A switch cabinet temperature rise early warning method and system

Abstract

The application discloses a switch cabinet temperature rise early warning method and system, which is used to solve the technical problem that the existing switch cabinet temperature rise early warning method is mainly based on a mathematical model established according to standard working conditions for prediction, resulting in poor engineering adaptability of the model. The method comprises the following steps: preprocessing an obtained historical annual switch cabinet sample data set, and outputting a historical annual normal sample data set and a historical annual abnormal sample data set; based on preset evaluation criteria, an initial temperature rise early warning model is constructed according to the historical annual normal sample data set and the historical annual abnormal sample data set; when a current annual switch cabinet sample data set is received, a preset temperature error criterion and the current annual switch cabinet sample data set are used to correct a weight matrix of the initial temperature rise early warning model, and a target temperature rise early warning model is output; and a preset early warning criterion is used to generate a switch cabinet temperature rise early warning result according to a weight matrix of the target temperature rise early warning model and a weight matrix of the initial temperature rise early warning model.

Description

Technical Field

[0001] This invention relates to the field of power equipment technology, and in particular to a method and system for early warning of temperature rise in switchgear. Background Technology

[0002] Switchgear is a crucial type of equipment in power distribution systems, and its safe and reliable operation has become a key indicator for evaluating the level of power distribution automation. A survey of the current operational status and potential safety hazards of existing switchgear reveals that temperature rise and overheating are among the most common types of faults and potential hazards in switchgear.

[0003] Excessive temperature in switchgear can lead to problems such as insulation material aging and thermal expansion, thereby affecting the safe and stable operation of the power system. To ensure the safe operation of the power system, real-time monitoring and prediction of switchgear temperature rise are necessary, with overheating faults particularly prevalent in the switchgear busbar compartment. Therefore, developing more efficient, accurate, and reliable methods for predicting switchgear busbar temperature rise has become a hot research topic in the current power system field.

[0004] Existing switchgear temperature rise early warning methods are mainly based on mathematical models established under standard operating conditions. However, engineering projects need to consider the impact of random disturbances such as service life, aging, and equipment usage. Mathematical models based on standard operating conditions only consider parameter relationships under ideal conditions, while random disturbances such as extended service life, material aging and deterioration, and equipment wear and tear that must be faced in actual engineering are not effectively incorporated into the model, resulting in poor engineering adaptability of the model. Summary of the Invention

[0005] This invention provides a method and system for early warning of temperature rise in switchgear, which solves the technical problem that existing methods for early warning of temperature rise in switchgear are mainly based on mathematical models established under standard operating conditions, resulting in poor engineering adaptability of the models.

[0006] The first aspect of this invention provides a method for early warning of temperature rise in switchgear, comprising:

[0007] Obtain historical switchgear sample datasets, preprocess the historical switchgear sample datasets, and output historical normal sample datasets and historical abnormal sample datasets.

[0008] Based on the pre-set evaluation criteria, an initial temperature rise early warning model is constructed according to the historical annual normal sample dataset and the historical annual abnormal sample dataset.

[0009] When the current year's switchgear sample dataset is received, the initial temperature rise warning model is corrected by weight matrix using a preset temperature error criterion and the current year's switchgear sample dataset, and the target temperature rise warning model is output.

[0010] The switchgear temperature rise warning result is generated based on the weight matrix of the target temperature rise warning model and the weight matrix of the initial temperature rise warning model using a pre-set warning criterion.

[0011] Optionally, the preprocessing of the historical year switchgear sample dataset to output a historical year normal sample dataset and a historical year abnormal sample dataset includes:

[0012] Based on multiple switch cabinet feature operation parameters and preset reference thresholds in the historical switch cabinet sample dataset, calculate multiple per-unit values ​​corresponding to each switch cabinet feature operation parameter in the historical switch cabinet sample dataset.

[0013] The per-unit values ​​corresponding to the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset are compared with a preset per-unit threshold to generate the comparison results corresponding to the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset.

[0014] Based on the comparison results of the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset, the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset are classified to generate a historical normal sample dataset and a historical abnormal sample dataset.

[0015] Optionally, the initial temperature rise warning model includes an initial normal sample temperature rise warning model and an initial abnormal sample temperature rise warning model; the step of constructing the initial temperature rise warning model based on preset evaluation criteria and according to the historical annual normal sample dataset and the historical annual abnormal sample dataset includes:

[0016] The historical annual normal sample dataset is divided into a first normal sample dataset and a second normal sample dataset;

[0017] Based on the first normal sample dataset and the second normal sample dataset, an initial normal sample temperature rise early warning model is constructed;

[0018] Based on the pre-set evaluation criteria, the initial normal sample temperature rise early warning model is corrected by adjusting the weight matrix using the historical annual abnormal sample dataset, and the initial abnormal sample temperature rise early warning model is determined.

[0019] Optionally, the step of constructing an initial normal sample temperature rise early warning model based on the first normal sample dataset and the second normal sample dataset includes:

[0020] Based on the operating parameters of multiple switchgear features in the first normal sample dataset, a first intermediate temperature rise early warning model is constructed.

[0021] The first intermediate temperature rise early warning model is used to output a first predicted temperature value based on the second normal sample dataset;

[0022] Calculate the first error value based on the first predicted temperature value;

[0023] The weight matrix of the first intermediate temperature rise early warning model is updated based on the first error value to determine the second intermediate temperature rise early warning model, and the number of model updates is counted in real time.

[0024] Determine whether the number of model updates has reached a preset threshold;

[0025] If so, the second intermediate temperature rise early warning model shall be used as the initial normal sample temperature rise early warning model.

[0026] Optionally, the step of adjusting the weight matrix of the initial normal sample temperature rise early warning model based on the preset evaluation criteria and using the historical annual abnormal sample dataset to determine the initial abnormal sample temperature rise early warning model includes:

[0027] The initial normal sample temperature rise early warning model is corrected by single parameter perturbation using multiple switch cabinet feature operation parameters in the historical annual abnormal sample dataset to determine the third intermediate temperature rise early warning model.

[0028] The third intermediate temperature rise early warning model is used to output multiple second predicted temperature values ​​based on multiple switch cabinet characteristic operating parameters in the historical annual abnormal sample dataset.

[0029] Based on each of the second predicted temperature values, a plurality of second error values ​​are calculated;

[0030] Determine whether multiple second error values ​​satisfy the preset evaluation criteria;

[0031] If so, the third intermediate temperature rise early warning model shall be used as the initial abnormal sample temperature rise early warning model.

[0032] Optionally, the target temperature rise early warning model includes a target normal sample temperature rise early warning model and a target abnormal sample temperature rise early warning model; the preset temperature error criterion includes a first error criterion and a second error criterion; the step of using the preset temperature error criterion and the current year's switchgear sample dataset to perform weight matrix correction on the initial temperature rise early warning model and output the target temperature rise early warning model includes:

[0033] The current year's switchgear sample dataset is preprocessed to output the current year's normal sample dataset and the current year's abnormal sample dataset;

[0034] Based on the first error criterion, the weight matrix of the initial normal sample temperature rise early warning model is corrected using the current year's normal sample dataset to determine the target normal sample temperature rise early warning model.

[0035] Based on the second error criterion, the weight matrix of the initial abnormal sample temperature rise early warning model is corrected using the current year's abnormal sample dataset to determine the target abnormal sample temperature rise early warning model.

[0036] Optionally, the step of correcting the weight matrix of the initial normal sample temperature rise early warning model based on the first error criterion and using the current year's normal sample dataset to determine the target normal sample temperature rise early warning model includes:

[0037] The initial normal sample temperature rise early warning model is used to output the first target predicted temperature value corresponding to each switch cabinet characteristic operating parameter in the current year's normal sample dataset based on multiple switch cabinet characteristic operating parameters in the current year's normal sample dataset.

[0038] Based on the first target predicted temperature value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset, calculate the first target error value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset;

[0039] Determine whether the first target error value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset satisfies the first error criterion;

[0040] The switch cabinet feature operation parameters corresponding to the first target error value that satisfies the first error criterion in the current year's normal sample dataset are used as the first error parameters to construct the first large error dataset;

[0041] Compare whether the data ratio between the first large error dataset and the current year's normal sample dataset is less than or equal to a preset first ratio threshold.

[0042] If so, the initial normal sample temperature rise early warning model shall be used as the target normal sample temperature rise early warning model;

[0043] If not, then the initial normal sample temperature rise early warning model is corrected by adjusting the weight matrix based on the current year's normal sample dataset to determine the target normal sample temperature rise early warning model.

[0044] Optionally, the step of correcting the weight matrix of the initial abnormal sample temperature rise early warning model based on the second error criterion and using the current year's abnormal sample dataset to determine the target abnormal sample temperature rise early warning model includes:

[0045] The initial abnormal sample temperature rise early warning model is used to output the second target predicted temperature value corresponding to each switch cabinet characteristic operating parameter in the current year's abnormal sample dataset based on multiple switch cabinet characteristic operating parameters in the current year's abnormal sample dataset.

[0046] Based on the second target predicted temperature value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset, calculate the second target error value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset;

[0047] Determine whether the second target error value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset satisfies the second error criterion;

[0048] The switch cabinet feature operation parameters corresponding to the second target error value that satisfies the second error criterion in the current year's abnormal sample dataset are used as the second error parameters to construct the second largest error dataset;

[0049] Compare whether the data ratio between the second largest error dataset and the current year's abnormal sample dataset is less than or equal to the preset first ratio threshold;

[0050] If so, then the initial abnormal sample temperature rise early warning model shall be used as the target abnormal sample temperature rise early warning model;

[0051] If not, then determine whether the data ratio between the second largest error dataset and the current year's abnormal sample dataset is less than a preset second ratio threshold.

[0052] If so, the initial abnormal sample temperature rise early warning model is modified by weight matrix based on the current year's abnormal sample dataset to determine the target abnormal sample temperature rise early warning model.

[0053] Optionally, it also includes:

[0054] If the data ratio between the second largest error dataset and the current year's abnormal sample dataset is greater than the preset second ratio threshold, then based on the switch cabinet temperature rise warning result, the linear fitting method is used to determine the warning parameter weight matrix according to the weight matrix of the target temperature rise warning model and the weight matrix of the initial temperature rise warning model.

[0055] The target temperature rise early warning model is updated by using the aforementioned early warning parameter weight matrix to determine the temperature rise early warning model for future times.

[0056] The future temperature rise early warning model is used to output multiple predicted temperature values ​​for third targets based on the current year's switchgear sample dataset;

[0057] Determine whether all of the predicted temperature values ​​of the third target are less than or equal to a preset temperature threshold.

[0058] If not, the preset reference threshold is updated based on the preset constant to determine a new preset reference threshold.

[0059] A second aspect of the present invention provides a switchgear temperature rise early warning system, comprising:

[0060] The acquisition module is used to acquire historical switchgear sample datasets, preprocess the historical switchgear sample datasets, and output historical normal sample datasets and historical abnormal sample datasets.

[0061] The construction module is used to construct an initial temperature rise early warning model based on the historical annual normal sample dataset and the historical annual abnormal sample dataset, according to the preset evaluation criteria.

[0062] The correction module is used to correct the weight matrix of the initial temperature rise warning model by using a preset temperature error criterion and the current year's switchgear sample dataset when the current year's switchgear sample dataset is received, and output the target temperature rise warning model.

[0063] The early warning module is used to generate a switchgear temperature rise early warning result based on the weight matrix of the target temperature rise early warning model and the weight matrix of the initial temperature rise early warning model using preset early warning criteria.

[0064] As can be seen from the above technical solutions, the present invention has the following advantages:

[0065] The above-mentioned technical solution of the present invention provides a method for early warning of switchgear temperature rise. It acquires a historical annual switchgear sample dataset, preprocesses the dataset, and outputs a historical annual normal sample dataset and a historical annual abnormal sample dataset. Based on a preset evaluation criterion, an initial temperature rise early warning model is constructed using the historical annual normal sample dataset and the historical annual abnormal sample dataset. When the current annual switchgear sample dataset is received, the initial temperature rise early warning model is corrected using a preset temperature error criterion and the current annual switchgear sample dataset, resulting in a target temperature rise early warning model. The switchgear temperature rise early warning result is generated using a preset early warning criterion based on the weight matrix of the target temperature rise early warning model and the weight matrix of the initial temperature rise early warning model. Based on the above solution, an initial temperature rise early warning model is established using the preprocessed historical annual switchgear sample dataset. For engineering application requirements, an annual weight matrix correction method is used, i.e., the initial temperature rise early warning model is corrected using a preset temperature error criterion and the current annual switchgear sample dataset, resulting in a target temperature rise early warning model. This enables interactive iteration between the model and the equipment, thereby improving the model's engineering adaptability. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. 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.

[0067] Figure 1 This is a flowchart illustrating the steps of a switchgear temperature rise early warning method according to Embodiment 1 of the present invention.

[0068] Figure 2 This is a schematic diagram of the sample training and model building process provided in Embodiment 1 of the present invention;

[0069] Figure 3 This is a schematic diagram of the iterative calculation and early warning output of the model weight matrix provided in Embodiment 1 of the present invention;

[0070] Figure 4 This is a flowchart illustrating the switchgear temperature rise early warning method provided in Embodiment 1 of the present invention;

[0071] Figure 5 This is a structural block diagram of a switchgear temperature rise early warning system provided in Embodiment 2 of the present invention. Detailed Implementation

[0072] This invention provides a method and system for early warning of temperature rise in switchgear, which solves the technical problem that existing methods for early warning of temperature rise in switchgear are mainly based on mathematical models established under standard operating conditions, resulting in poor engineering adaptability of the models.

[0073] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0074] Please see Figure 1 , Figure 1 The flowchart illustrates the steps of a switchgear temperature rise early warning method provided in Embodiment 1 of the present invention.

[0075] The present invention provides a method for early warning of temperature rise in switchgear, comprising:

[0076] Step 101: Obtain the historical switchgear sample dataset, preprocess the historical switchgear sample dataset, and output the historical normal sample dataset and the historical abnormal sample dataset.

[0077] It should be noted that the historical annual switchgear sample dataset is the A0 dataset of switchgear operation samples collected within the past three years of commissioning and in normal equipment condition; where normal equipment condition mainly refers to the absence of equipment accidents or structural maintenance during the three-year service period; each switchgear characteristic operation parameter in the historical annual switchgear sample dataset includes bus temperature T, TEV value x1 (Transient Earth Voltage), contact resistance increment x2, load rate x3, humidity x4, cooling method x5, circuit breaker operating frequency x6, and bus insulation resistance x7.

[0078] Furthermore, the process of preprocessing the historical switchgear sample dataset to output the historical normal sample dataset and the historical abnormal sample dataset can be achieved by executing the following steps S11 to S13:

[0079] Step S11: Based on the multiple switch cabinet feature operation parameters and preset reference thresholds in the historical switch cabinet sample dataset, calculate multiple per-unit values ​​corresponding to each switch cabinet feature operation parameter in the historical switch cabinet sample dataset.

[0080] Step S12: Compare the multiple per-unit values ​​corresponding to the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset with the preset per-unit threshold to generate the comparison results corresponding to the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset.

[0081] Step S13: Based on the comparison results of the operation parameters of each switch cabinet feature in the historical switch cabinet sample dataset, classify the operation parameters of each switch cabinet feature in the historical switch cabinet sample dataset to generate historical normal sample dataset and historical abnormal sample dataset.

[0082] It should be noted that, referring to Table 1, the preset reference thresholds include the reference thresholds corresponding to bus temperature T, TEV value x1 (Transient Earth Voltage), contact resistance increment x2, load rate x3, humidity x4, cooling method x5, circuit breaker operating frequency x6, and bus insulation resistance x7. Based on the original reference threshold 'a' set for each parameter in the switchgear characteristic operating parameters (i.e., bus temperature T, TEV value x1, contact resistance increment x2, load rate x3, humidity x4, cooling method x5, circuit breaker operating frequency x6, and bus insulation resistance x7), the per-unit results (per-unit values) corresponding to each parameter in the switchgear characteristic operating parameters can be obtained. The multiple per-unit values ​​corresponding to the switchgear characteristic operating parameters are then compared with the preset reference thresholds. A per-unit threshold (with a value of 1) is set for comparison. When the comparison result shows that multiple per-unit values ​​corresponding to the switch cabinet's characteristic operating parameters are all ≤1, the switch cabinet's characteristic operating parameters are defined as normal data samples, and all normal data samples are used to form a historical annual normal sample dataset. When the comparison result shows that multiple per-unit values ​​corresponding to the switch cabinet's characteristic operating parameters are greater than 1, the switch cabinet's characteristic operating parameters are defined as abnormal data samples, and all abnormal data samples are used to form a historical annual abnormal sample dataset. Based on this classification process, historical annual normal sample dataset A1 and historical annual abnormal sample dataset A2 are obtained.

[0083]

[0084] Step 102: Based on the pre-set evaluation criteria, construct an initial temperature rise early warning model according to the historical annual normal sample dataset and the historical annual abnormal sample dataset.

[0085] The initial temperature rise warning model includes an initial normal sample temperature rise warning model and an initial abnormal sample temperature rise warning model.

[0086] Specifically, step 102 may include the following sub-steps S21-S23:

[0087] Step S21: Divide the historical annual normal sample dataset into two parts and output the first normal sample dataset and the second normal sample dataset.

[0088] It should be noted that the historical annual normal sample dataset A1 is decomposed into a 0.8A1 training set (the first normal sample dataset) and a 0.2A1 validation set (the second normal sample dataset) for the establishment and optimization of the switch cabinet temperature rise prediction model.

[0089] Step S22: Construct an initial normal sample temperature rise early warning model based on the first normal sample dataset and the second normal sample dataset;

[0090] Furthermore, step S22 may include the following sub-steps S221-S226:

[0091] Step S221: Based on the operating parameters of multiple switchgear features in the first normal sample dataset, construct the first intermediate temperature rise early warning model;

[0092] Step S222: Using the first intermediate temperature rise early warning model, output the first predicted temperature value based on the second normal sample dataset;

[0093] Step S223: Calculate the first error value based on the first predicted temperature value;

[0094] Step S224: Update the weight matrix of the first intermediate temperature rise early warning model based on the first error value, determine the second intermediate temperature rise early warning model, and count the number of model updates in real time;

[0095] Step S225: Determine whether the number of model updates has reached the preset threshold.

[0096] Step S226: If so, the second intermediate temperature rise early warning model is used as the initial normal sample temperature rise early warning model.

[0097] Step S23: Based on the preset evaluation criteria, the weight matrix of the initial normal sample temperature rise early warning model is corrected using the historical annual abnormal sample dataset to determine the initial abnormal sample temperature rise early warning model.

[0098] Furthermore, step S23 may include the following sub-steps S231-S235:

[0099] Step S231: Use multiple switchgear characteristic operation parameters from the historical annual abnormal sample data to perform single parameter perturbation correction on the initial normal sample temperature rise early warning model, and determine the third intermediate temperature rise early warning model.

[0100] Step S232: Using the third intermediate temperature rise early warning model, multiple switch cabinet characteristic operating parameters are generated based on the historical annual abnormal sample data to output multiple second predicted temperature values.

[0101] Step S233: Calculate multiple second error values ​​based on each second predicted temperature value;

[0102] Step S234: Determine whether multiple second error values ​​meet the preset evaluation criteria;

[0103] Step S235: If so, the third intermediate temperature rise early warning model is used as the initial abnormal sample temperature rise early warning model.

[0104] It should be noted that you should refer to [link / reference]. Figure 2After completing the partitioning of the historical annual normal sample dataset A1, in establishing the temperature rise prediction model, the bus temperature T from the switchgear characteristic operating parameters is selected as the output observation result (predicted temperature value), and the parameter set X={x1, x2, x3, x4, x5, x6, x7} from the switchgear characteristic operating parameters is selected as the input parameter. The intermediate hidden layer is... H represents the output of the intermediate hidden layer. and These are two key quantities that need to be identified in the hidden layer. The hidden layer parameters represent functional relationships; through iterative solutions using the 0.8A1 training set, the parametric model, namely the first intermediate temperature rise early warning model, can be obtained. ,in, Let X be the weight matrix of the parameter set (i.e., a matrix with one row and seven columns, consisting of the weight matrix elements corresponding to x1, x2, x3, x4, x5, x6, and x7). This is a paranoid trait.

[0105] Furthermore, in the validation of the temperature rise prediction model, a 0.2A1 validation set was used for residual calculation, and the weight matrix results were validated and corrected. Specifically, the switchgear characteristic operating parameters in the second normal sample dataset were used as input to the first intermediate temperature rise early warning model, outputting a first predicted temperature value. Based on the first predicted temperature value, a first error value was calculated, i.e., the residual ΔT of the 0.2A1 dataset was calculated. i (ΔT) i (For the first error value of the i-th switchgear characteristic operating parameter in the second normal sample dataset), define the residual sequence ΔT = {ΔT1, ΔT2, ..., ΔT...} i , ...}.

[0106] Furthermore, the residual sequence ΔT is divided into a residual training set and a residual validation set. The residual training set is used for hyperfunction optimization, and the residual validation set is used to evaluate the merits of the corrected model. Specifically, the weight matrix of the first intermediate temperature rise early warning model is updated based on the first error value to determine the second intermediate temperature rise early warning model, and the number of model updates is counted in real time. or step is the step size. To correct The constructed random function, through these two parameters, can be used to construct the direction control for iterative solution; the updated value is calculated using the first error value. Use updated values The weight matrix of the first intermediate temperature rise early warning model is updated, and the number of model updates is counted in real time. When the preset number threshold is not reached, the second intermediate temperature rise early warning model is used as the new first intermediate temperature rise early warning model. New switch cabinet characteristic operating parameters are collected from the second normal sample dataset as input to the new first intermediate temperature rise early warning model to obtain a new first predicted temperature value. Then, step S223 is executed until the number of model updates reaches the preset number threshold. The second intermediate temperature rise early warning model determined when the number of model updates reaches the preset number threshold is used as the initial normal sample temperature rise early warning model.

[0107] Furthermore, based on the iterative solution of the 0.8A1 training set, the temperature rise prediction model can be established. Combined with the residual calculation and parameter correction based on the 0.2A1 validation set, the weight matrix results can be verified and optimized, thus obtaining the initial normal sample temperature rise early warning model.

[0108] Furthermore, based on the established temperature rise prediction model and weight matrix results, the temperature rise error of the abnormal sample set A2 can be calculated, and abnormal parameters can be extracted and error results calibrated. Each switchgear characteristic operating parameter in the historical annual abnormal sample dataset includes both abnormal and normal parameters. Abnormal parameters are those with a per-unit value greater than 1, while normal parameters are those with a per-unit value less than or equal to 1.

[0109] Furthermore, the initial normal sample temperature rise early warning model is corrected by single-parameter perturbation using multiple switchgear characteristic operating parameters from the historical annual abnormal sample dataset. Specifically, the error results are calculated using the following relationship. , These are the measured results from the abnormal sample set A2. The results are obtained from calculations based on the theoretical model. Includes normal parameters and abnormal parameters , The original value is the weight matrix. (In error calculation and optimization, normal parameters) The values ​​of , i.e., the elements of the weight matrix Remain unchanged, abnormal parameter The value of Optimization is required, and the corrected result (corrected weight matrix) is labeled as follows: ). In obtaining During the correction process, a single-parameter perturbation correction method is adopted (when a set of data has multiple outlier parameters, that set of data will be used in the data processing of each single-parameter correction). For example, if x1 in a switch cabinet characteristic operation parameter in a historical year's outlier sample dataset is an outlier parameter, then this switch cabinet characteristic operation parameter is first used as the input of the initial normal sample temperature rise early warning model, the predicted temperature value is output, and the corresponding error result is calculated. Based on the error result, the updated value is calculated. Utilize updated values For the weight matrix The weight matrix elements corresponding to x1 are updated to obtain the weight matrix elements corresponding to x1. When x1 is a normal parameter among the other switchgear feature operation parameters in the historical abnormal sample dataset, the element of the corrected weight matrix corresponding to x1 is... Save the output. If an abnormal parameter x1 still exists among the remaining switchgear feature parameters in the historical year's abnormal sample dataset, continue processing... After making corrections, we can similarly obtain the weight matrix elements corresponding to x2, x3, x4, x5, x6, and x7, thus obtaining the final corrected weight matrix. Thus, the third intermediate temperature rise early warning model was obtained.

[0110] Furthermore, the operating parameters of multiple switchgear features from the historical annual abnormal sample dataset are re-input into the third intermediate temperature rise early warning model to obtain multiple second predicted temperature values. Then, based on each second predicted temperature value, multiple second error values ​​are calculated to establish an error matrix. ={( , (), , For example, if x1 in the characteristic operating parameters of the switchgear is an abnormal parameter, while x2, x3, x4, x5, x6, and x7 are normal parameters, then the third intermediate temperature rise early warning model selects the parameter corresponding to x1. The weight matrix elements corresponding to x2, x3, x4, x5, x6, and x7 in the initial normal sample temperature rise early warning model are selected to output the second predicted temperature value, thereby calculating the corresponding second error value. If both x1 and x2 are abnormal parameters in the switchgear characteristic operation parameters, then the value corresponding to x1 is selected. x2 corresponds to The weight matrix elements corresponding to x3, x4, x5, x6, and x7 in the weight matrix of the initial normal sample temperature rise early warning model are selected to output the second predicted temperature value, thereby calculating the corresponding second error value and incorporating it into the error matrix.

[0111] Furthermore, the identification and optimization performed using the scatter-point fitting method is evaluated based on the following criteria: the pre-defined evaluation criteria are as follows:

[0112] ;

[0113] in, This is the j-th second error value; This is the average value calculated based on all second error values; Let be the standard deviation of ΔT, where ΔT is the second error value; and n be the total number of second error values.

[0114] Furthermore, if there are 80% or more of the second error values ​​that satisfy the preset evaluation criterion, the correction process is stopped, and the third intermediate temperature rise warning model determined when 80% of the second error values ​​satisfy the preset evaluation criterion is used as the initial abnormal sample temperature rise warning model. If there are less than 80% of the second error values ​​that satisfy the preset evaluation criterion, the third intermediate temperature rise warning model is used as the new initial normal sample temperature rise warning model, and the process jumps to step S231 until there are 80% or more of the second error values ​​that satisfy the preset evaluation criterion.

[0115] Step 103: When the current year's switchgear sample dataset is received, the initial temperature rise warning model is corrected by weight matrix using the preset temperature error criterion and the current year's switchgear sample dataset, and the target temperature rise warning model is output.

[0116] The target temperature rise early warning model includes a target normal sample temperature rise early warning model and a target abnormal sample temperature rise early warning model.

[0117] The preset temperature error criterion includes a first error criterion and a second error criterion.

[0118] Specifically, step 103 may include the following sub-steps S31-S33:

[0119] Step S31: Preprocess the current year's switchgear sample dataset and output the current year's normal sample dataset and the current year's abnormal sample dataset;

[0120] Step S32: Based on the first error criterion, the weight matrix of the initial normal sample temperature rise early warning model is corrected using the current year's normal sample dataset to determine the target normal sample temperature rise early warning model;

[0121] Furthermore, step S32 may include the following sub-steps S321-S327:

[0122] Step S321: Using the initial normal sample temperature rise early warning model, based on the operating parameters of multiple switchgear features in the current year's normal sample dataset, output the first target predicted temperature value corresponding to each switchgear feature operating parameter in the current year's normal sample dataset.

[0123] Step S322: Based on the first target predicted temperature value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset, calculate the first target error value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset.

[0124] Step S323: Determine whether the first target error value corresponding to the characteristic operating parameters of each switch cabinet in the current year's normal sample dataset meets the first error criterion;

[0125] Step S324: Take the switch cabinet feature operation parameter corresponding to the first target error value that satisfies the first error criterion in the current year's normal sample dataset as the first error parameter and construct the first large error dataset;

[0126] Step S325: Compare the data ratio between the first large error dataset and the current year's normal sample dataset to see if it is less than or equal to a preset first ratio threshold.

[0127] Step S326: If so, the initial normal sample temperature rise early warning model shall be used as the target normal sample temperature rise early warning model.

[0128] Step S327: If not, then correct the weight matrix of the initial normal sample temperature rise early warning model based on the current year's normal sample dataset to determine the target normal sample temperature rise early warning model.

[0129] Step S33: Based on the second error criterion, the weight matrix of the initial abnormal sample temperature rise early warning model is corrected using the current year's abnormal sample dataset to determine the target abnormal sample temperature rise early warning model.

[0130] Further, step S33 may include the following sub-steps S331-S338:

[0131] Step S331: Using the initial abnormal sample temperature rise early warning model, based on the operating parameters of multiple switchgear features in the current year's abnormal sample dataset, output the second target predicted temperature value corresponding to each switchgear feature operating parameter in the current year's abnormal sample dataset.

[0132] Step S332: Based on the second target predicted temperature value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset, calculate the second target error value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset.

[0133] Step S333: Determine whether the second target error value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset satisfies the second error criterion;

[0134] Step S334: Take the switch cabinet feature operation parameter corresponding to the second target error value that satisfies the second error criterion in the current year's abnormal sample dataset as the second error parameter and construct the second largest error dataset;

[0135] Step S335: Compare whether the data ratio between the second largest error dataset and the current year's abnormal sample dataset is less than or equal to a preset first ratio threshold;

[0136] Step S336: If so, the initial abnormal sample temperature rise early warning model shall be used as the target abnormal sample temperature rise early warning model.

[0137] Step S337: If not, determine whether the data ratio between the second largest error dataset and the current year's abnormal sample dataset is less than the preset second ratio threshold.

[0138] Step S338: If so, then the weight matrix of the initial abnormal sample temperature rise early warning model is corrected based on the current year's abnormal sample dataset to determine the target abnormal sample temperature rise early warning model.

[0139] It should be noted that you should refer to [link / reference]. Figure 3 When sample data B0 of switchgear from year k is received, i.e., the current year's switchgear sample dataset, the sample data is classified based on the reference threshold 'a' of each parameter in the parameter set X, into a normal sample set B1 and an abnormal sample set B2. This involves preprocessing the current year's switchgear sample dataset to obtain the current year's normal sample dataset B1 and the current year's abnormal sample dataset B2. The principle behind the preprocessing of the current year's switchgear sample dataset is the same as that described above for preprocessing historical year's switchgear sample datasets, and will not be elaborated further in this invention.

[0140] Furthermore, based on the established temperature rise prediction models—namely, the initial normal sample temperature rise early warning model and the initial abnormal sample temperature rise early warning model—error analysis is performed on the normal sample set B1 and the abnormal sample set B2. The original parameters for year k are... and These two parameters will be iteratively updated during error optimization.

[0141] Furthermore, using the first error criterion, "temperature error ≤ 5%", data screening is performed on the normal sample set B1 to output the first large error dataset B3. Specifically, the first target error value less than or equal to 5%, i.e., the switch cabinet characteristic operation parameter corresponding to the first target error value that satisfies the first error criterion, is used as the first error parameter. The first large error dataset is constructed using all the first error parameters. When the data ratio B3 / B1 between the first large error dataset and the current year's normal sample dataset is ≤ 10% (preset first ratio threshold), no correction is required, and the initial normal sample temperature rise warning model is used as the target normal sample temperature rise warning model. When B3 / B1 > 10%, the initial normal sample temperature rise warning model is corrected by weight matrix based on the current year's normal sample dataset to determine the target normal sample temperature rise warning model. The principle of this weight matrix correction process is the same as the principle of the above-mentioned determination of the initial normal sample temperature rise warning model, and will not be elaborated further in this invention.

[0142] Furthermore, using the second error criterion, "temperature error ≤ 10%", data screening is performed on the abnormal sample set B2, outputting the second largest error dataset B4. When the data ratio between the second largest error dataset and the current year's abnormal sample dataset, B4 / B2, is ≤ 10%, no weight matrix correction is needed, and the initial abnormal sample temperature rise warning model is used as the target abnormal sample temperature rise warning model. When 30% > B4 / B2 > 10%, the weight matrix of the initial abnormal sample temperature rise warning model is corrected based on the current year's abnormal sample dataset to determine the target abnormal sample temperature rise warning model. The principle of this weight correction process is consistent with the principle of determining the initial abnormal sample temperature rise warning model, and will not be elaborated further in this invention. When B4 / B2 > 30%, the k-year reference threshold a is used. k The correction involves adjusting the reference thresholds corresponding to the following values ​​in the preset reference thresholds: TEV value x1 (Transient Earth Voltage), contact resistance increment x2, load rate x3, humidity x4, cooling method x5, circuit breaker operating frequency x6, and bus insulation resistance x7.

[0143] Step 104: Using preset early warning criteria, generate switchgear temperature rise early warning results based on the weight matrix of the target temperature rise early warning model and the weight matrix of the initial temperature rise early warning model.

[0144] It should be noted that, based on the iterative calculations and parameter adjustments for year k, the model and parameters for year k+1 are obtained after the solution converges. and The output is the weight matrix of the target normal sample temperature rise early warning model. The weight matrix of the target abnormal sample temperature rise early warning model The parameter results output by this model will be used as the raw values ​​for model calculations in year k+1. Iterative calculations for the annual model application are performed sequentially. and The iterative updates can be used to characterize the evolution of the aging characteristics of the switchgear during its service life.

[0145] Furthermore, this invention employs a pre-set early warning criterion for monitoring parameter x in parameter set X. i The warning determination, the specific criterion for the pre-set warning is as follows: W Tk+1 The monitoring parameter x in the weight matrix of the target temperature rise early warning model for year k+1 i The corresponding weight matrix elements, W Tk-1 The monitoring parameter x for the temperature rise early warning model in year k-1 i The corresponding weight matrix elements, W Tk The monitoring parameter x in the weight matrix of the initial temperature rise warning model for year k i The corresponding weight matrix elements, when a certain monitoring parameter x i Whether it is a normal or abnormal parameter, the monitoring parameter x in the initial normal sample temperature rise early warning model is considered. i The corresponding weight matrix elements, and the monitoring parameter x in the target normal sample temperature rise early warning model. i The corresponding weight matrix elements, or the monitoring parameter x in the initial abnormal sample temperature rise early warning model. i The corresponding weight matrix elements, and the monitoring parameter x in the target abnormal sample temperature rise early warning model. i The corresponding weight matrix elements are used to determine whether the preset early warning criteria are met. If they are met, the generated switchgear temperature rise early warning result is determined to require an early warning, and this monitoring parameter x needs to be closely monitored in the following year. i The characteristic changes; if there are no parameters in the parameter set X that satisfy the preset early warning criterion, the generated switchgear temperature rise early warning result is determined to be that no early warning is required.

[0146] Optionally, a reference threshold a is set for year k. k The correction process also includes:

[0147] If the data ratio between the second largest error dataset and the current year's abnormal sample dataset is greater than the preset second ratio threshold, then based on the switch cabinet temperature rise warning result, the linear fitting method is used to determine the warning parameter weight matrix according to the weight matrix of the target temperature rise warning model and the weight matrix of the initial temperature rise warning model.

[0148] The target temperature rise early warning model is updated by using the early warning parameter weight matrix to determine the temperature rise early warning model at future times.

[0149] The future temperature rise early warning model is used to output multiple predicted temperature values ​​for third targets based on the current year's switchgear sample dataset;

[0150] Determine whether the predicted temperature values ​​of multiple third targets are all less than or equal to a preset temperature threshold;

[0151] If not, the preset reference threshold is updated based on the preset constant to determine a new preset reference threshold.

[0152] It should be noted that this is based on the monitoring parameter x that requires early warning. i The output, using linear fitting method based on W Tk , W Tk+1 , The value relationships are used to determine the early warning parameter weight matrix, which includes the normal early warning parameter weight matrix W. Tk+2 and abnormal early warning parameter weight matrix The normal sample set B1 and the abnormal sample set B2 collected within k years are used as the input sample data for the simulation calculation. That is, the data uses the normal sample set B1 and the abnormal sample set B2, and the weights are W. Tk+2 and Calculate the predicted temperature T for all data, and determine if T ≤ 60℃ (preset temperature threshold). If all T values ​​are less than or equal to 60℃, the condition is met, and there is no need to update the value of a (there is no monitoring parameter x that requires an early warning). i In general, the conditions are met. If there are parameters that do not meet the criteria, then the preset reference threshold is updated based on the preset constant to determine a new preset reference threshold, where the monitoring parameter x that needs to be alerted is used. i Subtracting the preset constant from the corresponding threshold yields the monitoring parameter x that requires early warning. i The corresponding new threshold is then used to obtain a new preset reference threshold.

[0153] It is worth mentioning that if there is no monitoring parameter x that requires early warning... i If so, there is no need to update the preset reference threshold.

[0154] In this embodiment, based on the aforementioned model parameter W T , The model data for year k+1 is updated by correcting the reference threshold a in parameter set X and updating the reference threshold a in parameter set X, so as to be used for the investigation and early warning of potential temperature rise hazards in year k+1.

[0155] For comparison of technical effects, existing technologies can be used as a reference. Switchgear is a crucial piece of equipment in distribution network systems, and its safe and reliable operation has become a key indicator for evaluating the level of distribution network automation. A survey of the current operational status and safety hazards of existing switchgear reveals that temperature rise and overheating are among the most common fault hazards in switchgear. These not only affect the safe and reliable operation of the equipment but also directly impact the insulation and aging of conductive components, shortening the equipment's lifespan and increasing maintenance risks. It should be noted that many factors influence the temperature rise characteristics within switchgear, making the tracing of its origins difficult and hindering the fundamental solution for identifying and warning of temperature rise hazards within switchgear, resulting in a very passive approach to engineering maintenance.

[0156] In existing technologies, temperature is merely used as a monitoring object for assessing the operating status of switchgear. When the temperature reaches a certain value, contingency plans can only be passively implemented through power outages for maintenance or operation under high-temperature loads. The origins of temperature rise in switchgear are known to be related to many factors, including partial discharge, busbar conductivity, load rate, cooling, and aging. Therefore, establishing a multi-dimensional training model for identifying potential temperature rise hazards would have significant engineering implications for tracing the source and optimizing temperature rise.

[0157] Based on the above, the shortcomings of the existing technology are as follows: (1) There are many factors affecting the temperature rise of switchgear, and there is a lack of effective mathematical models to predict the temperature rise characteristics. (2) There is a lack of effective evaluation criteria and methods for tracing the source of temperature rise in switchgear. The temperature rise control and optimization strategies are too blind and lack technical support. (3) How to ensure the adaptability of a theoretical model in different equipment, and how to ensure the reliability of the model and method in engineering applications, which need to be considered in engineering applications, are problems that must be solved in engineering applications.

[0158] To address the aforementioned issues, this invention proposes a switchgear temperature rise early warning method. Starting with the key factors influencing switchgear temperature rise, it proposes a multi-source data fusion approach to establish a temperature rise prediction model. During the process, the weights of monitoring parameters are calculated and corrected to improve the model's adaptability to different equipment and application scenarios. To address the difficulty of a single model meeting the engineering application adaptability requirements of different equipment and service scenarios, this invention collects relatively ideal sample data and establishes a basic temperature rise prediction model through sample classification and training. To enhance the model's engineering adaptability, an annual weight iterative correction method is used to achieve interactive iteration between the model and the equipment, thereby achieving optimized equipment parameter configuration and application. Simultaneously, based on the established temperature rise prediction model and the setting of weights for each monitoring parameter, the high-weight potential sources of temperature rise impact are extracted through annual dynamic weight change trend analysis, thus realizing the output of potential monitoring parameter hazards. Furthermore, the weights and reference thresholds in the model can be calculated and updated according to the annual collected data and weight change patterns, facilitating the investigation and early warning application of temperature rise hazard risks.

[0159] Specifically, please refer to Figure 4 First, the operation sample data of the switchgear is collected and classified. This part mainly involves collecting operation sample data of switchgear within three years of commissioning and in normal equipment condition, and classifying the data into normal and abnormal sample sets based on the original threshold settings of each collected parameter. Second, the sample data is trained and a temperature rise prediction model is generated. This part mainly involves training the collected sample data to establish a mathematical model for predicting temperature rise hazards. The normal sample set is used for generating the basic model and calculating weights, with optimization achieved through the setting of training and validation sets; the abnormal sample set is used to correct the relevant weights when abnormal parameters are input. Next, error analysis and parameter weight correction are performed through the application of the k-year model (initial temperature rise early warning model). This part mainly involves applying the established model to the engineering application of switchgear temperature rise hazard assessment, and correcting the relevant parameter weight values ​​through comparison and optimization of error results. Simultaneously, the parameters of the k+1-year model (target temperature rise early warning model) are updated, and early warnings for monitoring parameters are issued. This section primarily involves revising the model parameter weights for the aforementioned year k, and finally updating the model parameter data for year k+1. This update is used for assessing the potential temperature rise hazard in the switchgear for year k+1. Based on the changing trends of the monitoring parameter weight values, early warning outputs for sensitive detection parameters are generated. The thresholds for the early warning parameters for year k+1 are also calculated and updated. This section primarily involves updating the model parameter weights for the aforementioned year k+1, calculating and updating the thresholds for relevant early warning parameters, and outputting the temperature rise hazard investigation model for year k+1. The output of the temperature rise hazard investigation model for year k+1 will be generated after the model parameter revision and threshold update are completed.

[0160] In summary, this invention proposes a switchgear temperature rise early warning method for predicting temperature rise and identifying potential risks under various monitoring parameters. Based on the analysis of factors influencing switchgear temperature rise, it proposes multi-source parameter monitoring and reference threshold settings, and establishes a temperature rise prediction model under multi-source disturbances using sample data components and training. The use of sample sets and training sets improves the accuracy of iterative calculations. Furthermore, considering engineering application needs, it proposes a method for adjusting the weights of monitoring parameters in the model using annual sample data iterative calculations, improving the model's adaptability to different equipment and application scenarios. In addition, this invention has applications for temperature rise tracing and early warning output. During the model's iterative calculations, the annual dynamic weight change trend can be obtained, which can be used for tracing the sources of temperature rise influencing factors and providing early warning output for monitoring parameters.

[0161] In this embodiment of the invention, a method for early warning of switchgear temperature rise is provided. The method involves acquiring a historical annual switchgear sample dataset, preprocessing the dataset to output a historical annual normal sample dataset and a historical annual abnormal sample dataset. Based on a pre-set evaluation criterion, an initial temperature rise warning model is constructed using the historical annual normal sample dataset and the historical annual abnormal sample dataset. When the current annual switchgear sample dataset is received, the initial temperature rise warning model is corrected using a pre-set temperature error criterion and the current annual switchgear sample dataset, resulting in a target temperature rise warning model. The method also uses a pre-set warning criterion to generate a switchgear temperature rise warning result based on the weight matrix of the target temperature rise warning model and the initial temperature rise warning model. Based on the above scheme, an initial temperature rise warning model is established using the pre-processed historical annual switchgear sample dataset. To meet engineering application requirements, an annual weight matrix correction method is used, i.e., the initial temperature rise warning model is corrected using a pre-set temperature error criterion and the current annual switchgear sample dataset, resulting in a target temperature rise warning model. This enables interactive iteration between the model and the equipment, thereby improving the model's engineering adaptability.

[0162] Please see Figure 5 , Figure 5 This is a structural block diagram of a switchgear temperature rise early warning system provided in Embodiment 2 of the present invention.

[0163] The present invention provides a switchgear temperature rise early warning system, comprising:

[0164] The acquisition module 501 is used to acquire historical switchgear sample datasets, preprocess the historical switchgear sample datasets, and output historical normal sample datasets and historical abnormal sample datasets.

[0165] Module 502 is used to construct an initial temperature rise early warning model based on a pre-set evaluation criterion and the historical annual normal sample dataset and the historical annual abnormal sample dataset.

[0166] The correction module 503 is used to correct the weight matrix of the initial temperature rise warning model by using a preset temperature error criterion and the current year's switchgear sample dataset when the current year's switchgear sample dataset is received, and output the target temperature rise warning model.

[0167] The early warning module 504 is used to generate a switchgear temperature rise early warning result based on the weight matrix of the target temperature rise early warning model and the weight matrix of the initial temperature rise early warning model using preset early warning criteria.

[0168] Furthermore, module 501 is specifically used for:

[0169] Based on the multiple switch cabinet feature operation parameters and preset reference thresholds in the historical switch cabinet sample dataset, calculate multiple per-unit values ​​corresponding to each switch cabinet feature operation parameter in the historical switch cabinet sample dataset.

[0170] The per-unit values ​​corresponding to the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset are compared with the preset per-unit threshold to generate the comparison results corresponding to the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset.

[0171] Based on the comparison results of the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset, the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset are classified to generate historical normal sample datasets and historical abnormal sample datasets.

[0172] Furthermore, the initial temperature rise early warning model includes an initial normal sample temperature rise early warning model and an initial abnormal sample temperature rise early warning model; construction module 502 includes:

[0173] The first submodule is used to divide the historical annual normal sample dataset and output the first normal sample dataset and the second normal sample dataset.

[0174] The second submodule is used to construct an initial normal sample temperature rise early warning model based on the first normal sample dataset and the second normal sample dataset.

[0175] The third submodule is used to correct the weight matrix of the initial normal sample temperature rise early warning model based on the pre-set evaluation criteria and the historical annual abnormal sample dataset, so as to determine the initial abnormal sample temperature rise early warning model.

[0176] Furthermore, the second submodule is specifically used for:

[0177] Based on the operating parameters of multiple switchgear features in the first normal sample dataset, a first intermediate temperature rise early warning model is constructed.

[0178] The first intermediate temperature rise early warning model is used to output the first predicted temperature value based on the second normal sample dataset;

[0179] Calculate the first error value based on the first predicted temperature value;

[0180] The weight matrix of the first intermediate temperature rise early warning model is updated based on the first error value to determine the second intermediate temperature rise early warning model, and the number of model updates is counted in real time.

[0181] Determine whether the number of model updates has reached the preset threshold;

[0182] If so, the second intermediate temperature rise early warning model will be used as the initial normal sample temperature rise early warning model.

[0183] Furthermore, the third submodule is specifically used for:

[0184] The initial normal sample temperature rise early warning model was corrected by single parameter perturbation using multiple switch cabinet characteristic operating parameters in the historical annual abnormal sample dataset, and the third intermediate temperature rise early warning model was determined.

[0185] The third intermediate temperature rise early warning model is adopted to output multiple second predicted temperature values ​​based on the operating parameters of multiple switchgear characteristics in the historical annual abnormal sample dataset.

[0186] Based on each of the second predicted temperature values, multiple second error values ​​are calculated;

[0187] Determine whether multiple second error values ​​meet the preset evaluation criteria;

[0188] If so, the third intermediate temperature rise early warning model will be used as the initial abnormal sample temperature rise early warning model.

[0189] Furthermore, the target temperature rise early warning model includes a target normal sample temperature rise early warning model and a target abnormal sample temperature rise early warning model; the preset temperature error criteria include a first error criterion and a second error criterion; the correction module 503 includes:

[0190] The fourth submodule is used to preprocess the current year's switchgear sample dataset and output the current year's normal sample dataset and the current year's abnormal sample dataset.

[0191] The fifth submodule is used to correct the weight matrix of the initial normal sample temperature rise early warning model based on the first error criterion and the current year's normal sample dataset, and to determine the target normal sample temperature rise early warning model.

[0192] The sixth submodule is used to correct the weight matrix of the initial abnormal sample temperature rise early warning model based on the second error criterion and the current year's abnormal sample dataset, and to determine the target abnormal sample temperature rise early warning model.

[0193] Furthermore, the fifth submodule is specifically used for:

[0194] The initial normal sample temperature rise early warning model is adopted to output the first target predicted temperature value corresponding to each switch cabinet characteristic operating parameter in the current year's normal sample dataset based on multiple switch cabinet characteristic operating parameters in the current year's normal sample dataset.

[0195] Based on the first target predicted temperature value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset, calculate the first target error value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset.

[0196] Determine whether the first target error value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset satisfies the first error criterion;

[0197] The switch cabinet feature operation parameter corresponding to the first target error value that meets the first error criterion in the current year's normal sample dataset is used as the first error parameter and the first large error dataset is constructed.

[0198] Compare the proportion of data between the largest error dataset and the current year's normal sample dataset to see if it is less than or equal to a preset first proportion threshold.

[0199] If so, the initial normal sample temperature rise early warning model will be used as the target normal sample temperature rise early warning model;

[0200] If not, then the weight matrix of the initial normal sample temperature rise early warning model is corrected based on the current year's normal sample dataset to determine the target normal sample temperature rise early warning model.

[0201] Furthermore, the sixth submodule is specifically used for:

[0202] The initial abnormal sample temperature rise early warning model is adopted to output the second target predicted temperature value corresponding to the operating parameters of each switch cabinet in the current year's abnormal sample dataset based on the operating parameters of multiple switch cabinets in the current year's abnormal sample dataset.

[0203] Based on the second target predicted temperature value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset, calculate the second target error value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset.

[0204] Determine whether the second target error value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset satisfies the second error criterion;

[0205] The switch cabinet feature operation parameters corresponding to the second target error value that meets the second error criterion in the current year's abnormal sample dataset are used as the second error parameters to construct the second largest error dataset;

[0206] Compare the data proportion between the second largest error dataset and the current year's abnormal sample dataset to see if it is less than or equal to a preset first proportion threshold.

[0207] If so, the initial abnormal sample temperature rise early warning model will be used as the target abnormal sample temperature rise early warning model;

[0208] If not, then determine whether the data ratio between the second largest error dataset and the current year's abnormal sample dataset is less than the preset second ratio threshold.

[0209] If so, the initial abnormal sample temperature rise early warning model is modified by adjusting the weight matrix based on the current year's abnormal sample dataset to determine the target abnormal sample temperature rise early warning model.

[0210] In one optional system embodiment, it further includes:

[0211] The first module is used to determine the warning parameter weight matrix based on the switch cabinet temperature rise warning result if the data ratio between the second largest error dataset and the current year's abnormal sample dataset is greater than the preset second ratio threshold. The module uses a linear fitting method to determine the warning parameter weight matrix according to the weight matrix of the target temperature rise warning model and the weight matrix of the initial temperature rise warning model.

[0212] The second module is used to update the weight matrix of the target temperature rise early warning model using the early warning parameter weight matrix, and to determine the temperature rise early warning model at future times.

[0213] The third module is used to output multiple predicted temperature values ​​for the third target based on the current year's switchgear sample dataset using a future temperature rise early warning model.

[0214] The fourth module is used to determine whether the predicted temperature values ​​of multiple third targets are all less than or equal to a preset temperature threshold.

[0215] The fifth module is used to update the preset reference threshold based on the preset constant if no, and to determine the new preset reference threshold.

[0216] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and sub-modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0217] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0218] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0219] The above-described 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for early warning of temperature rise in switchgear, characterized in that, include: Obtain historical switchgear sample datasets, preprocess the historical switchgear sample datasets, and output historical normal sample datasets and historical abnormal sample datasets. Based on the pre-set evaluation criteria, an initial temperature rise early warning model is constructed according to the historical annual normal sample dataset and the historical annual abnormal sample dataset. When the current year's switchgear sample dataset is received, the initial temperature rise warning model is corrected by weight matrix using a preset temperature error criterion and the current year's switchgear sample dataset, and the target temperature rise warning model is output. The switchgear temperature rise warning result is generated based on the weight matrix of the target temperature rise warning model and the weight matrix of the initial temperature rise warning model using a pre-set warning criterion.

2. The switchgear temperature rise early warning method according to claim 1, characterized in that, The preprocessing of the historical switchgear sample dataset to output historical normal sample datasets and historical abnormal sample datasets includes: Based on multiple switch cabinet feature operation parameters and preset reference thresholds in the historical switch cabinet sample dataset, calculate multiple per-unit values ​​corresponding to each switch cabinet feature operation parameter in the historical switch cabinet sample dataset. The per-unit values ​​corresponding to the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset are compared with a preset per-unit threshold to generate the comparison results corresponding to the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset. Based on the comparison results of the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset, the operating parameters of each switch cabinet feature in the historical switch cabinet sample dataset are classified to generate a historical normal sample dataset and a historical abnormal sample dataset.

3. The switchgear temperature rise early warning method according to claim 1, characterized in that, The initial temperature rise warning model includes an initial normal sample temperature rise warning model and an initial abnormal sample temperature rise warning model; the initial temperature rise warning model, constructed based on preset evaluation criteria and according to the historical annual normal sample dataset and the historical annual abnormal sample dataset, includes: The historical annual normal sample dataset is divided into a first normal sample dataset and a second normal sample dataset; Based on the first normal sample dataset and the second normal sample dataset, an initial normal sample temperature rise early warning model is constructed; Based on the pre-set evaluation criteria, the initial normal sample temperature rise early warning model is corrected by adjusting the weight matrix using the historical annual abnormal sample dataset, and the initial abnormal sample temperature rise early warning model is determined.

4. The switchgear temperature rise early warning method according to claim 3, characterized in that, The step of constructing an initial normal sample temperature rise early warning model based on the first normal sample dataset and the second normal sample dataset includes: Based on the operating parameters of multiple switchgear features in the first normal sample dataset, a first intermediate temperature rise early warning model is constructed. The first intermediate temperature rise early warning model is used to output a first predicted temperature value based on the second normal sample dataset; Calculate the first error value based on the first predicted temperature value; The weight matrix of the first intermediate temperature rise early warning model is updated based on the first error value to determine the second intermediate temperature rise early warning model, and the number of model updates is counted in real time. Determine whether the number of model updates has reached a preset threshold; If so, the second intermediate temperature rise early warning model shall be used as the initial normal sample temperature rise early warning model.

5. The switchgear temperature rise early warning method according to claim 3, characterized in that, The step of adjusting the weight matrix of the initial normal sample temperature rise early warning model based on the preset evaluation criteria and using the historical annual abnormal sample dataset to determine the initial abnormal sample temperature rise early warning model includes: The initial normal sample temperature rise early warning model is corrected by single parameter perturbation using multiple switch cabinet feature operation parameters in the historical annual abnormal sample dataset to determine the third intermediate temperature rise early warning model. The third intermediate temperature rise early warning model is used to output multiple second predicted temperature values ​​based on multiple switch cabinet characteristic operating parameters in the historical annual abnormal sample dataset. Based on each of the second predicted temperature values, a plurality of second error values ​​are calculated; Determine whether multiple second error values ​​satisfy the preset evaluation criteria; If so, the third intermediate temperature rise early warning model shall be used as the initial abnormal sample temperature rise early warning model.

6. The switchgear temperature rise early warning method according to claim 3, characterized in that, The target temperature rise early warning model includes a target normal sample temperature rise early warning model and a target abnormal sample temperature rise early warning model; the preset temperature error criterion includes a first error criterion and a second error criterion. The initial temperature rise warning model is corrected using a weight matrix based on a preset temperature error criterion and the current year's switchgear sample dataset, resulting in a target temperature rise warning model, including: The current year's switchgear sample dataset is preprocessed to output the current year's normal sample dataset and the current year's abnormal sample dataset; Based on the first error criterion, the weight matrix of the initial normal sample temperature rise early warning model is corrected using the current year's normal sample dataset to determine the target normal sample temperature rise early warning model. Based on the second error criterion, the weight matrix of the initial abnormal sample temperature rise early warning model is corrected using the current year's abnormal sample dataset to determine the target abnormal sample temperature rise early warning model.

7. The switchgear temperature rise early warning method according to claim 6, characterized in that, The step of correcting the weight matrix of the initial normal sample temperature rise early warning model based on the first error criterion and using the current year's normal sample dataset to determine the target normal sample temperature rise early warning model includes: The initial normal sample temperature rise early warning model is used to output the first target predicted temperature value corresponding to each switch cabinet characteristic operating parameter in the current year's normal sample dataset based on multiple switch cabinet characteristic operating parameters in the current year's normal sample dataset. Based on the first target predicted temperature value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset, calculate the first target error value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset; Determine whether the first target error value corresponding to the characteristic operating parameters of each switchgear in the current year's normal sample dataset satisfies the first error criterion; The switch cabinet feature operation parameters corresponding to the first target error value that satisfies the first error criterion in the current year's normal sample dataset are used as the first error parameters to construct the first large error dataset; Compare whether the data ratio between the first large error dataset and the current year's normal sample dataset is less than or equal to a preset first ratio threshold. If so, the initial normal sample temperature rise early warning model shall be used as the target normal sample temperature rise early warning model; If not, then the initial normal sample temperature rise early warning model is corrected by adjusting the weight matrix based on the current year's normal sample dataset to determine the target normal sample temperature rise early warning model.

8. The switchgear temperature rise early warning method according to claim 7, characterized in that, The step of using the current year's abnormal sample dataset to correct the weight matrix of the initial abnormal sample temperature rise early warning model based on the second error criterion, and determining the target abnormal sample temperature rise early warning model, includes: The initial abnormal sample temperature rise early warning model is used to output the second target predicted temperature value corresponding to each switch cabinet characteristic operating parameter in the current year's abnormal sample dataset based on multiple switch cabinet characteristic operating parameters in the current year's abnormal sample dataset. Based on the second target predicted temperature value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset, calculate the second target error value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset; Determine whether the second target error value corresponding to the characteristic operating parameters of each switchgear in the current year's abnormal sample dataset satisfies the second error criterion; The switch cabinet feature operation parameters corresponding to the second target error value that satisfies the second error criterion in the current year's abnormal sample dataset are used as the second error parameters to construct the second largest error dataset; Compare whether the data ratio between the second largest error dataset and the current year's abnormal sample dataset is less than or equal to the preset first ratio threshold; If so, then the initial abnormal sample temperature rise early warning model shall be used as the target abnormal sample temperature rise early warning model; If not, then determine whether the data ratio between the second largest error dataset and the current year's abnormal sample dataset is less than a preset second ratio threshold. If so, the initial abnormal sample temperature rise early warning model is modified by weight matrix based on the current year's abnormal sample dataset to determine the target abnormal sample temperature rise early warning model.

9. The switchgear temperature rise early warning method according to claim 8, characterized in that, Also includes: If the data ratio between the second largest error dataset and the current year's abnormal sample dataset is greater than the preset second ratio threshold, then based on the switch cabinet temperature rise warning result, the linear fitting method is used to determine the warning parameter weight matrix according to the weight matrix of the target temperature rise warning model and the weight matrix of the initial temperature rise warning model. The target temperature rise early warning model is updated by using the aforementioned early warning parameter weight matrix to determine the temperature rise early warning model for future times. The future temperature rise early warning model is used to output multiple predicted temperature values ​​for third targets based on the current year's switchgear sample dataset; Determine whether all of the predicted temperature values ​​of the third target are less than or equal to a preset temperature threshold. If not, the preset reference threshold is updated based on the preset constant to determine a new preset reference threshold.

10. A switchgear temperature rise early warning system, characterized in that, include: The acquisition module is used to acquire historical switchgear sample datasets, preprocess the historical switchgear sample datasets, and output historical normal sample datasets and historical abnormal sample datasets. The construction module is used to construct an initial temperature rise early warning model based on the historical annual normal sample dataset and the historical annual abnormal sample dataset, according to the preset evaluation criteria. The correction module is used to correct the weight matrix of the initial temperature rise warning model by using a preset temperature error criterion and the current year's switchgear sample dataset when the current year's switchgear sample dataset is received, and output the target temperature rise warning model. The early warning module is used to generate a switchgear temperature rise early warning result based on the weight matrix of the target temperature rise early warning model and the weight matrix of the initial temperature rise early warning model using preset early warning criteria.

Images