An insulation shield monitoring system based on temperature detection

Abstract

The present application relates to the technical field of temperature measurement, in particular to an insulation shield monitoring system based on temperature detection, in the present application, the continuous temperature trace sample column is formed after the surface multi-measuring position temperature sequence of the insulation shield is time sequence number corrected, breakpoint removed and adjacent time temperature difference calculated, the continuous temperature section and the temperature difference section are taken as input to participate in temperature sequence modeling calculation, the support vector machine performs prediction operation on the same measuring position temperature evolution track and generates a reference temperature sequence, then the Kalman filter is used to perform state recursion on the continuous period deviation record and divide the stable section, the slow rising section and the high temperature section, the temperature state evolution link is constructed by counting the repetition number of the high temperature section and the interruption number in the fall, the abnormal temperature rise evolution trend can be identified under the conditions of load fluctuation, environmental temperature change and shield thermal inertia, the thermal abnormality identification capability of the connecting part is enhanced, and the positioning accuracy of the abnormal position is improved.

Description

Technical Field

[0001] This invention relates to the field of temperature measurement technology, and in particular to an insulating cover monitoring system based on temperature detection. Background Technology

[0002] The field of temperature measurement technology aims to acquire the temperature value of the target object, establish a temperature change calculation model, and determine the operating status of the object based on temperature parameters. By constructing a temperature detection device and executing a temperature calculation method, it realizes the quantitative analysis of the thermal state of equipment operation, and identifies the operating status of the equipment, determines the fault risk, and conducts an operational safety assessment based on temperature threshold determination, temperature change rate calculation, and temperature distribution calculation.

[0003] The purpose of a temperature-detection-based insulating cover monitoring system is to continuously detect the operating temperature status of the conductor connection points covered by the insulating cover, calculate temperature changes, and determine overheating anomalies. This system isolates energized conductors and prevents external contact. If increased contact resistance occurs at the conductor connection points, Joule heating is generated when current passes through the connection points. This heat is conducted through the conductor and the insulating cover, causing a temperature rise on the cover surface. The insulating cover includes an insulating body that covers the outside of the conductor connection points. The insulating body forms a stable heat conduction path from the conductor to the outer surface of the cover and introduces a time delay characteristic in the heat conduction process. Simultaneously, it restricts external airflow from directly affecting the conductor connection points, thus creating a controlled heat diffusion environment. By arranging multiple temperature detection nodes on the outer surface of the insulating cover to obtain temperature values, a temperature time series is constructed, and the temperature growth rate, temperature gradient, and temperature deviation are calculated to identify thermal anomalies at the conductor connection points.

[0004] Traditional temperature detection methods primarily identify equipment operating status through temperature threshold determination, temperature change rate calculation, and temperature distribution analysis. However, in actual operating environments, the surface temperature of the insulating cover is simultaneously affected by changes in current load, ambient temperature, and differences in heat conduction paths, resulting in periodic temperature fluctuations. As the heat conduction medium between the conductor connection and the external environment, the insulating cover's internal heat diffusion paths include conduction paths from the conductor to the insulation body and diffusion paths from the insulation body to the outer surface. These multiple paths are thermally coupled, leading to spatial correlation and time lag in temperature responses between different measurement points. Relying solely on instantaneous temperature values ​​and fixed thresholds can easily result in misjudgments. As the load gradually increases and the contact resistance gradually increases, the temperature rise rate may be slow and may not reach the threshold range. Early abnormal stages are difficult to identify. Temperature change rate calculation is usually performed on a single measuring point. There is a lack of joint judgment on the thermal diffusion correlation between multiple measuring points. The process of local heating at the connection point spreading to the surrounding measuring points is difficult to capture accurately. In actual operation, if the contact resistance of the wiring clamp gradually increases, the temperature in the local area gradually increases and is conducted to the adjacent protective area. It is difficult to distinguish between occasional temperature fluctuations and real thermal anomalies by judging the temperature at a single point. Under long-term operation, the abnormal part may continue to heat up without being identified in time, affecting the operational safety of the conductor connection and increasing the risk of equipment failure. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology and propose an insulation cover monitoring system based on temperature detection.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an insulation cover monitoring system based on temperature detection includes:

[0007] Temperature trace archiving module: Based on the temperature value sequence of the insulation cover surface measurement, the wiring clamp area, and the cover overlap area, the time sequence number is checked, the breakpoint record is removed, the temperature difference between adjacent time moments is calculated, and the time sequence temperature and temperature difference records are organized according to the partition number and measurement position number to obtain the temperature trace sample column.

[0008] Reference temperature track module: Based on the temperature track sample column, extract the continuous temperature value and temperature difference sequence of the same measurement position, use support vector machine to write the temperature record of the next time moment, generate a reference temperature sequence for each measurement position, and align the reference temperature sequence with the measured temperature sequence according to time to obtain the reference temperature track table.

[0009] Deviation Status Module: Based on the reference temperature trace table, calculate the difference between the measured temperature and the reference temperature of the junction box area and the cover overlap area, divide the continuous window into fixed durations, record the positive deviation length of the window, the peak dwell time, and the number of synchronous deviations of adjacent measurement positions to obtain the deviation behavior cluster;

[0010] Transition identification module: Based on the deviation behavior cluster, Kalman filtering is used to divide the stable segment, the gradual rise segment, and the high temperature segment. The sequence of state changes in the measurement cycle is recorded, the number of repetitions and the number of interruptions in the high temperature segment are counted, and the state segments are connected in series according to the cycle to obtain the thermal transition chain.

[0011] Thermal zone characterization module: Based on the thermal transition chain and calling the deviation behavior cluster, compare the linkage range and duration of the junction box area and the cover overlap area, merge the abnormal measurement number and the partition number of the same connection part, and obtain the thermal anomaly position of the cover.

[0012] As a further embodiment of the present invention, the temperature trace sample sequence includes a measurement location number sequence, a time sequence, a temperature measurement value sequence, and a temperature difference sequence between adjacent times; the reference temperature trace table includes a measurement location number set, a reference temperature sequence, a corresponding measured temperature sequence, and a time sequence; the deviation behavior cluster includes a deviation value sequence, a positive deviation duration sequence, a peak dwell time sequence, and a measurement location linkage quantity sequence; the thermal transition chain includes a temperature state sequence, a state jump record sequence, a cycle sequence, and a high-temperature state repetition record; and the protective cover thermal anomaly position includes a partition number, a measurement location number, an anomaly cycle sequence, and a linkage measurement location number sequence.

[0013] As a further aspect of the present invention, the temperature trace archiving module includes:

[0014] Measurement sequence submodule: Based on the temperature value sequence of the surface measurement of the insulating cover, the junction box area, and the cover overlap area, a corresponding relationship is established according to the measurement position number and the partition number. The temperature values ​​are arranged item by item according to the time sequence number. Continuous temperature records of the same measurement position are spliced ​​in sequence. Temperature records of adjacent measurement positions are archived side by side. Temperature records of the junction box area and the cover overlap area are written in partitions to obtain the measurement temperature column.

[0015] The time sequence processing submodule: Based on the measured temperature column, it checks the continuity of the time sequence number, deletes the time records with breakpoints, calculates the temperature difference between adjacent time times, writes the temperature difference value and the corresponding temperature record in parallel for each time time, and forms a continuous time sequence combination according to the partition number and the measurement position number to obtain the temperature trace sample column.

[0016] As a further aspect of the present invention, the reference temperature track module includes:

[0017] Sample Extraction Submodule: Based on the temperature trace sample column, extract continuous temperature values ​​and temperature difference sequences from the same measurement location, collect continuous time period records according to measurement location number, arrange temperature segments and temperature difference segments according to time sequence, verify the correspondence between partition number and measurement location number, and generate temperature difference sequence group;

[0018] Sequence mapping submodule: Based on the temperature difference sequence group, a support vector machine is used to write the temperature record of the next moment, match the temperature segment, temperature difference segment and the temperature value of the next moment according to the correspondence of the same measurement position, map the temperature results of each measurement position item by item in time order, organize the temperature sequence corresponding to the measurement position, and obtain the reference sequence group.

[0019] Table entry alignment submodule: Based on the reference sequence group, the reference temperature sequence and the measured temperature sequence are aligned by time, the correspondence between the time numbers of the same measurement position is checked, the reference temperature value and the measured temperature value are merged, and the corresponding table entries are organized according to the measurement position number, partition number and time sequence number to obtain the reference temperature trace table.

[0020] As a further aspect of the present invention, the support vector machine first constructs an input vector based on the continuous temperature segment and temperature difference segment of each measurement location in the temperature difference sequence group. It then combines the temperature values ​​and corresponding temperature difference values ​​of multiple consecutive time moments in chronological order to form training samples. The output of the training samples is set to the temperature record of the next time moment. The input vector is mapped to a high-dimensional feature space through a kernel function. A regression function is established in the high-dimensional feature space and a set of support vectors is calculated. The input vector is then regressed based on the set of support vectors and the parameters of the regression function to output the predicted temperature value at the corresponding time moment. The predicted temperature value is then written into the temperature sequence of the corresponding measurement location to form a reference sequence group.

[0021] As a further aspect of the present invention, the deviation situation module includes:

[0022] Temperature difference comparison submodule: Based on the reference temperature trace table, extract the measured temperature sequence of the junction box area and the reference temperature sequence, calculate the difference between the two at each time step and write it into the difference sequence. Perform the same difference calculation on the temperature sequence of the cover overlap area. The difference is arranged in order of measurement position number. The difference between adjacent measurement positions is archived side by side to obtain the temperature difference deviation column.

[0023] Window Deviation Submodule: Based on the temperature difference deviation column, the continuous window segment is divided according to a fixed duration, the number of positive deviation moments within the window is counted, the peak dwell time of the window is recorded, the synchronous deviation records of adjacent measurement positions are collected according to the window number, and the deviation records of each window are organized according to the measurement position order to obtain the deviation behavior cluster.

[0024] As a further aspect of the present invention, the transition identification module includes:

[0025] State segmentation submodule: Based on the deviation behavior cluster, Kalman filtering is used to divide the state into a stable segment, a gradual rise segment, and a high temperature segment. Continuous periodic deviation records are read according to the measurement position number. The deviation is assigned to the corresponding state segment according to the duration of positive deviation, the peak dwell time, and the number of linkages. The state distribution of the same measurement position is sorted along the time sequence to generate a state segment column.

[0026] The sequence recording submodule records the sequence of state changes in the measurement cycle based on the state segment column, counts the number of times the high temperature segment is repeated and the number of times the fall-off interruption is counted, writes the state transition order according to the connection relationship between the previous and subsequent cycles, merges the repeated high temperature records, fall-off interruption records and upgrade records of the same measurement position, and obtains the transition sequence book.

[0027] Link serialization submodule: Based on the transfer sequence book, it serializes state segments according to the period, connects adjacent periodic state sequences, merges continuous state paths of the same measurement position, collects multiple measurement position state evolution links, and organizes the link entries according to the measurement position number and periodic order to obtain the thermal transition chain.

[0028] As a further aspect of the present invention, the Kalman filter first establishes a state variable sequence based on the continuous periodic deviation records of each measurement position in the deviation behavior cluster. The state estimate of the previous period and the deviation record of the current period are input into the state recursive equation to obtain the predicted state value. The deviation record of the current period is used as the observation value and the predicted state value to calculate the difference to form the observation residual. The gain coefficient is calculated based on the observation residual and the covariance matrix. The predicted state value is corrected by the gain coefficient to obtain the updated state value. The updated state value is used as the recursive input for the next period. The smooth state sequence is obtained by continuous calculation along the periodic order.

[0029] As a further aspect of the present invention, the thermal zone characterization module includes:

[0030] Linked screening submodule: Based on the thermal transition chain and calling the deviation behavior cluster, extract the state sequence of the wiring clamp area and the state sequence of the shield overlap area, group the deviation records of adjacent test positions according to the connection position, count the number of continuous cycles item by item, and organize the test position numbers of the same connection position to obtain the linked test position group;

[0031] Anomaly location submodule: Based on the linked measurement group, the measurement number and the partition number are checked item by item, the abnormal measurement points of the same connection part are screened item by item, the continuous periodic records are written into the measurement sequence, and the abnormal measurement points are merged and organized according to the connection part to obtain the thermal anomaly points of the protective cover.

[0032] As a further aspect of the present invention, the abnormal measurement points are grouped and organized according to the connection parts. The connection part number table records the wiring clamp installation point number, the cover overlap area number, and the set of adjacent measurement point numbers. Based on the correspondence between the measurement point number and the connection part number, the measurement point numbers that have continuous abnormal cycle records within the same connection part number range are centrally classified. The measurement point numbers that have continuous abnormal cycles reaching a set number of cycles are written into the corresponding connection part number entry. Multiple measurement point numbers in the same connection part number entry are arranged in spatial layout order to form an abnormal measurement point set. The corresponding thermal abnormal position of the cover is output according to the connection part number.

[0033] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0034] In this invention, the temperature sequence from multiple measurement points on the surface of the insulating cover is processed by time sequence number verification, breakpoint removal, and temperature difference calculation between adjacent time points to form a continuous temperature track sample series. Continuous temperature segments and temperature difference segments serve as inputs for temperature sequence modeling and calculation. A support vector machine performs prediction calculations on the temperature evolution trajectory of the same measurement point and generates a reference temperature sequence, transforming the temperature judgment criterion from a single temperature value to a comparison of temperature evolution trajectories. The measured temperature sequence and the reference temperature sequence are aligned moment-by-moment, and the deviation difference is calculated. A fixed-duration window is used to statistically analyze the duration of positive deviation, the peak dwell time, and the number of synchronous deviations at adjacent measurement points, forming a set of deviation behaviors reflecting the characteristics of temperature change. Subsequently, Kalman filtering is used to analyze the continuous periodic deviation records. The system records state recursion and divides the data into stable, gradually rising, and high-temperature segments. By statistically analyzing the number of repetitions and interruptions in the high-temperature segment, a temperature state evolution link is constructed. Combined with the spatial distribution of the connection points, a merge analysis is performed on the multi-measurement linkage range and continuously. Before the temperature reaches the high threshold, abnormal temperature rise trends can be identified and abnormal areas in the connection points can be located. Temperature trajectory prediction, residual behavior statistics, and state transition analysis form a multi-level thermal state identification mechanism, transforming temperature monitoring from single-point numerical judgment to temperature evolution process analysis. It can identify abnormal temperature rise evolution trends under load fluctuations, ambient temperature changes, and the presence of thermal inertia of the protective cover, enhancing the ability to identify thermal anomalies in connection points and improving the accuracy of anomaly location. Attached Figure Description

[0035] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0037] Example 1

[0038] Please see Figure 1 The present invention provides a technical solution: an insulation cover monitoring system based on temperature detection includes:

[0039] Temperature trace archiving module: Based on the temperature value sequence of the insulation cover surface measurement, the wiring clamp area, and the cover overlap area, the time sequence number is checked, the breakpoint record is removed, the temperature difference between adjacent time moments is calculated, and the time sequence temperature and temperature difference records are organized according to the partition number and measurement position number to obtain the temperature trace sample column.

[0040] Reference temperature trace module: Based on the temperature trace sample column, extract the continuous temperature value and temperature difference sequence of the same measurement position, use support vector machine to write the temperature record of the next time moment, generate a reference temperature sequence for each measurement position, and align the reference temperature sequence with the measured temperature sequence according to time to obtain the reference temperature trace table.

[0041] Deviation Status Module: Based on the reference temperature trace table, calculate the difference between the measured temperature and the reference temperature of the junction box area and the cover overlap area, divide the continuous window into fixed durations, record the positive deviation length of the window, the peak dwell time, and the number of synchronous deviations of adjacent measurement positions to obtain the deviation behavior cluster;

[0042] Transition identification module: Based on deviation behavior clusters, Kalman filtering is used to divide the system into a stable segment, a gradual rise segment, and a high temperature segment. The sequence of state changes during the measurement cycle is recorded, the number of repetitions and interruptions during the high temperature segment is counted, and the state segments are connected in series according to the cycle to obtain the thermal transition chain.

[0043] Thermal zone characterization module: Based on the thermal transition chain and calling the deviation behavior cluster, compare the linkage range and duration of the junction box area and the cover overlap area, merge the abnormal measurement number and the partition number of the same connection part, and obtain the thermal anomaly position of the cover.

[0044] The temperature trace sample series includes a measurement location number sequence, a time sequence, a temperature measurement value sequence, and a temperature difference sequence between adjacent times. The reference temperature trace table includes a measurement location number set, a reference temperature sequence, a corresponding measured temperature sequence, and a time sequence. The deviation behavior cluster includes a deviation value sequence, a positive deviation duration sequence, a peak dwell time sequence, and a measurement location linkage quantity sequence. The thermal transition chain includes a temperature state sequence, a state jump record sequence, a cycle number sequence, and a high-temperature state repetition record. The protective cover thermal anomaly location includes a zone number, a measurement location number, an anomaly cycle sequence, and a linkage measurement location number sequence.

[0045] The temperature tracking module includes:

[0046] Measurement sequence submodule: Based on the temperature value sequence of the surface measurement of the insulating cover, the junction box area, and the cover overlap area, a corresponding relationship is established according to the measurement position number and the partition number. The temperature values ​​are arranged item by item according to the time sequence number. Continuous temperature records of the same measurement position are spliced ​​in sequence. Temperature records of adjacent measurement positions are archived side by side. Temperature records of the junction box area and the cover overlap area are written in partitions to obtain the measurement temperature column.

[0047] Time series processing submodule: Based on the measured temperature column, it checks the continuity of time sequence numbers, deletes time records with breakpoints, calculates the temperature difference between adjacent time times, writes the temperature difference value and the corresponding temperature record in parallel time time, and forms a continuous time series combination according to the partition number and the measured position number to obtain the temperature trace sample column.

[0048] Measurement Sequence Submodule: Based on the temperature value sequences of the insulation cover surface measurement, junction box area, and cover overlap area, a partition mapping table is used to establish a correspondence between the measurement position number and the partition number. A merge sorting algorithm is used to arrange the temperature value sequence item by item according to the time sequence number. The partition mapping table is written with partition numbers 01 to 16, measurement position numbers M001 to M128, and area identifiers A1 and A2. The merge sorting uses the partition number as the first sorting key, the measurement position number as the second sorting key, and the time sequence number as the third sorting key. The sampling interval for the time sequence number is set to 60 seconds. The condition for splicing continuous temperature records of the same measurement position is that the difference between adjacent time sequence numbers is equal to 60. The condition for archiving adjacent measurement positions is that the difference between measurement position numbers within the same partition is equal to 1. The temperature record of the junction box area is written to the area identifier A1, and the temperature record of the cover overlap area is written to the area identifier A2. Records of the same partition, the same measurement position, and the same time sequence are sequentially merged into the measurement position temperature column to obtain the measurement position temperature column.

[0049] The time-series processing submodule: Based on the measured temperature column, a dual-pointer time-series verification method is used to check the continuity of time sequence numbers, and a first-order difference calculation method is used to calculate the temperature difference between adjacent time points. The two pointers point to the previous and next records of the same measurement position, respectively. The continuity judgment condition is set to the difference between the next time sequence number and the previous time sequence number equals 60, and the breakpoint judgment condition is set to the difference being greater than 120. Records that meet the breakpoint judgment condition are directly deleted. The temperature difference is calculated by subtracting the previous temperature value from the next temperature value. The difference step size is set to 1, and the number of decimal places retained for the temperature difference is set to 2 decimal places. The temperature difference value and the corresponding temperature record are written side by side according to the same measurement position number, the same partition number, and the same time sequence number, and a time sequence combination is formed according to the continuous time sequence number from small to large to generate a temperature trace sample column.

[0050] The reference temperature trace module includes:

[0051] Sample Extraction Submodule: Based on the temperature trace sample column, extract continuous temperature values ​​and temperature difference sequences from the same measurement location, collect continuous time period records according to measurement location number, arrange temperature segments and temperature difference segments according to time sequence, verify the correspondence between the partition number and the measurement location number, and generate temperature difference sequence groups;

[0052] The sequence mapping submodule is based on the temperature difference sequence group and uses a support vector machine to write the temperature record of the next time moment. It matches the temperature segment, temperature difference segment and temperature value of the next time moment according to the correspondence of the same measurement position. It maps the temperature results of each measurement position item by item in time sequence, organizes the temperature sequence corresponding to the measurement position, and obtains the reference sequence group.

[0053] The table entry alignment submodule: Based on the reference sequence group, the reference temperature sequence and the measured temperature sequence are aligned by time, the correspondence between the time numbers of the same measurement position is checked, the reference temperature value and the measured temperature value are merged, and the corresponding table entries are organized by measurement position number, partition number and time sequence number to obtain the reference temperature trace table;

[0054] Sample Extraction Submodule: Based on the temperature trace sample column, a fixed-length sliding window segmentation method is used to extract continuous temperature values ​​and temperature difference sequences at the same measurement location. The sliding window length is set to 6 time points, the step size is set to 1 time point, the measurement location number range is set to M001 to M128, and the partition number range is set to 01 to 16. Six consecutive temperature values ​​at the same measurement location are written into the temperature segment according to the time sequence number from smallest to largest. The corresponding six adjacent temperature difference values ​​are written into the temperature difference segment according to the same sequence number. The partition number and measurement location number are compared using the number equivalence verification rule. If both fields are equal, the current window record is retained. If either field is unequal, the current window record is deleted. The temperature segment and temperature difference segment are arranged according to the time sequence to generate a temperature difference sequence group.

[0055] The sequence mapping submodule uses a support vector machine (SVM) regression algorithm based on temperature difference sequence groups to write the temperature record of the next time step. The SVM type is set to ε regression, the kernel function is set to radial basis function, the penalty coefficient C is set to 32, the kernel parameter γ is set to 0.125, and the loss parameter ε is set to 0.01. The input is a 12-dimensional vector formed by concatenating 6 consecutive temperature values ​​and 6 consecutive temperature difference values ​​at the same measurement point. The output is the temperature value at the 7th time step. The 12-dimensional vector is read item by item according to the measurement point number and fed into the regression equation. The kernel matrix, Lagrange multiplier, and bias terms are calculated, and the corresponding predicted temperature value is output. The predicted temperature values ​​are mapped to the temperature results of each measurement point in chronological order. The corresponding temperature sequence of the measurement points is organized according to the correspondence of the same measurement point to obtain the reference sequence group.

[0056] The table entry alignment submodule: Based on the reference sequence group, a dual-index time-series alignment method is used to align the reference temperature sequence and the measured temperature sequence by time. The reference sequence index fields are set as measurement site number, partition number, and time sequence number, and the measured sequence index fields are set to the same three fields. The correspondence between the time numbers of the same measurement site is checked using the three-field equality matching rule. When the measurement site number, partition number, and time sequence number are all equal, they are written into the same table entry. When any field is not equal, it is shifted one position to the right along the smaller time sequence number to complete the merging of the reference temperature value and the measured temperature value. The corresponding table entries are then sorted in ascending order by measurement site number, partition number, and time sequence number to obtain the reference temperature trace table.

[0057] Support Vector Machine (SVM) first constructs an input vector based on the continuous temperature segments and temperature difference segments of each measurement location in the temperature difference sequence group. It then combines the temperature values ​​and corresponding temperature difference values ​​of multiple consecutive time moments in chronological order to form training samples. The output of the training samples is set as the temperature record of the next time moment. The input vector is mapped to a high-dimensional feature space through a kernel function. A regression function is established in the high-dimensional feature space and a set of support vectors is calculated. The input vector is then regressed based on the set of support vectors and the parameters of the regression function to output the predicted temperature value at the corresponding time moment. The predicted temperature value is then written into the temperature sequence of the corresponding measurement location to form a reference sequence group.

[0058] The deviation from the situation module includes:

[0059] Temperature difference comparison submodule: Based on the reference temperature trace table, extract the measured temperature sequence of the junction box area and the reference temperature sequence, calculate the difference between the two at each time step and write it into the difference sequence. Perform the same difference calculation on the temperature sequence of the cover overlap area. The difference is arranged in the order of the measurement position number. The difference between adjacent measurement positions is archived side by side to obtain the temperature difference deviation column.

[0060] Window Deviation Submodule: Based on the temperature difference deviation column, continuous window segments are divided according to a fixed duration. The number of positive deviation moments within the window is counted, the peak dwell time of the window is recorded, and synchronous deviation records of adjacent measurement positions are collected by window number. The deviation records of each window are organized according to the measurement position order to obtain a cluster of deviation behaviors.

[0061] Temperature difference comparison submodule: Based on the reference temperature trace table, the measured temperature sequence and reference temperature sequence of the junction box area are extracted using the time-series difference calculation method. The temperature value is read and compared with the corresponding position at each time step. The read fields are set as measurement position number, zone number, time sequence number, measured temperature value, and reference temperature value. The measured temperature value minus the reference temperature value is written into the difference record and the difference sequence is generated. At the same time, the same reading method and difference writing operation are performed on the temperature sequence of the cover overlap area, and the difference record is filled in. The difference is arranged in ascending order of measurement position number, and adjacent measurement position number differences of 1 are archived side by side. The difference records of the junction box area and the difference records of the cover overlap area are grouped and organized according to the same numbering structure to generate a temperature difference deviation column.

[0062] Window Deviation Submodule: Based on the temperature difference deviation column, a fixed time-series window statistical method is used to divide continuous difference records into window segments. The window length is set to 5 consecutive time intervals and the window movement step size is set to 1 time interval. Continuous difference records of the same measurement position are read window by window, and records with greater than zero difference are counted. The count results are written to the window deviation quantity field. At the same time, all difference records within the window are read and the difference is identified. The number of consecutive occurrences of the difference is recorded as the peak dwell time. At the same time, difference records of adjacent measurement positions at the same time are read and greater than zero synchronization records are determined. The number of synchronization deviations is accumulated and collected by window number and sorted by both window number and measurement position number. The deviation quantity records, peak dwell time records and synchronization deviation records within the window are merged and organized to generate deviation behavior clusters.

[0063] The transition identification module includes:

[0064] State segmentation submodule: Based on deviation behavior clusters, Kalman filtering is used to divide the state into stable segment, gradual rise segment, and high temperature segment. Continuous periodic deviation records are read according to the measurement position number. The positive deviation duration, peak dwell time and linkage quantity are assigned to the corresponding state segment. The state distribution of the same measurement position is sorted along the time sequence to generate a state segment column.

[0065] The sequence recording submodule records the sequence of state changes in the measurement cycle based on the state segment column, counts the number of times the high temperature segment is repeated and the number of times the fall-off interruption is counted, writes the state transition order according to the connection relationship between the previous and subsequent cycles, merges the repeated high temperature records, fall-off interruption records and upgrade records of the same measurement position, and obtains the transition sequence book.

[0066] Link serialization submodule: Based on the transfer sequence book, it serializes state segments according to the period, connects adjacent periodic state sequences, merges continuous state paths of the same measurement position, collects multiple measurement position state evolution links, and organizes the link entries according to the measurement position number and periodic order to obtain the thermal transition chain.

[0067] State segmentation submodule: Based on deviation behavior clusters, Kalman filtering is used to recursively analyze the state of continuous periodic deviation records. It reads four records: measurement location number, cycle number, positive deviation duration, peak dwell time, and number of linkages. The state variable is set as a three-dimensional temperature rise state vector. The initial state value is taken from the first periodic deviation record. The diagonal values ​​of the initial covariance matrix, the main diagonal values ​​of the state transition matrix, and the main diagonal values ​​of the observation matrix are all set to 1. The process noise covariance is set to 0.01, and the observation noise covariance is set to 0.05. Predicted value writing, observed value reading, gain update, and state correction are performed sequentially to obtain a smooth deviation sequence. Threshold segmentation is then performed: records with a positive deviation duration of less than 2 cycles and a peak dwell time of less than 2 cycles are classified into the stable segment; records with a positive deviation duration of 2 to 4 cycles and a number of linkages of less than 3 cycles are classified into the gradual rise segment; and records with a positive deviation duration of more than 4 cycles and a peak dwell time of more than 3 cycles are classified into the high temperature segment. The state distribution of the same measurement location is organized chronologically to generate a state segment column.

[0068] The sequence recording submodule, based on the state segment column, uses a first-order Markov state transition statistics method to record the sequence of state changes in the measurement cycle. It performs state pairing of adjacent cycles, accumulation of state transition counts, merging of repeated high-temperature records, merging of fallback interruption records and writing them into the order-up record. The state set is set as three items: stable segment, gradual rise segment, and high-temperature segment. The transition statistics window is set as two adjacent cycles. The state of the previous cycle and the state of the next cycle are read cycle by cycle according to the same measurement position. Multiple transitions from stable segment to gradual rise segment, from gradual rise segment to high-temperature segment, from high-temperature segment to high-temperature segment, and from high-temperature segment to gradual rise segment are written into the state transition table. Two or more consecutive high-temperature segments are written into the repeated high-temperature record. Records of high-temperature segments that do not have a gradual rise segment after the high-temperature segment but directly remain in the high-temperature segment are written into the fallback interruption record. Records of stable segments directly entering the high-temperature segment are written into the order-up record. The transition sequence is merged and arranged according to the measurement position number and cycle number to obtain the transition sequence book.

[0069] Link Serialization Submodule: Based on the transition sequence book, the state segments are serialized in series according to the period using the state path serialization method. It performs sequential connection of adjacent period states, merging of continuous state paths at the same measurement position, collection of state evolution links at multiple measurement positions, and organization of link entries. The link start point is set as the first non-stationary segment period, and the link end point is set as the period in which two consecutive stationary segments appear. All state transition records in the transition sequence book are read according to the measurement position number. Records with a difference of one period between the end period of the previous record and the start period of the next record are connected end to end. Records of continuous stationary segments to gradual rise segments to high temperature segments at the same measurement position are merged into a single state path. Multiple measurement position state paths at the same connection point are grouped into the same link set according to the partition number. The link number, measurement position number, start period, end period, and state sequence are written into the link entry one by one, and the link entries are organized according to the measurement position number and period order to obtain the thermal transition chain.

[0070] Kalman filtering first establishes a state variable sequence based on the continuous periodic deviation records of each measurement position in the deviation behavior cluster. The state estimate of the previous period and the deviation record of the current period are input into the state recursive equation to obtain the predicted state value. The deviation record of the current period is used as the observation value and the difference between the observed value and the predicted state value to form the observation residual. The gain coefficient is calculated based on the observation residual and the covariance matrix. The predicted state value is corrected by the gain coefficient to obtain the updated state value. The updated state value is used as the recursive input for the next period. The smooth state sequence is obtained by continuous calculation along the periodic order.

[0071] The hot zone qualitative module includes:

[0072] Linked screening submodule: Based on thermal transition chain and calling deviation behavior cluster, extract the state sequence of the wiring clamp area and the state sequence of the shield overlap area, group the deviation records of adjacent test positions according to the connection part, count the number of continuous cycles item by item, and organize the test position numbers of the same connection part to obtain the linked test position group;

[0073] Anomaly location submodule: Based on the linkage measurement group, the measurement number and the partition number are checked one by one, the abnormal measurement points of the same connection part are filtered one by one, the continuous periodic record is written into the measurement sequence, and the abnormal measurement points are merged and organized according to the connection part to obtain the thermal anomaly points of the protective cover.

[0074] The linkage screening submodule, based on the thermal transition chain and calling the deviation behavior cluster, uses the connection part adjacency matrix grouping method to extract the state sequence of the wiring clamp area and the state sequence of the shield overlap area. It then reads the connection part number table, loads the measurement position adjacency relationship, partitions and splits the state sequence, and synchronizes and registers the deviation records. The connection part number range is set to J01 to J32, the partition number range is set to 01 to 16, and the row and column items of the adjacency matrix are set to the measurement position number M001 to M128. The matrix value 1 indicates adjacent, and the matrix value 0 indicates non-adjacent. Adjacent measurement position deviation records are read according to the same connection part number. Records with consecutive high temperature segments or slow rise segments and window synchronous deviation times greater than or equal to 3 are written into the grouping table. The continuous cycle counting method is used to perform cycle-by-cycle accumulation for each measurement position state sequence. Records with consecutive state cycle counts greater than or equal to 4 cycles are written into the continuous cycle field. Finally, the measurement position number, continuous cycle count, state segment order, and synchronous deviation count of the same group are summarized according to the connection part number to complete the set of measurement position numbers of the same connection part and generate the linkage measurement position group.

[0075] The anomaly location submodule, based on the linkage measurement group, uses a partition verification and filtering method to check the measurement number and partition number item by item. It also performs screening of abnormal measurement locations at connection points, writing of continuous cycle records, merging and organizing connection points, and outputting anomaly location numbers. The verification fields are set as measurement number, partition number, connection point number, number of continuous cycles, number of synchronous deviations, and number of high-temperature segment repetitions. The filtering conditions are set as the number of continuous cycles greater than or equal to 4, the number of synchronous deviations greater than or equal to 3, and the number of high-temperature segment repetitions greater than or equal to 2. Records that meet all conditions are written into the anomaly measurement table, and records that do not meet all conditions are deleted. Then, the anomaly measurement table is merged according to the connection point number using the connection point merging rule. Multiple anomaly measurement locations under the same connection point number are arranged in spatial layout order. The continuous cycle records are written into the measurement sequence. The partition number, measurement number, connection point number, start cycle, and end cycle are written into the location table item by item to generate the thermal anomaly location of the protective cover.

[0076] Abnormal measurement points are grouped and organized according to connection locations. The connection location number table records the wiring clamp installation point number, the cover overlap area number, and the set of adjacent measurement point numbers. Based on the correspondence between measurement point numbers and connection location numbers, measurement point numbers with consecutive abnormal cycles within the same connection location number range are centrally categorized. Measurement point numbers with consecutive abnormal cycles reaching a set number of cycles are written into the corresponding connection location number entry. Multiple measurement point numbers in the same connection location number entry are arranged according to spatial layout order to form an abnormal measurement point set. The corresponding thermal abnormality location of the cover is output according to the connection location number.

[0077] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An insulating cover monitoring system based on temperature detection, characterized in that, The system includes: Temperature trace archiving module: Based on the temperature value sequence of the insulation cover surface measurement, the wiring clamp area, and the cover overlap area, the time sequence number is checked, the breakpoint record is removed, the temperature difference between adjacent time moments is calculated, and the time sequence temperature and temperature difference records are organized according to the partition number and measurement position number to obtain the temperature trace sample column. Reference temperature track module: Based on the temperature track sample column, extract the continuous temperature value and temperature difference sequence of the same measurement position, use support vector machine to write the temperature record of the next time moment, generate a reference temperature sequence for each measurement position, and align the reference temperature sequence with the measured temperature sequence according to time to obtain the reference temperature track table. Deviation Status Module: Based on the reference temperature trace table, calculate the difference between the measured temperature and the reference temperature of the junction box area and the cover overlap area, divide the continuous window into fixed durations, record the positive deviation length of the window, the peak dwell time, and the number of synchronous deviations of adjacent measurement positions to obtain the deviation behavior cluster; Transition identification module: Based on the deviation behavior cluster, Kalman filtering is used to divide the stable segment, the gradual rise segment, and the high temperature segment. The sequence of state changes in the measurement cycle is recorded, the number of repetitions and the number of interruptions in the high temperature segment are counted, and the state segments are connected in series according to the cycle to obtain the thermal transition chain. The thermal zone qualitative module: Based on the thermal transition chain and calling the deviation behavior cluster, it compares the linkage range and duration of the junction box area and the cover overlap area, merges the abnormal measurement number and the partition number of the same connection part, and obtains the thermal anomaly position of the cover. The thermal zone qualitative module includes: Linked screening submodule: Based on the thermal transition chain and calling the deviation behavior cluster, extract the state sequence of the wiring clamp area and the state sequence of the shield overlap area, group the deviation records of adjacent test positions according to the connection position, count the number of continuous cycles item by item, and organize the test position numbers of the same connection position to obtain the linked test position group; Anomaly location submodule: Based on the linkage measurement group, the measurement number and the partition number are checked one by one, the abnormal measurement points of the same connection part are filtered one by one, the continuous periodic records are written into the measurement sequence, and the abnormal measurement points are merged and sorted according to the connection part to obtain the thermal anomaly points of the protective cover. The abnormal measurement points are grouped and organized according to the connection parts. The connection part number table records the wiring clamp installation point number, the cover overlap area number, and the set of adjacent measurement point numbers. Based on the correspondence between the measurement point number and the connection part number, the measurement point numbers that have continuous abnormal cycle records within the same connection part number range are centrally classified. The measurement point numbers that have continuous abnormal cycles reaching a set number of cycles are written into the corresponding connection part number entry. Multiple measurement point numbers in the same connection part number entry are arranged in spatial layout order to form an abnormal measurement point set. The corresponding thermal abnormal position of the cover is output according to the connection part number.

2. The insulation cover monitoring system based on temperature detection according to claim 1, characterized in that, The temperature trace sample column includes a measurement location number sequence, a time sequence, a temperature measurement value sequence, and a temperature difference sequence between adjacent times. The reference temperature trace table includes a measurement location number set, a reference temperature sequence, a corresponding measured temperature sequence, and a time sequence. The deviation behavior cluster includes a deviation value sequence, a positive deviation duration sequence, a peak dwell time sequence, and a measurement location linkage quantity sequence. The thermal transition chain includes a temperature state sequence, a state jump record sequence, a cycle number sequence, and a high-temperature state repetition record. The protective cover thermal anomaly position includes a partition number, a measurement location number, an anomaly cycle sequence, and a linkage measurement location number sequence.

3. The insulation cover monitoring system based on temperature detection according to claim 1, characterized in that, The temperature tracking module includes: Measurement sequence submodule: Based on the temperature value sequence of the surface measurement of the insulating cover, the junction box area, and the cover overlap area, a corresponding relationship is established according to the measurement position number and the partition number. The temperature values ​​are arranged item by item according to the time sequence number. Continuous temperature records of the same measurement position are spliced ​​in sequence. Temperature records of adjacent measurement positions are archived side by side. Temperature records of the junction box area and the cover overlap area are written in partitions to obtain the measurement temperature column. The time sequence processing submodule: Based on the measured temperature column, it checks the continuity of the time sequence number, deletes the time records with breakpoints, calculates the temperature difference between adjacent time times, writes the temperature difference value and the corresponding temperature record in parallel for each time time, and forms a continuous time sequence combination according to the partition number and the measurement position number to obtain the temperature trace sample column.

4. The insulation cover monitoring system based on temperature detection according to claim 1, characterized in that, The reference temperature track module includes: Sample Extraction Submodule: Based on the temperature trace sample column, extract continuous temperature values ​​and temperature difference sequences from the same measurement location, collect continuous time period records according to measurement location number, arrange temperature segments and temperature difference segments according to time sequence, verify the correspondence between partition number and measurement location number, and generate temperature difference sequence group; Sequence mapping submodule: Based on the temperature difference sequence group, a support vector machine is used to write the temperature record of the next moment, match the temperature segment, temperature difference segment and the temperature value of the next moment according to the correspondence of the same measurement position, map the temperature results of each measurement position item by item in time order, organize the temperature sequence corresponding to the measurement position, and obtain the reference sequence group. Table entry alignment submodule: Based on the reference sequence group, the reference temperature sequence and the measured temperature sequence are aligned by time, the correspondence between the time numbers of the same measurement position is checked, the reference temperature value and the measured temperature value are merged, and the corresponding table entries are organized according to the measurement position number, partition number and time sequence number to obtain the reference temperature trace table.

5. The insulation cover monitoring system based on temperature detection according to claim 4, characterized in that, The support vector machine first constructs an input vector based on the continuous temperature segment and temperature difference segment of each measurement location in the temperature difference sequence group. It then combines the temperature values ​​and corresponding temperature difference values ​​of multiple consecutive time moments in chronological order to form training samples. The output of the training samples is set to the temperature record of the next time moment. The input vector is mapped to a high-dimensional feature space through a kernel function. A regression function is established in the high-dimensional feature space and a set of support vectors is calculated. The input vector is then regressed based on the set of support vectors and the parameters of the regression function to output the predicted temperature value at the corresponding time moment. The predicted temperature value is then written into the temperature sequence of the corresponding measurement location to form a reference sequence group.

6. The insulation cover monitoring system based on temperature detection according to claim 1, characterized in that, The deviation status module includes: Temperature difference comparison submodule: Based on the reference temperature trace table, extract the measured temperature sequence of the junction box area and the reference temperature sequence, calculate the difference between the two at each time step and write it into the difference sequence. Perform the same difference calculation on the temperature sequence of the cover overlap area. The difference is arranged in order of measurement position number. The difference between adjacent measurement positions is archived side by side to obtain the temperature difference deviation column. Window Deviation Submodule: Based on the temperature difference deviation column, the continuous window segment is divided according to a fixed duration, the number of positive deviation moments within the window is counted, the peak dwell time of the window is recorded, the synchronous deviation records of adjacent measurement positions are collected according to the window number, and the deviation records of each window are organized according to the measurement position order to obtain the deviation behavior cluster.

7. The insulation cover monitoring system based on temperature detection according to claim 1, characterized in that, The transition identification module includes: State segmentation submodule: Based on the deviation behavior cluster, Kalman filtering is used to divide the state into a stable segment, a gradual rise segment, and a high temperature segment. Continuous periodic deviation records are read according to the measurement position number. The deviation is assigned to the corresponding state segment according to the duration of positive deviation, the peak dwell time, and the number of linkages. The state distribution of the same measurement position is sorted along the time sequence to generate a state segment column. The sequence recording submodule records the sequence of state changes in the measurement cycle based on the state segment column, counts the number of times the high temperature segment is repeated and the number of times the fall-off interruption is counted, writes the state transition order according to the connection relationship between the previous and subsequent cycles, merges the repeated high temperature records, fall-off interruption records and upgrade records of the same measurement position, and obtains the transition sequence book. Link serialization submodule: Based on the transfer sequence book, it serializes state segments according to the period, connects adjacent periodic state sequences, merges continuous state paths of the same measurement position, collects multiple measurement position state evolution links, and organizes the link entries according to the measurement position number and periodic order to obtain the thermal transition chain.

8. The insulation cover monitoring system based on temperature detection according to claim 7, characterized in that, The Kalman filter first establishes a state variable sequence based on the continuous periodic deviation records of each measurement position in the deviation behavior cluster. The state estimate of the previous period and the deviation record of the current period are input into the state recursive equation to obtain the predicted state value. The deviation record of the current period is used as the observation value and the predicted state value to calculate the difference to form the observation residual. The gain coefficient is calculated based on the observation residual and the covariance matrix. The predicted state value is corrected by the gain coefficient to obtain the updated state value. The updated state value is used as the recursive input for the next period. The smooth state sequence is obtained by continuous calculation along the periodic order.

Images