha string Notes The probe parameters of lines 4 10 are written for subsequent positions only if the positions were used otherwise the lines are omitted gt The drift study data lines two lines per position are always written to the file regardless wheth
14. e probes The TK04 measuring system features a self test at the beginning of each measuring cycle including probe number validation registration of the source temperature and its drift and calculation of the heating power used The following executable programs are used to operate the system e TKMEAS EXE to acquire time temperature data series creating DWL files e TKEVA for standard lt 5 uncertainty or special lt 2 uncertainty re evaluation of data creating short DAT or long ERG lists and parameter files and e TKGRAPH to display all solutions and assess the quality of the calculated solutions In addition the following parameter files are used e TKMEAS MNU a list of standard menu settings for TKMEAS EXE e INI list of parameters for probes where is the number engraved on the probe and e TKEVA INI list of user modifiable parameters required for TKEVA EXE Multiple measurements can be taken under identical conditions The instrument cycles through the measurements automatically creating files with the user defined root name e g Core Section Interval only six characters allowed and a two digit serial number incrementing by one for each measurement within a cycle The following files are created by the TK04 system e lt Rootname SerialNo gt DWL if Save data was selected contains measurement parameters and temperature time series raw data required for extended evaluations
15. er positions were used or not If a position was not used all values are zero Data are written on one line for each measurement cycle On each line there are the following readings separated in time by 3 s hard coded in the program 1 cycle number 2 internal reference voltage 3 to 7 up to five probe voltage readings no reading for unused positions 8 heater current Total time for one cycle is 2 lt number of positions used gt times 3 s 2 stands for reference and heater current readings It varies between 6 s no position used and 21 s five positions used Database Model Standard Queries Table 8 4 Database model TCON section TCON probe proc data TCON run tcon_id PK1 FK tcon_probe_num PK2 FK tcon_id PK1 FK tcon_probe_num PK2 icon_id PK1 top_interval bottom_interval section_id TCON control tcon_id PK1 FK tcon_probe_num PK2 FK standard_id PK3 FK TCON drift raw data tcon_id PK1 FK tcon_probe_num PK2 FK tcon_raw_drift_time PK3 tcon_raw_drift_temp tcon_comment tcon_meas_calib_m0 tcon_meas_calib_m1 tcon_meas_calib_time_first tcon_meas_calib_time_last tcon_meas_drift_lsq_fit tcon_meas_drift_rate_final tcon_meas_drift_temp_final tcon_probe_alpha tcon_probe_beta tcon_probe_gamma tcon_probe_half_full tcon_probe_specific_res tcon_proc_drift_corr_flag tcon_proc_point_first tcon_proc_point_last tcon
16. er 1997 8 9 Extended Evaluation 8 10 LET In f ax 243 A4 A3 14 The time dependent terms in previous equation are T tmax AI nax A3 Un tmax tmaxl Ast max 15 A can be substituted with previous Equation 118 to give Tit ma Al nay 2A30n t nax tmaxl 16 max max This equation shows that the purely logarithmic dependence of the approximated temperature required by the theory is stronger the larger tnax gets For large tmaw the second term in Equation on page 10 approaches zero The evaluation procedure approximates the heating curve in as many time intervals as possible and examines each interval for its suitability for thermal conductivity calculation using the following criteria 1 k t is located above a given value of time defined by LET 2 standard deviation of the function for A is below a given value 3 k t is a maximum A gt 0 and 4 derivation k t is continuous for t tmaxi A2tmax A3 9 If these criteria are met thermal conductivity can be calculated as k q 4mA 17 The evaluation interval is restricted by the dimension of the line source It must be within the interval of 20 to 80 s to avoid boundary effects and at least 25 s long for a stable calculation of the coefficients The input parameters for standard evaluation are e minimum duration of approximation interval 25 s e start of first approximation interval 20 s e end of last approximation
17. ient k can be determined at any time increment dt as k 2i2 R Ldin t 4 dT At BY or 7 k i R 2 AL dln t d7 8 The first group of terms in these equations is an instrument constant including generated heat and needle geometry The second group of terms is calculated for each measurement The optimum time segment for calculating thermal conductivity is selected interactively by the user by placing cross hairs on a In vs T plot of the data Information on the quality of the fit is updated on the screen as the cross hairs are moved The curve fit parameter is the root mean square of the temperature deviation and should not exceed 0 04 C min However it is more important to choose a consistent sampling time than it is to reduce the drift as much as possible DATA SPECIFICATIONS TC PC Output Files At present the TC PC data are not integrated in the new ODP database The following two program output files are archived the Processed Data or Results DAT files and the Raw Data TC files Data in the DAT files are fixed format mixed string and numeric with one record line per position per TC run If a given position on a run is not processed then there is no entry in this file However if a given position is processed more than once there are multiple lines in this file for that position The file name is the hole identifier Data in the RAW files are free format in which each line
18. interval 80 s e lower limit for LET 4 and e maximum standard deviation of calculated temperature curve from measured heating curve 0 0003 With the default parameters the heating curve is approximated for the following time intervals 20 45 20 46 20 47 20 78 20 79 20 80 21 46 21 47 21 78 21 79 21 80 22 47 22 78 22 79 22 80 53 78 53 79 53 80 54 79 54 80 55 80 Among all time intervals that fulfill the listed criteria the one with the largest LET is used to calculate thermal conductivity No solutions may be found if the measurement is disturbed by poor sample condition or ambient temperature changes An extended evaluation is required if PP Handbook Peter Blum November 1997 e the valid solutions are to be plotted against the calculation parameters to judge the results graphically or e the measurements are to be reevaluated with different parameters e g a stronger criterion for the LET In both cases the DWL files containing the temperature time data are required The ERG files long result lists that can be created contain all valid solutions for the thermal conductivity and a line entry in the TC LIST DAT file is created with the asymptotic optimal thermal conductivity value There are three options for extended evaluation e single evaluation typing lt TKSAM gt prompts for filename e batch mode with filename as parameter typing lt TKSAM filen
19. is required Accuracy About 5 Systematic evaluation is required MEASUREMENT 1 Bring cores to temperature equilibrium about 4 hr Hard rock specimens DATA PROCESSING 10 11 should be placed in a water bath to equilibrate Soft sediment drill holes into core liner Also drill a small hole in semiconsolidated sediment if necessary Apply thermal joint compound if necessary Insert full space probes carefully into sediment Rocks prepare smooth surface on a split core specimen at least 5 cm long Treat the needles gently and store them properly when not in use Insert one probe into a standard material for a control measurement to be used for later corrections if necessary Start the TCMENU program and follow the prompts for parameters Default values are provided for each prompt Press the reset button on the Thermcon 85 unit to start the drift study After a couple of minutes the drift data will be displayed The drift study is performed in phases of 25 minutes the maximum time the box can be programmed The drift study is terminated if all positions are equilibrated or if the user overrides the drift study Press the reset button twice to start the process of heating data acquisition and creation of the raw data file Messages will be displayed if there are data or hardware problems The user has the option of acquiring more data and processing batches of data later or processing the data collected immediately It
20. it is not necessary but strongly recommended to save the heating curves for routine evaluation These files allow later extended evaluation and graphical display of the solutions e lt Rootname gt LST short list of results from evaluating one root name batch of DWL files using either the special approximation method PP Handbook Peter Blum November 1997 SAM or conventional CON method contains evaluation parameters and the optimal calculated thermal conductivity value This is the standard results file e TC LIST DAT multiline short list optional contains the same information as previous file lt Rootname gt LST but for multiple root names This file is updated as new evaluations are performed This file is created only by the optional extended evaluation e lt Rootname gt ERG long lists of results from evaluating DWL files with the SAM method contains evaluation parameters and all valid calculated thermal conductivity values This file is optional and required only if graphical evaluation of all valid solutions is desired It can be created at any time if the DWL files are saved This file is created only by the optional extended evaluation CALIBRATION No calibration is required The unit conducts a self test at the beginning of each measurement cycle Macor standards are used to confirm the 1 65 W m K value DATA PROCESSING The Special The main advantage of the Teka data reduction program
21. pecimen at least 5 cm long Treat needles gently store them properly when not in use PP Handbook Peter Blum November 1997 3 On the computer change to directory containing the TKMEAS EXE file press enter 4 Type TKERG ON press enter 5 Type the command tkmeas press enter 6 Set the parameters on the screen Heating power should be about 5 W m adjust if necessary measuring time should be about 80 s enter Y to save time temperature data DATA SPECIFICATIONS TK04 Output Files Table 8 5 TK04 raw data file lt Rootname Serial gt DWL program output files are archived Currently TK04 data are not integrated in the new ODP database The following Short description Description Data file designation Header Filename Root name custom sample id serial TK04 Raw Data rootname_serial Probe Probe ID TK04 date TK04 Raw Data probe Comment Comment used to identify sample TK04 Raw Data comment Heat Heating power W m TK04 Raw Data heating_power Fit Slope Std dev temperature TK04 Raw Data fit 2 Something Reserved TK04 Raw Data something Value1 Some drift value 1 TK04 Raw Data value1 Value2 Some drift value 2 TK04 Raw Data value2 Data Temp Temperature C TK04 Raw Data temperature Time Time s TK04 Raw Data time Resistance Resistance ohm TK04 Raw Data resistance Table amp 6 TK04
22. rst time Time at lower end point s TC PC Results 102 104 time_at_first_point Upper end Upper end point used TC PC Results 106 107 upper_end_point Last time Time at upper end point s TC PC Results 109 111 time_at_last_point Drift status Drift study status TC PC Results 113 126 drift_status T drift Temp at drift study termination C TC PC Results 128 132 drift_temperature Drift rate Drift rate at termination C s TC PC Results 134 142 drift_rate Drift fit Least squares fit for drift TC PC Results 144 151 drift_fit Run status Run status NORMAL TC PC Results 153 160 run_status Alpha Probe alpha constant TC PC Results 162 180 probe_alpha Beta Probe beta constant TC PC Results 182 200 probe_beta Gamma Probe gamma constant TC PC Results 202 220 probe_gamma Resistance Probe wire resistance ohm cm TC PC Results 222 227 probe_wire_resistance Half space Probe half space flag 1 true TC PC Results 229 230 half_space_flag Probe m1 Probe secondary calibration slope TC PC Results 232 238 probe_m1 Probe mO Probe secondary calibration intercept TC PC Results 240 246 probe _m0 Lower end Upper end point probe calibration s TC PC Results 248 250 time_at_first_point Upper end Lower end point probe calibration s TC PC Results 252 254 time_at_last_point Drift corr Drift correction status TC PC Results 256 268 drift_correction_status Version Version of TC PC program TC PC Results 270 274 tcpc_version Commen
23. s from DWL file TK04 Results hints PP Handbook Peter Blum November 1997 8 13 Table 8 8 amp TKO4 extended results file ERG files Short description Description Data file designation Header SAM Evaluation Parameters TKSAM EXE Filename Root name serial sample ID TK04 Results rootname_serial Comment Comment used to identify sample TK04 Raw Data comment Time Time interval minimum s TK04 Results eval_interval_min Start time Start of evaluation s TK04 Results eval_time_start End time End of optimal time interval s TK04 Results eval_time_end LET Nat log of time TK04 Results eval_let Std Dev Limit of std dev optional 0 0003 TK04 Results eval_limit_sd Table 8 9 Valid solutions Short description Description Data file designation TC Calculated thermal conductivity TK04 Results calculated_tc LET Natural logarithm of time at max therm al condition TK04 Results let Start time Start of approx time interval s TK04 Results time_start Time Length of approx time interval s TK04 Results time_length End time End of optimal time interval s TK04 Results time_end Std Dev Standard deviation of fit TK04 Results std deviation Notes ERG files are optional They are created by extended evaluation and are required only for graphical evaluation They can be recreated from DWL files at any
24. t Comment TC PC Results 276 356 comment Notes The numbers following the file name TC PC Results are positions in the fixed space format of the output file Corrected thermal conductivity is corrected using the secondary probe calibration coefficients m4 and mp obtained from standard measurements Corrected thermal conductivity is added only if the user selects this option when specifying data reduction If correction is not selected the position numbers are reduced by 8 spaces starting with the Standard error field Table 8 3 TC PC Raw Data file free format 8 6 Short description Description Data file designations Run parameters Title Title string TC PC Raw 1 title Run Run number TC PC Raw 2 run_number Positions No of positions used length min TC PC Raw 3 no_of_positions_length Parameters for first position Sample ID ODP sample identification TC PC Raw 4 sample_id Piece Piece TC PC Raw 5 piece Subpiece Subpiece TC PC Raw 5 sub_piece Space Space model TC PC Raw 7 full_or_half Position no Position number TC PC Raw 8 position_number Alpha Probe alpha constant TC PC Raw 9 1 probe_alpha Beta Probe beta constant TC PC Raw 9 2 probe_beta Gamma Probe gamma constant TC PC Raw 9 3 probe_gamma Resistance Probe wire resistance ohm cm TC PC Raw 9 4 probe_wire_resistance Half space Prob
25. ted for in situ conditions USE OF THERMAL CONDUCTIVITY Thermal conductivity is an intrinsic material property for which the values depend on the chemical composition porosity density structure and fabric of the material e g Jumikis 1966 In marine geophysics mainly thermal conductivity profiles of sediment and rock sections are used along with temperature measurements to determine heat flow Heat flow is not only characteristic of the material but an indicator of type and age of ocean crust and fluid circulation processes at shallow and great depths 8 2 Thermcon 85 System EQUIPMENT The Thermcon 85 system consists of the following components e Thermcon 85 unit e calibrated needle probes e personal computer e TC PC control and data reduction program and e calibration file for TC PC The Thermcon 85 unit was purchased from Woods Hole Oceanographic Institution It is under the control of PROM based programming and an RS 232 serial interface is available One to five needle probes can be connected to the rear panel An eight channel multiplexer selects the appropriate input for each measurement See the Thermcon 85 manual for more details The needle probes are either assembled at ODP or purchased preassembled In either case they contain factory calibrated thermistors The TC PC program was developed at ODP in 1991 using Quick Basic v 4 5 and runs on a PC clone The following programs are involved e TCM
26. time Database Model 8 14 A database model and integration into the database are difficult to implement without writing an ODP sample identification routine linked to the TK04 output A better approach is to write an entirely new user interface for the system preferably for an upgraded version with multiple channel capability PP Handbook Peter Blum November 1997

Download Pdf Manuals

image

Related Search

Related Contents

DCN Next Generation Text/Status Display  32-bit PCI-Bus 10/100Mbps ACPI Fast Ethernet Card - Atlantis-Land  Participant User Manual p  神戸市民のくらしをまもる条例施行規則の一部を改正する規則をここに  ストラブ・クランプ Cタイプ  2755-6.6, Hand-Held Cordless Bar Code Scanners Programming  Sony DSC-S600 User's Guide  Descargar  RNA Assay User Guide  Toshiba SR-C8002 CD Player User Manual  

Copyright © All rights reserved.
Failed to retrieve file