A thermocouple calibrating furnace automatic control method based on temperature data acquisition

Abstract

This invention discloses an automatic control method for a thermocouple calibration furnace based on temperature data acquisition, belonging to the field of temperature control technology. The method includes: constructing an error decomposition model using a dynamic inertial prediction layer, a residual structure separation layer, and a disturbance suppression layer based on temperature-related dynamic response parameters, and performing frequency separation to obtain the intrinsic measurement error components; correcting the apparent temperature sequence based on the intrinsic measurement error components, and performing dynamic compensation inversion in conjunction with the temperature-related dynamic response parameters to generate a true temperature sequence and a temperature reliability index; evaluating the stability of the calibration temperature field based on the true temperature sequence and the temperature reliability index, generating heating control commands, adjusting the heating power in the calibration furnace, and generating a calibration temperature field control report. This invention, by constructing temperature-related dynamic response parameters and establishing an error decomposition model, achieves a structured analysis of the dynamic hysteresis behavior and intrinsic measurement errors of thermocouples by temperature sensors.

Description

Technical Field

[0001] This invention relates to the field of temperature control technology, and in particular to an automatic control method for a thermocouple calibration furnace based on temperature data acquisition. Background Technology

[0002] Thermocouples, as one of the most widely used temperature sensors in industrial temperature measurement, are widely applied in metallurgy, metrology calibration, energy equipment, and high-end manufacturing due to their simple structure, wide measurement range, and excellent high-temperature resistance. To ensure the accuracy and traceability of thermocouple measurement results, periodic calibration of the thermocouples under test is typically performed using a thermocouple calibration furnace. Traditional thermocouple calibration furnace control methods mainly rely on a combination of constant temperature control and manual stabilization assessment. This involves acquiring thermoelectric potential signals and performing temperature conversion using a standard calibration table, then reading the temperature data after a settling time to complete the calibration. With the increasing demands for temperature measurement accuracy, existing technologies are gradually incorporating digital sampling, automatic data acquisition and recording, and simple filtering methods to improve data acquisition efficiency and stability.

[0003] Existing automatic control methods for thermocouple calibration furnaces have two main shortcomings: First, current technologies typically fail to finely characterize the dynamic response process of thermocouples and do not distinguish between trend errors caused by sensor thermal inertia and fluctuation errors caused by random disturbances. This leads to error compensation often relying on empirical corrections or simple smoothing methods, making it difficult to achieve dynamic inversion of the true temperature. Second, existing calibration processes often use a single stability criterion as the basis for the validity of the temperature field, lacking a comprehensive evaluation mechanism for the spatiotemporal coupling characteristics of the temperature field and the reliability of measurements. Consequently, they cannot perform structured analysis and zonal identification of the dynamic stability state of the calibration temperature field. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an automatic control method for thermocouple calibration furnace based on temperature data acquisition to solve the problems of insufficient reconstruction of temperature measurement accuracy and determination of temperature field validity.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides an automatic control method for a thermocouple calibration furnace based on temperature data acquisition. The method includes: acquiring thermocouple temperature data and performing smoothing compensation processing to generate an apparent temperature sequence and time-varying characteristic data; identifying temperature-related dynamic response characteristics in the apparent temperature sequence and time-varying characteristic data to generate temperature-related dynamic response parameters; constructing an error decomposition model based on the temperature-related dynamic response parameters using a dynamic inertial prediction layer, a residual structure separation layer, and a disturbance suppression layer, and performing frequency separation to obtain intrinsic measurement error components; correcting the apparent temperature sequence based on the intrinsic measurement error components, and performing dynamic compensation inversion in conjunction with the temperature-related dynamic response parameters to generate a true temperature sequence and a temperature reliability index; evaluating the stability of the calibration temperature field based on the true temperature sequence and temperature reliability index, generating heating control commands, adjusting the heating power inside the calibration furnace, and generating a calibration temperature field control report.

[0007] As a preferred embodiment of the automatic control method for thermocouple calibration furnace based on temperature data acquisition described in this invention, the thermocouple temperature data includes thermoelectric potential signal, timestamp, location identifier and cold junction temperature measurement value.

[0008] As a preferred embodiment of the automatic control method for thermocouple calibration furnace based on temperature data acquisition according to the present invention, the specific steps of acquiring thermocouple temperature data and performing smoothing compensation processing to generate apparent temperature sequence and time change characteristic data are as follows: Cold junction compensation is performed on the thermocouple temperature data, and calibration conversion is carried out to generate an apparent temperature sequence. The apparent temperature series is continuously mean smoothed and then subjected to a fixed time interval difference operation to generate time-varying feature data.

[0009] As a preferred embodiment of the automatic control method for thermocouple calibration furnace based on temperature data acquisition described in this invention, the step of generating temperature-related dynamic response parameters by identifying the temperature-related dynamic response characteristics in the apparent temperature sequence and time-varying feature data is as follows: Identify the starting and turning points of temperature changes in the apparent temperature sequence and divide the apparent temperature sequence into temperature response event segment sequences. Perform cross-delay estimation on the temperature response event segment sequence to generate a dynamic hysteresis feature set of the response segment; The dynamic hysteresis feature set of the response segment is encoded using hysteresis fingerprinting, and a cyclic consistency check is performed to establish a temperature mapping relationship, generating temperature-related dynamic response parameters.

[0010] As a preferred embodiment of the automatic control method for thermocouple calibration furnace based on temperature data acquisition described in this invention, the specific steps for constructing an error decomposition model using a dynamic inertial prediction layer and a residual structure separation layer based on temperature-related dynamic response parameters are as follows: The dynamic inertial prediction layer reconstructs the response trajectory of the apparent temperature sequence based on temperature-related dynamic response parameters, and generates a dynamic inertial predicted temperature sequence by constructing a predicted temperature evolution path. The residual structure separation layer extracts the residual signal between the apparent temperature sequence and the dynamic inertial prediction temperature sequence, and performs trend decomposition to obtain the inertial residual trend component and fluctuation residual component. The dynamic inertial prediction layer and the residual structure separation layer are decomposed into an error decomposition model.

[0011] As a preferred embodiment of the automatic control method for thermocouple calibration furnace based on temperature data acquisition described in this invention, the measured intrinsic error component refers to the random disturbance component obtained by performing frequency separation on the fluctuating residual component based on the error decomposition model and combining it with the inertial residual trend component for point-by-point superposition.

[0012] As a preferred embodiment of the automatic control method for thermocouple calibration furnace based on temperature data acquisition described in this invention, the error correction of the apparent temperature sequence based on the intrinsic measurement error component refers to using an error cancellation mapping algorithm to subtract the intrinsic measurement error component from the apparent temperature sequence point by point to generate an error-corrected temperature sequence.

[0013] As a preferred embodiment of the automatic control method for thermocouple calibration furnace based on temperature data acquisition described in this invention, the specific steps for generating the true temperature sequence and temperature reliability index are as follows: Based on temperature-related dynamic response parameters, the error-corrected temperature sequence is compensated for response time backtracking through a dynamic response inverse mapping algorithm to generate a true temperature sequence. The error residual ratio is calculated based on the actual temperature sequence and the measurement endogenous error components, generating error residual characteristics, which are then converted into normalized confidence indicators to generate the temperature confidence index.

[0014] As a preferred embodiment of the automatic control method for thermocouple calibration furnace based on temperature data acquisition described in this invention, the specific steps for evaluating the stability of the calibration temperature field based on the actual temperature sequence and temperature reliability index, and generating heating control commands, are as follows: A temporal spatial expansion matrix is ​​constructed based on the real temperature sequence and the temperature confidence index, and a confidence gating mapping algorithm is executed to filter the temporal confidence and generate a reliable temperature evolution matrix. Slowly varying energy components and rapidly perturbation energy components are extracted from the reliable temperature evolution matrix, and cross-mapping calculations are performed to generate a coupled temperature field energy distribution map. The energy field topology decomposition operation is performed on the coupled temperature field energy distribution map to identify continuous stable regions and abrupt boundary regions in the spatiotemporal dimension, and then partitioned control mapping is performed to generate heating control commands.

[0015] As a preferred embodiment of the automatic control method for thermocouple calibration furnace based on temperature data acquisition described in this invention, the calibration temperature field control report is generated by adjusting the heating power inside the calibration furnace according to the heating control command, and combining the changes in temperature field zoning before and after adjustment.

[0016] The beneficial effects of this invention are as follows: by constructing temperature-related dynamic response parameters and establishing an error decomposition model, a structured analysis of the dynamic hysteresis behavior and intrinsic measurement error of thermocouples is realized; by generating a true temperature sequence and temperature reliability index through dynamic compensation inversion, the traditional verification method based on apparent temperature is upgraded to a precise control method based on dynamic modeling and error inversion, thereby improving the accuracy of temperature measurement in the dynamic stage and the ability to determine the effectiveness of the verification temperature field. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of an automatic control method for a thermocouple calibration furnace based on temperature data acquisition.

[0019] Figure 2 A flowchart generated for dynamic response parameters.

[0020] Figure 3 A flowchart for constructing the error decomposition model.

[0021] Figure 4 This is a flowchart for evaluating the stability of the temperature field.

[0022] Figure 5 This is a comparison chart of the dynamic compensation inversion effect.

[0023] Figure 6 This is a statistical comparison chart of temperature measurement errors. Detailed Implementation

[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0026] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0027] Reference Figures 1-6 This is one embodiment of the present invention, which provides an automatic control method for a thermocouple calibration furnace based on temperature data acquisition, comprising the following steps: S1: Collect thermocouple temperature data and perform smoothing compensation processing to generate apparent temperature sequence and time change characteristic data.

[0028] S1.1: Thermocouple temperature data includes thermoelectric potential signal, timestamp, location identifier and cold junction temperature measurement value.

[0029] Specifically, the thermoelectric potential signal is obtained by connecting the two ends of the thermocouple to the differential voltage input terminal, and the analog-to-digital converter continuously samples the millivolt-level voltage generated at the two ends of the thermocouple at a fixed sampling time interval (e.g., once per second). The analog-to-digital converter converts the analog voltage into the corresponding digital voltage value at each sampling to obtain the thermoelectric potential signal.

[0030] The timestamp is generated by a unified clock and the current value is read each time the thermoelectric potential signal is acquired; after establishing a one-to-one correspondence with the corresponding digital voltage values, they are stored synchronously to obtain the timestamp.

[0031] The position identifier is obtained by recording the actual measurement position of the thermocouple in the calibration furnace. When the thermocouple is in a fixed measurement state, the corresponding measurement hole number is recorded. When the thermocouple moves along the axis of the calibration furnace, the current position value is read and the current position value is synchronously associated with the corresponding timestamp to obtain the position identifier corresponding to the thermoelectric potential signal data sequence.

[0032] A temperature sensor is installed at the location of the thermocouple reference end to sense the changes in the ambient temperature of the thermocouple reference end in real time. The analog temperature signal output by the temperature sensor is converted into a digital temperature value by a voltage sampling circuit. At the moment of sampling, a corresponding timestamp is generated by a unified clock. The digital temperature value and the corresponding timestamp are stored one-to-one to generate the cold junction temperature measurement value.

[0033] S1.2: Perform cold junction compensation on the thermocouple temperature data and perform calibration conversion to generate an apparent temperature sequence.

[0034] Specifically, the thermoelectric potential signal and corresponding timestamp are extracted from the thermocouple temperature data. At the same time, the cold junction temperature measurement value and corresponding timestamp are extracted. The thermoelectric potential signal and cold junction temperature measurement value are matched point by point according to the timestamp to form a time-consistent thermoelectric potential signal sequence and cold junction temperature measurement value sequence.

[0035] The cold junction temperature measurement value is converted by looking up the standard calibration table corresponding to the type of thermocouple used (e.g., type K thermocouple and type S thermocouple) to obtain the equivalent thermoelectric potential value of the cold junction. The thermoelectric potential signal and the equivalent thermoelectric potential value of the cold junction are matched point by point according to the timestamp, so that each timestamp corresponds to a unique set of thermoelectric potential signal value and equivalent thermoelectric potential value of the cold junction. When the direction of the thermoelectric potential signal is consistent with the direction of the equivalent thermoelectric potential of the cold junction, the values ​​are directly superimposed. When the direction of the thermoelectric potential signal is opposite to the direction of the equivalent thermoelectric potential of the cold junction, the difference between the thermoelectric potential signal and the equivalent thermoelectric potential of the cold junction is used as the compensated thermoelectric potential value. The compensated thermoelectric potential values ​​are arranged in the order of the timestamps to form a sequence of compensated thermoelectric potential values.

[0036] The compensated thermoelectric potential value sequence is converted point by point according to the standard calibration table to obtain a temperature value sequence that corresponds one-to-one with the timestamp; the temperature value sequence is arranged according to the timestamp order to form an apparent temperature sequence that is continuously distributed over time.

[0037] It should be noted that the standard calibration table is a table showing the correspondence between thermocouple potential and temperature. It is used to convert the thermoelectric potential value generated by the thermocouple under a specified reference junction temperature condition into the corresponding temperature value. The standard calibration table gives the potential value corresponding to each temperature point or the temperature value corresponding to each potential value according to different thermocouple types, so as to obtain the temperature value corresponding to the actual thermoelectric potential.

[0038] S1.3: Perform continuous mean smoothing on the apparent temperature series and perform fixed time interval difference operations to generate time-varying feature data.

[0039] Specifically, the apparent temperature sequence is traversed point by point according to the timestamp order, and a sliding window is formed by a fixed number of consecutive temperature values ​​(e.g., five consecutive sampling points form a sliding window). The temperature value corresponding to the center position of the sliding window is replaced with the average value of the sliding window, thereby obtaining the smoothed temperature sequence.

[0040] Based on the smoothed temperature sequence, the temperature difference between adjacent time nodes is taken as the temperature change. Combined with the temperature change sequence at a fixed time interval, the ratio of the temperature change to the fixed sampling time interval is taken as the temperature change rate and arranged in the order of timestamps to obtain the temperature change rate sequence. The temperature change sequence and the temperature change rate sequence together constitute the time change feature data.

[0041] S2: By identifying the temperature-related dynamic response characteristics in the apparent temperature sequence and time-varying feature data, temperature-related dynamic response parameters are generated.

[0042] S2.1: Identify the starting and turning points of temperature changes in the apparent temperature sequence and divide the apparent temperature sequence into temperature response event segment sequences.

[0043] Specifically, the apparent temperature sequence is traversed point by point according to the timestamp order, while retrieving the temperature change quantity sequence and temperature change rate sequence from the time change feature data; the starting point of temperature change is identified based on the sign change of the temperature change rate sequence. When the temperature change rate changes from a fluctuating state of zero to a continuous positive or negative value, the corresponding timestamp position is determined as the starting point of temperature change.

[0044] During the continuous temperature change, the temperature change rate sequence is traversed. When the temperature change rate changes from a positive value to a negative value or from a negative value to a positive value, the corresponding timestamp position is determined as the temperature change inflection point. The continuous time interval between the temperature change start point and the next temperature change inflection point is determined as an independent temperature response event segment, and the temperature response event segment sequence is obtained.

[0045] S2.2: Perform cross-delay estimation on the temperature response event segment sequence to generate a dynamic lag feature set of the response segment.

[0046] Specifically, each temperature response event segment sequence is extracted segment by segment according to the timestamp order, and the temperature change rate sequence in the time change feature data is extracted synchronously within the corresponding time range; within each temperature response event segment sequence, the apparent temperature sequence is standardized, and the temperature change rate sequence is standardized for the same length to obtain the standardized apparent temperature sequence and the standardized temperature change rate sequence.

[0047] The standardized apparent temperature sequence and the standardized temperature change rate sequence within the same temperature response event segment sequence are compared point by point, and the correlation under different time offset conditions is calculated. The time offset corresponding to the maximum correlation is obtained as the dynamic hysteresis characteristic value of the temperature response event segment sequence.

[0048] The expression for calculating the correlation of the time offset condition is: ; in, The degree of correlation of the time offset condition. This represents the total number of sampling points in the current temperature response event segment. This is the time offset. For the sampling point index, To standardize the apparent temperature series, This is a standardized temperature change rate sequence.

[0049] Record the dynamic hysteresis feature values ​​corresponding to each temperature response event segment sequence in the order of timestamps, and generate a set of dynamic hysteresis features of response segments arranged by time.

[0050] S2.3: Perform hysteresis fingerprint encoding on the dynamic hysteresis feature set of the response segment, and perform cyclic consistency check to establish temperature mapping relationship and generate temperature-related dynamic response parameters.

[0051] Specifically, the dynamic hysteresis feature set of the response segment is arranged according to the timestamp order, and the apparent temperature sequence within the corresponding time range is retrieved. According to the time order of the temperature response event segment sequence, the dynamic hysteresis feature value corresponding to each temperature response event segment sequence is paired with the apparent temperature value in the corresponding segment, and a hysteresis fingerprint sequence is formed according to the order of occurrence.

[0052] Based on the temporal sequence of temperature response event segments, the hysteresis fingerprint sequences corresponding to the heating process and the cooling process are compared one by one. By comparing the degree of difference in dynamic hysteresis characteristic values ​​within the corresponding temperature range, it is determined that the dynamic hysteresis characteristic values ​​change in the same direction within the temperature range are cyclically consistent, and that they change in different directions are cyclically inconsistent. A mapping relationship between temperature and dynamic hysteresis characteristic values ​​is established, and temperature-related dynamic response parameters are generated by arranging them in order of apparent temperature values.

[0053] It should be noted that the hysteresis fingerprint sequence refers to the serialized characterization obtained by continuously depicting the time delay characteristics formed by sensor thermal inertia, heat transfer hysteresis and response delay during the temperature measurement process, which provides a basis for subsequent dynamic response inverse mapping, real temperature backtracking compensation and temperature field stability analysis.

[0054] S3: Based on the temperature-related dynamic response parameters, an error decomposition model is constructed using a dynamic inertial prediction layer and a residual structure separation layer, and frequency separation is performed to obtain the intrinsic measurement error components.

[0055] S3.1: The dynamic inertial prediction layer reconstructs the response trajectory of the apparent temperature sequence based on temperature-related dynamic response parameters, and generates a dynamic inertial predicted temperature sequence by constructing a predicted temperature evolution path.

[0056] Specifically, the apparent temperature sequence and temperature-related dynamic response parameters are synchronously arranged according to the timestamp order, and a one-to-one correspondence between the apparent temperature sequence and the temperature-related dynamic response parameters is established under the same timestamp condition. Based on the dynamic lag characteristic value contained in the temperature-related dynamic response parameters and the temperature change corresponding to each time node in the apparent temperature sequence, the temperature change at the current time node is adjusted backward according to the time delay indicated by the corresponding dynamic lag characteristic value, and the temperature change at the backtracked time node is used as the predicted temperature change.

[0057] The predicted temperature changes are statistically analyzed in chronological order to form a continuous predicted temperature evolution path. The predicted temperature evolution path is then arranged in chronological order to obtain a dynamic inertial predicted temperature sequence that corresponds one-to-one with the timestamp.

[0058] S3.2: The residual structure separation layer extracts the residual signal between the apparent temperature sequence and the dynamic inertial prediction temperature sequence, and performs trend decomposition to obtain the inertial residual trend component and the fluctuation residual component.

[0059] Specifically, the apparent temperature sequence and the dynamic inertial prediction temperature sequence are arranged synchronously according to the timestamp order. Under the same timestamp condition, the difference between the apparent temperature sequence and the dynamic inertial prediction temperature sequence is counted point by point to form a residual signal sequence arranged continuously in time.

[0060] Subsequently, the residual signal sequence is traversed point by point according to the timestamp order. A sliding window is formed by a fixed number of continuous residual signal values, and the residual signal value corresponding to the center position of the sliding window is replaced with the average value of the sliding window to form an inertial residual trend component that is continuously distributed over time.

[0061] After obtaining the inertial residual trend component, the difference between the residual signal value at the corresponding timestamp position in the residual signal sequence and the inertial residual trend component value at the same timestamp position is taken as the fluctuation residual component arranged continuously in time, thus obtaining the inertial residual trend component and the fluctuation residual component.

[0062] S3.3: Decompose the errors of the dynamic inertial prediction layer and the residual structure separation layer to construct an error decomposition model.

[0063] Specifically, the dynamic inertial prediction layer reconstructs the response trajectory of the apparent temperature sequence based on temperature-related dynamic response parameters and generates a dynamic inertial prediction temperature sequence; the residual structure separation layer uses the difference between the apparent temperature sequence and the dynamic inertial prediction temperature sequence point by point according to the timestamp order as the residual signal sequence.

[0064] After the residual signal sequence is formed, the residual structure separation layer forms a sliding window with a fixed number of continuous residual signal values, and the median of the continuous residual signal values ​​in the sliding window is determined as the inertial residual trend component; the difference between the value at the corresponding timestamp position in the residual signal sequence and the value at the same timestamp position in the inertial residual trend component is taken as the fluctuation residual component.

[0065] The inertial residual trend component and the fluctuation residual component are used as the results of trend and fluctuation separation, and the inertial residual trend component and the fluctuation residual component are jointly determined as the decomposition result of the error decomposition model to generate the error decomposition model.

[0066] S3.4: Based on the error decomposition model, frequency separation is performed on the fluctuation residual component to extract the random disturbance component, and the inertial residual trend component is combined with point-by-point superposition to obtain the measurement intrinsic error component.

[0067] It should be noted that, based on the error decomposition model, a fixed number of continuous sampling points are selected in the order of timestamps to form a spectrum analysis segment, and fast Fourier transform processing is performed to obtain the amplitude sequence corresponding to the frequency; the amplitude sequence is traversed point by point from low to high frequency, and the average value of the current amplitude and the amplitude of the adjacent frequency points is calculated at each frequency point, and the difference between the current amplitude and the adjacent average amplitude is compared.

[0068] When the current amplitude is higher than the adjacent average amplitude at multiple consecutive frequency points and forms a continuous peak-shaped interval, the corresponding frequency interval is determined as a periodic structure frequency component. When the amplitude is discretely distributed on the frequency axis and does not form a continuous peak-shaped interval, the corresponding frequency point is determined as a random disturbance component.

[0069] The frequency components corresponding to the random disturbance components are retained in the frequency domain, while the frequency components of the periodic structure are set to zero in the frequency domain. The frequency components corresponding to the retained random disturbance components are subjected to inverse fast Fourier transform processing to obtain a sequence of random disturbance components that correspond one-to-one with the timestamp. The random disturbance component sequence is then superimposed point by point with the inertial residual trend component obtained from the error decomposition model in the order of the timestamps to form the measurement endogenous error component that corresponds one-to-one with the timestamp.

[0070] S4: Based on the measurement endogenous error components, the apparent temperature sequence is corrected for errors, and combined with the temperature-related dynamic response parameters for dynamic compensation inversion to generate the true temperature sequence and temperature confidence index.

[0071] S4.1: An error cancellation mapping algorithm is used to subtract the intrinsic measurement error components from the apparent temperature sequence point by point to generate an error-corrected temperature sequence. Specifically, the apparent temperature value and the measurement intrinsic error component value at the same timestamp position are read simultaneously in the order of timestamps and synchronized in time sequence; starting from the position corresponding to the minimum timestamp value, the system traverses point by point backwards, extracting the corresponding apparent temperature value and measurement intrinsic error component value at each timestamp position; when the measurement intrinsic error component is positive, the difference between the apparent temperature value and the measurement intrinsic error component value is calculated; when the measurement intrinsic error component is negative, the sum of the absolute values ​​of the apparent temperature value and the measurement intrinsic error component value is calculated.

[0072] The difference between the apparent temperature value and the measured intrinsic error component value, and the sum of the absolute values ​​of the apparent temperature value and the measured intrinsic error component value, are written into the corresponding timestamp position in the new temperature sequence, and arranged in the order of the timestamps to form an error correction temperature sequence.

[0073] It should be noted that the error cancellation mapping algorithm is a numerical mapping method that corrects the apparent temperature sequence point by point based on the measurement endogenous error component. It uses the decomposed measurement endogenous error component to cancel the apparent temperature value under the condition of one-to-one correspondence of timestamps, deducts the error component from the apparent temperature value, eliminates the deviation caused by dynamic lag and random disturbance, and realizes the mapping conversion of apparent temperature to error correction temperature.

[0074] S4.2: Based on temperature-related dynamic response parameters, the error-corrected temperature sequence is compensated for by response time backtracking through the dynamic response inverse mapping algorithm to generate the true temperature sequence; Specifically, the temperature values ​​at each timestamp in the error correction temperature sequence are read in the order of timestamps. At the same time, the dynamic hysteresis characteristic values ​​contained in the temperature-related dynamic response parameters corresponding to the same timestamp position are read, and the dynamic hysteresis characteristic values ​​are used as the corresponding time delay.

[0075] Search for the historical timestamp position corresponding to the time delay amount from the current timestamp position towards the leading time axis. When the time delay amount is an integer multiple of the fixed sampling time interval, directly locate the corresponding timestamp position and read the temperature value of the error-corrected temperature sequence at the timestamp position; use the temperature value at the timestamp position as the true temperature value at the current timestamp position.

[0076] When the time delay is not an integer multiple of the fixed sampling time interval, linear interpolation is performed between two adjacent timestamp positions according to the time ratio to obtain the true temperature value at the current timestamp position; the obtained true temperature value is written into the timestamp position corresponding to the new temperature sequence, arranged according to the timestamp order to generate the true temperature sequence.

[0077] It should be noted that the dynamic response inverse mapping algorithm is based on the time backtracking compensation of the error correction temperature sequence using temperature-related dynamic response parameters. It uses the identified dynamic hysteresis feature value as the time delay, searches forward in the timestamp sequence for the corresponding temperature position, obtains the temperature value at the corresponding time, and thus reverses the response hysteresis of the temperature sensor to achieve the time dimension mapping transformation from the error correction temperature to the true temperature.

[0078] S4.3: Calculate the residual error ratio based on the actual temperature sequence and the intrinsic error components of the measurement, generate residual error characteristics, and convert them into normalized confidence index to generate the temperature confidence index.

[0079] Specifically, the actual temperature value at each timestamp in the real temperature sequence and the error value at the same timestamp in the measurement endogenous error component are read simultaneously in the order of timestamps, maintaining synchronous traversal in time order; the ratio of the actual temperature value at each timestamp to the measurement endogenous error component is used as the error residual ratio, and they are arranged in the order of timestamps to form an error residual ratio sequence.

[0080] Traverse the error residual ratio sequence in timestamp order, determine the maximum and minimum ratio values ​​in the error residual ratio sequence, fix the maximum and minimum ratio values ​​as the upper and lower boundaries of the linear scaling interval, and calculate the normalized confidence index, expressed as: ; in, As a normalized credibility index, For the first The percentage of residual error at each timestamp location. For timestamp index, It is the minimum value in the error residual proportion sequence. This represents the maximum value in the error residual proportion sequence.

[0081] The normalized confidence index was arranged in chronological order according to timestamps and determined as the temperature confidence index.

[0082] Figure 5The results show a comparison of temperature response during the dynamic stages. The upper part is the full-time temperature evolution curve, and the lower part is a magnified view of the dynamic segment indicated by the red dashed box. The first curve, the "real temperature sequence," originates from the set temperature control curve of the calibration furnace and the sampling results of the high-precision reference temperature measuring device, serving as a benchmark sequence in simulation and verification. The second curve, the "apparent temperature sequence," is the raw temperature sampling data obtained by collecting the thermocouple output signal. The apparent temperature sequence has not undergone dynamic compensation processing and only includes the dynamic hysteresis behavior and intrinsic measurement error components of the thermocouple under actual operating conditions. The third curve, the "real temperature sequence," is generated by constructing temperature-related dynamic response parameters based on the apparent temperature sequence and establishing an error decomposition model to decompose the apparent temperature sequence into dynamic hysteresis components and intrinsic measurement error components. Based on the temperature-related dynamic response parameters and the error decomposition model, a dynamic compensation inversion operation is performed to generate the real temperature sequence. Figure 5 As can be seen, the three curves basically overlap in the steady-state region, while in the step event region, the apparent temperature sequence lags significantly behind the true temperature sequence. After dynamic compensation and inversion processing, the rise rate of the true temperature sequence tends to be consistent with that of the true temperature sequence, and the maximum deviation decreases. A magnified view further shows that at the point of maximum difference, the deviation between the true temperature sequences is smaller than the deviation between the apparent temperature sequence and the true temperature sequence, thus verifying the structured analytical capability for the dynamic hysteresis behavior of thermocouples and the compensation effect for intrinsic measurement error components.

[0083] S5: Evaluate the stability of the calibration temperature field based on the actual temperature sequence and temperature reliability index, generate heating control commands, adjust the heating power in the calibration furnace, and generate a calibration temperature field control report.

[0084] S5.1: Construct a time-series spatial expansion matrix based on the real temperature sequence and temperature confidence index, and perform a confidence-gated mapping algorithm to filter the time-domain confidence and generate a reliable temperature evolution matrix. Specifically, the real temperature sequence and temperature confidence index are read in the order of timestamps. The real temperature value and temperature confidence index are extracted at the same timestamp position and arranged in the order of timestamps and columns of position identifiers. The real temperature values ​​corresponding to different timestamps are filled into the matrix row vector in the order of position identifiers. The temperature confidence index of the same timestamp is stored in the corresponding position to complete the construction of the temporal spatial expansion matrix.

[0085] In the temporal spatial expansion matrix, the actual temperature value and temperature confidence index of each time node are traversed row by row in timestamp order. The temperature confidence index is then compared with a confidence threshold (based on the numerical range obtained by linear normalization of the temperature confidence index, with a value range of: The system compares the temperature confidence index with the real temperature value. When the temperature confidence index is not less than the confidence threshold, the real temperature value is retained. When the temperature confidence index is less than the confidence threshold, the real temperature value is marked as null.

[0086] After completing the credibility screening for each time node, the real temperature values ​​retained by the credibility gating mapping algorithm are rearranged according to the original timestamp and location identifier order to form a credibility temperature evolution matrix unfolded by time and space.

[0087] It should be noted that the credibility gating mapping algorithm uses the temperature credibility index as a gating criterion to judge the credibility of the real temperature value under the condition of one-to-one correspondence of timestamps. When the temperature credibility index meets the credibility threshold, the corresponding real temperature value is retained. When the temperature credibility index does not meet the credibility threshold, the corresponding real temperature value is marked. This realizes the automatic filtering of low credibility data, so that the subsequent temperature field analysis is only based on high credibility temperature data, thereby improving the reliability of the temperature field stability evaluation.

[0088] S5.2: Extract the slowly varying energy components and the rapidly perturbation energy components from the reliable temperature evolution matrix, and perform cross-mapping calculations to generate a coupled temperature field energy distribution map.

[0089] Specifically, the reliable temperature evolution matrix is ​​traversed row by row according to the timestamp order. In each row of the reliable temperature evolution matrix, the reliable temperature values ​​are read in order according to the position identifier and a time temperature vector is formed. A time sliding window is formed by a fixed number of consecutive timestamps corresponding to the time temperature vectors. The continuous mean of the reliable temperature values ​​corresponding to the same position identifier in the time sliding window is calculated to obtain the slowly varying temperature sequence. The square value of the slowly varying temperature sequence at each timestamp position is taken as the slowly varying energy component.

[0090] The difference between the reliable temperature value corresponding to the timestamp and location identifier in the reliable temperature evolution matrix and the slowly varying temperature value at the same timestamp and location identifier in the slowly varying temperature sequence is taken as the fast perturbation temperature. These values ​​are then arranged in chronological order to generate a fast perturbation temperature sequence. The square of the fast perturbation temperature sequence at each timestamp is taken as the fast perturbation energy component.

[0091] Cross-mapping calculations are performed on the slow-varying energy components and the fast-disturbance energy components according to the two-dimensional index of timestamp and location identifier. The product of the slow-varying energy component and the fast-disturbance energy component under the same timestamp and location identifier is used as the coupling energy value. The coupling energy values ​​are filled in the energy matrix according to the timestamp as the row order and the location identifier as the column order.

[0092] The energy matrix is ​​arranged continuously along the time dimension and expanded according to the position labels to obtain the coupled temperature field energy distribution map.

[0093] It should be noted that the coupled temperature field energy distribution map is used to uniformly correlate and express the slowly varying energy components and rapidly perturbation energy components in the reliable temperature evolution matrix, reflecting the stable accumulation characteristics and local perturbation characteristics of the calibration temperature field in both temporal and spatial dimensions. This provides a unified basis for subsequent energy field topology decomposition operations, identification of continuous stable regions and abrupt boundary regions, and generation of heating control commands, thereby improving the accuracy of the calibration temperature field stability evaluation and the targeted nature of heating regulation.

[0094] S5.3: Perform energy field topological decomposition operation on the coupled temperature field energy distribution map, identify continuous stable regions and abrupt boundary regions in the spatiotemporal dimension, perform partitioned control mapping, and generate heating control commands.

[0095] Specifically, the coupled energy values ​​in the coupled temperature field energy distribution map are read according to the timestamp as the row order and the position identifier as the column order, and a two-dimensional adjacency relationship is constructed based on the adjacency relationship of the timestamp and the adjacency relationship of the position identifier. By calculating the difference between each coupled energy value in the coupled temperature field energy distribution map and the corresponding coupled energy value of the adjacent timestamp position and the adjacent position identifier position, a coupled energy change matrix is ​​formed.

[0096] The coupled energy change matrix is ​​processed by connected region extraction. Grids with continuous coupled energy changes and consistent trends are divided into the same connected region, and the timestamp range and location identifier range covered by the connected region are recorded. Connected regions in the coupled energy change matrix whose change amplitude remains unchanged in the local neighborhood and covers multiple continuous timestamps and location identifiers are marked as continuous stable regions. Regions where coupled energy changes jump between adjacent grids and form boundary morphologies are marked as abrupt boundary regions.

[0097] The continuous stable region and the abrupt boundary region are calibrated according to their temporal and spatial positions to form a temperature field partitioning correspondence; the stability maintenance requirements corresponding to the continuous stable region are extracted to determine the power maintenance direction corresponding to the continuous stable region; the boundary suppression requirements corresponding to the abrupt boundary region are extracted to determine the power adjustment direction corresponding to the abrupt boundary region.

[0098] The stability maintenance requirement and the boundary suppression requirement are mapped to the corresponding heating position and the corresponding time position, respectively. The power adjustment amplitude of each zone is sequentially integrated to form a zone adjustment relationship. Based on the zone adjustment relationship, the maintenance control content is output to the corresponding position of the continuous stable zone, and the compensation control content is output to the corresponding position of the abrupt boundary zone. The heating control command is generated by uniformly arranging the commands according to the temporal and spatial order.

[0099] S5.4: Adjust the heating power in the calibration furnace according to the heating control command, and generate a calibration temperature field control report based on the changes in temperature field zones before and after adjustment.

[0100] It should be noted that the corresponding heating position, corresponding time position, and corresponding power adjustment content in the heating control command are extracted, and the heating power of each heating position in the calibration furnace is adjusted to form the adjusted operating state. Based on the adjusted operating state, the temperature change results are collected, and the true temperature sequence, temperature confidence index, confidence temperature evolution matrix, coupled temperature field energy distribution map, and continuous stable region and abrupt boundary region are obtained in sequence according to the previous steps to form the adjusted temperature field partition.

[0101] By comparing the continuous stable region and abrupt boundary region obtained before adjustment with the continuous stable region and abrupt boundary region obtained after adjustment, the range change of the continuous stable region and the distribution change of the abrupt boundary region are extracted to form the temperature field zoning change.

[0102] The power adjustment content in the heating control command, the changes in temperature field zones before and after adjustment, and the corresponding changes are sequentially integrated to generate a verification temperature field control report.

[0103] Figure 6 The chart is a grouped bar chart, with the horizontal axis representing the stage type and the vertical axis representing the mean absolute error, showing the statistical comparison results of errors. The first group of bars, "Dynamic Stage - Average Apparent Temperature Error," calculates the absolute error between the apparent temperature sequence and the true temperature sequence within the step event segment, and averages this error across all sampling points in that segment to obtain the average apparent temperature error for the dynamic stage. The second group of bars, "Dynamic Stage - Average True Temperature Error," calculates the absolute error between the true temperature sequences within the same step event segment, and averages this error across all sampling points in that segment to obtain the average true temperature error for the dynamic stage. Figure 6 As can be seen, during the dynamic phase, the average value of the actual temperature error is lower than the average value of the apparent temperature error, with an error reduction rate of approximately 8% or more. In the steady-state phase, the difference between the two is small, and no error amplification occurs. Therefore, this invention effectively optimizes the dynamic phase while maintaining steady-state measurement accuracy, achieving a quantitative improvement in the accuracy of temperature measurement during the dynamic phase, and simultaneously enhancing the ability to determine the effectiveness of the temperature field verification.

[0104] In summary, this invention achieves structured analysis of the dynamic hysteresis behavior and intrinsic measurement errors of thermocouples by constructing temperature-related dynamic response parameters and establishing an error decomposition model; and by generating a true temperature sequence and temperature reliability index through dynamic compensation inversion, it upgrades the traditional verification method based on apparent temperature to a precise control method based on dynamic modeling and error inversion, thereby improving the accuracy of temperature measurement in the dynamic stage and the ability to determine the effectiveness of the verification temperature field.

[0105] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An automatic control method for a thermocouple calibration furnace based on temperature data acquisition, characterized in that, include: Thermocouple temperature data is collected and smoothed to generate apparent temperature series and time-varying characteristic data. By identifying the temperature-related dynamic response characteristics in apparent temperature series and time-varying feature data, temperature-related dynamic response parameters are generated. Based on the temperature-related dynamic response parameters, an error decomposition model is constructed using a dynamic inertial prediction layer, a residual structure separation layer, and a disturbance suppression layer. Frequency separation is performed on the apparent temperature sequence to obtain the intrinsic measurement error components. Error correction is performed on the apparent temperature sequence based on the intrinsic error component of the measurement, and dynamic compensation inversion is performed in combination with temperature-related dynamic response parameters to generate the true temperature sequence and temperature confidence index. The stability of the calibration temperature field is evaluated based on the actual temperature sequence and temperature reliability index. Heating control commands are generated, the heating power in the calibration furnace is adjusted, and a calibration temperature field control report is generated.

2. The automatic control method for thermocouple calibration furnace based on temperature data acquisition as described in claim 1, characterized in that: The thermocouple temperature data includes thermoelectric potential signal, timestamp, location identifier, and cold junction temperature measurement value.

3. The automatic control method for thermocouple calibration furnace based on temperature data acquisition as described in claim 2, characterized in that, The process of collecting thermocouple temperature data and performing smoothing compensation to generate apparent temperature sequences and time-varying characteristic data is as follows: Cold junction compensation is performed on the thermocouple temperature data, and calibration conversion is carried out to generate an apparent temperature sequence. The apparent temperature series is continuously mean smoothed and then subjected to a fixed time interval difference operation to generate time-varying feature data.

4. The automatic control method for thermocouple calibration furnace based on temperature data acquisition as described in claim 1, characterized in that, The step of generating temperature-related dynamic response parameters by identifying temperature-related dynamic response characteristics in apparent temperature sequences and time-varying feature data is as follows: Identify the starting and turning points of temperature changes in the apparent temperature sequence and divide the apparent temperature sequence into temperature response event segment sequences. Perform cross-delay estimation on the temperature response event segment sequence to generate a dynamic hysteresis feature set of the response segment; The dynamic hysteresis feature set of the response segment is encoded using hysteresis fingerprinting, and a cyclic consistency check is performed to establish a temperature mapping relationship, generating temperature-related dynamic response parameters.

5. The automatic control method for thermocouple calibration furnace based on temperature data acquisition as described in claim 4, characterized in that, The error decomposition model is constructed based on temperature-related dynamic response parameters using a dynamic inertial prediction layer and a residual structure separation layer. The specific steps are as follows: The dynamic inertial prediction layer reconstructs the response trajectory of the apparent temperature sequence based on temperature-related dynamic response parameters, and generates a dynamic inertial predicted temperature sequence by constructing a predicted temperature evolution path. The residual structure separation layer extracts the residual signal between the apparent temperature sequence and the dynamic inertial prediction temperature sequence, and performs trend decomposition to obtain the inertial residual trend component and fluctuation residual component. The dynamic inertial prediction layer and the residual structure separation layer are decomposed into an error decomposition model.

6. The automatic control method for thermocouple calibration furnace based on temperature data acquisition as described in claim 5, characterized in that, The intrinsic error component of the measurement refers to the random disturbance component obtained by performing frequency separation on the fluctuating residual component based on the error decomposition model, and then combining it with the inertial residual trend component for point-by-point superposition.

7. The automatic control method for thermocouple calibration furnace based on temperature data acquisition as described in claim 6, characterized in that, The aforementioned error correction of the apparent temperature sequence based on the intrinsic measurement error component refers to using an error cancellation mapping algorithm to subtract the intrinsic measurement error component from the apparent temperature sequence point by point, thereby generating an error-corrected temperature sequence.

8. The automatic control method for thermocouple calibration furnace based on temperature data acquisition as described in claim 7, characterized in that, The specific steps for generating the true temperature sequence and temperature confidence index are as follows: Based on temperature-related dynamic response parameters, the error-corrected temperature sequence is compensated for response time backtracking through a dynamic response inverse mapping algorithm to generate a true temperature sequence. The error residual ratio is calculated based on the actual temperature sequence and the measurement endogenous error components, generating error residual characteristics, which are then converted into normalized confidence indicators to generate the temperature confidence index.

9. The automatic control method for thermocouple calibration furnace based on temperature data acquisition as described in claim 1, characterized in that, The specific steps for verifying the temperature field stability of the actual temperature sequence and temperature confidence index, and generating heating control commands are as follows: A temporal spatial expansion matrix is ​​constructed based on the real temperature sequence and the temperature confidence index, and a confidence gating mapping algorithm is executed to filter the temporal confidence and generate a reliable temperature evolution matrix. Slowly varying energy components and rapidly perturbation energy components are extracted from the reliable temperature evolution matrix, and cross-mapping calculations are performed to generate a coupled temperature field energy distribution map. The energy field topology decomposition operation is performed on the coupled temperature field energy distribution map to identify continuous stable regions and abrupt boundary regions in the spatiotemporal dimension, and then partitioned control mapping is performed to generate heating control commands.

10. The automatic control method for thermocouple calibration furnace based on temperature data acquisition as described in claim 9, characterized in that, The calibration temperature field control report is generated by adjusting the heating power inside the calibration furnace according to the heating control command, and combining the changes in temperature field zones before and after the adjustment.

Images