Circuit breaker state discrimination method and system based on improved D-S evidence theory

Abstract

The invention relates to the technical field of circuit breaker state discrimination, in particular to a circuit breaker state discrimination method and system based on an improved D-S evidence theory. Comprising the following steps of collecting data and extracting features, using current feature data for support vector machine model training, using vibration feature data for decision tree model training, determining basic probability distribution of a D-S evidence theory through an ant colony algorithm, performing D-S evidence theory fusion on two models, testing the fused models and verifying the effect. According to the method, current signals and vibration signals are acquired to facilitate analysis and discrimination of various types of data, and the basic probability distribution of the D-S evidence theory is determined by the ant colony algorithm to improve the accuracy of evidence fusion; a circuit breaker state discrimination model based on a support vector machine and a circuit breaker state discrimination model based on a decision tree are fused through a D-S evidence theory, outputs of different models are reasonably fused, a more credible comprehensive result is generated according to the accuracy and confidence of the models, and the accuracy of the system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of circuit breaker state determination, and in particular to a circuit breaker state determination method and system based on improved DS evidence theory. Background Art

[0002] In the power system, circuit breakers are key devices used to protect electrical equipment from faults such as overload and short circuit. They are a necessary line of defense in substations, so ensuring the safety of circuit breakers is crucial. However, due to long-term operation, environmental factors and manufacturing defects, circuit breakers are prone to various types of failures, resulting in reduced reliability of the power system and even possible accidents and damage.

[0003] The state identification technology of the circuit breaker can timely detect and identify the fault state of the circuit breaker, which helps to prevent equipment damage and power system accidents, ensure personal safety and equipment safety, and ensure that the circuit breaker can act in time under fault conditions to prevent the fault from spreading. It is of great significance to maintain the stability and safety of the entire power system. For example, a method, system and device for determining the contact state of a gas circuit breaker contact disclosed in the invention patent with publication number CN115184788A uses a vibration signal emitted by the gas circuit breaker contact, and then converts the vibration signal into a fractal box dimension, and compares it with the fractal box dimension of the sample signal in the database, thereby determining the current state of the gas circuit breaker contact.

[0004] However, it is found during use that most of the current state judgment systems rely solely on a certain type of data or model, and the state judgment accuracy obtained is usually subject to various limitations, resulting in low judgment accuracy and poor practicality. Therefore, there is an urgent need for a circuit breaker state judgment method and system that improves the DS evidence theory to improve the above problems. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a method for obtaining current signals and vibration signals, which is convenient for analyzing and distinguishing through multiple types of data, and the ant colony algorithm determines the basic probability distribution of DS evidence theory to improve the accuracy, robustness and adaptability of evidence fusion. Furthermore, the DS evidence theory is used to fuse the circuit breaker state discrimination model based on support vector machine and the circuit breaker state discrimination model based on decision tree, which can reasonably fuse the outputs of different models, generate more reliable comprehensive results according to their accuracy and confidence, reduce the occurrence of erroneous discrimination, and improve the accuracy of the system. A circuit breaker state discrimination method and system with improved DS evidence theory.

[0006] A circuit breaker state determination method based on an improved DS evidence theory of the present invention comprises the following steps:

[0007] S1. Collect data and extract features:

[0008] Collect the continuous coil current signal and the vibration signal generated by the mechanical movement, and extract relevant characteristic data from the current signal and the vibration signal;

[0009] S2. Use current characteristic data for support vector machine model training:

[0010] After S1 obtains the current characteristic data, a circuit breaker state discrimination model based on a support vector machine is established, the collected current characteristic data is used as a training set for the model, the circuit breaker state is discriminated according to the current characteristic data, and the result of the model's discrimination of the circuit breaker state on the training set is obtained;

[0011] S3. Use vibration feature data for decision tree model training:

[0012] After S1 obtains the vibration feature data, a circuit breaker state discrimination model based on a decision tree is established, the extracted vibration feature data is used as a training set for the model, the circuit breaker state is discriminated according to the vibration feature data, and the result of the circuit breaker state discrimination of the model on the training set is obtained;

[0013] S4. Ant colony algorithm determines the basic probability distribution of DS evidence theory:

[0014] The DS evidence theory is used to fuse the circuit breaker state discrimination model based on support vector machine and the circuit breaker state discrimination model based on decision tree to obtain the fused circuit breaker discrimination model. The fusion effect based on DS evidence theory depends on the basic probability distribution. The global optimal basic probability distribution scheme is automatically found through multiple iterations of the ant colony algorithm. The ant colony algorithm can find the best probability value distribution method in the search space by simulating the process of ants finding the optimal path. Through multiple rounds of iterations, the ant colony algorithm gradually adjusts and optimizes the probability distribution, making the final fusion result more reliable. In addition, it has the characteristics of parallel search. Multiple ants can explore different solutions at the same time, which provides strong support for the search of the global optimal solution. The basic probability distribution of DS evidence theory is optimized by the ant colony algorithm, which effectively improves the accuracy, robustness and adaptability of evidence fusion.

[0015] S5. Combine the two models with DS evidence theory:

[0016] According to the basic probability distribution scheme generated by each iteration of the ant colony algorithm, multi-source data fusion is performed on the circuit breaker state discrimination model based on the support vector machine and the circuit breaker state discrimination model based on the decision tree. The comprehensive probability value is calculated according to the synthesis rule of the DS evidence theory, and the event or category with the maximum confidence is selected as the final decision result. The DS evidence theory effectively handles these uncertainties and improves the robustness of diagnosis by calculating the confidence and evidence fusion. Since different algorithms perform differently under different data characteristics or conditions, the DS evidence theory can reasonably fuse the outputs of different models to generate more reliable comprehensive results based on their accuracy and confidence, avoiding the limitations of relying on a single model. In addition, in actual circuit breaker state monitoring, sensors may experience noise, errors or failures. The DS evidence theory has strong fault tolerance and reduces the impact of a single erroneous judgment on the overall system through redundant information from different sources.

[0017] S6. Test the fused model and verify the effect:

[0018] After the completion of the previous five steps, a circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory was obtained. The circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory was tested by collecting the real-time current signal and vibration signal of the circuit breaker and extracting relevant features. After passing the test, the circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory was applied to use in a real environment.

[0019] Preferably, the DS evidence theory includes the following parts:

[0020] a1. Identification framework:

[0021] First, define the set of objects or events S = {a, b, c...}, a total of n mutually exclusive events, then the identification framework Total 2 n elements;

[0022] a2. Basic probability distribution:

[0023] DS evidence theory uses basic probability distribution to express the degree of trust in different hypotheses. Basic probability distribution is to distribute a total probability value to all possible hypotheses. Function m is used to represent the basic probability distribution at each time, and the relationship is satisfied:

[0024]

[0025] That is, the sum of the basic probability distributions of all events is 1;

[0026] a3. Synthesis rules:

[0027] The probability distribution value of the combined set is obtained by combining different events, satisfying the relationship:

[0028]

[0029] Among them, m A (a) and m B (b) is the probability value from different sources, and K is the conflict degree, which is used to measure the inconsistency or conflict in the synthesis process. Specifically, it is calculated based on the degree of conflict between evidence from different sources, satisfying the relationship:

[0030]

[0031] That is, it represents the sum of the combined probability values ​​of all mutually exclusive event sets.

[0032] Preferably, the process of determining the basic probability distribution of DS evidence theory by the ant colony algorithm includes the following steps:

[0033] b1. Initialization:

[0034] Set parameters such as the number of ants, maximum number of iterations, number of events, and initialize the probability value array to store the probability value of each event;

[0035] b2. Iteration process:

[0036] Each iteration is performed by all ants simultaneously. Each ant evaluates the available data, selects the current algorithm model for judgment, and updates the probability value after obtaining the judgment result. According to the judgment result and the true result label, the trust of the corresponding event is increased accordingly.

[0037] b3. Normalization:

[0038] After each iteration, the probability values ​​need to be normalized to ensure that their sum is 1, that is, to meet the definition of basic probability distribution;

[0039] b4. Output:

[0040] The optimized basic probability distribution results are returned to the DS evidence theory fusion model for multi-source data fusion.

[0041] Preferably, the method of automatically finding the global optimal basic probability allocation scheme through multiple iterations of the ant colony algorithm in S4 is:

[0042] Each ant individual will find a basic probability allocation scheme, substitute this scheme into the multi-source data fusion of DS evidence theory, and obtain the discrimination result of the fused circuit breaker discrimination model on the training set. The scheme will be evaluated according to the discrimination result, and the probability value will be updated. Through multiple multi-individual iterations, the global optimal basic probability allocation scheme can be quickly obtained.

[0043] Preferably, the specific method of testing the fused model and verifying the effect in S6 includes:

[0044] The real-time current signal and vibration signal of the circuit breaker are collected and relevant features are extracted. The fusion model is tested as a test set to obtain the result of the fusion model on the circuit breaker state discrimination. At the same time, the circuit breaker state discrimination model based on support vector machine and the circuit breaker state discrimination model based on decision tree are tested, and the circuit breaker discrimination results corresponding to the two models are obtained respectively. The results are compared and analyzed with the circuit breaker state discrimination model results based on ant colony algorithm to improve DS evidence theory. After the test, the circuit breaker state discrimination model based on ant colony algorithm to improve DS evidence theory is applied to use in real environment.

[0045] Preferably, the current signal is collected from a current sensor, and the vibration signal is collected from an acceleration sensor;

[0046] The current peak value and the current gradient difference are extracted from the current signal.

[0047] A circuit breaker state determination system based on an improved DS evidence theory of the present invention comprises:

[0048] Data acquisition module: collects continuous coil current signals from the current sensor and collects vibration signals generated by mechanical motion from the acceleration sensor;

[0049] Data processing module: extract relevant feature data from the data;

[0050] Data analysis module: The data is analyzed and processed by using the circuit breaker status discrimination model based on the ant colony algorithm to improve the DS evidence theory, and the status of the circuit breaker is discriminated.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. The ant colony algorithm determines the basic probability distribution of DS evidence theory to improve the accuracy, robustness and adaptability of evidence fusion;

[0053] 2. By integrating the circuit breaker state discrimination model based on support vector machine and the circuit breaker state discrimination model based on decision tree through DS evidence theory, the outputs of different models can be reasonably integrated to generate more reliable comprehensive results based on their accuracy and confidence, thereby reducing the occurrence of erroneous discrimination and improving the accuracy of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a schematic diagram of the flow structure of the circuit breaker state determination method based on the improved DS evidence theory of the present invention;

[0055] Figure 2 This is a comparison chart of the test results of three types of models of the test set data of the present invention;

[0056] Figure 3 It is a structural schematic diagram of the circuit breaker state judgment system based on the improved DS evidence theory of the present invention. DETAILED DESCRIPTION

[0057] In order to facilitate understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. The present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0058] like Figure 1 and Figure 2 As shown, a circuit breaker state determination method and system based on improved DS evidence theory includes the following steps:

[0059] S1. Collect data and extract features:

[0060] Collect the continuous coil current signal and the vibration signal generated by the mechanical movement, and extract relevant characteristic data from the current signal and the vibration signal;

[0061] S2. Use current characteristic data for support vector machine model training:

[0062] After S1 obtains the current characteristic data, a circuit breaker state discrimination model based on a support vector machine is established, the collected current characteristic data is used as a training set for the model, the circuit breaker state is discriminated according to the current characteristic data, and the result of the model's discrimination of the circuit breaker state on the training set is obtained;

[0063] S3. Use vibration feature data for decision tree model training:

[0064] After S1 obtains the vibration feature data, a circuit breaker state discrimination model based on a decision tree is established, the extracted vibration feature data is used as a training set for the model, the circuit breaker state is discriminated according to the vibration feature data, and the result of the circuit breaker state discrimination of the model on the training set is obtained;

[0065] S4. Ant colony algorithm determines the basic probability distribution of DS evidence theory:

[0066] The DS evidence theory is used to fuse the circuit breaker state discrimination model based on support vector machine and the circuit breaker state discrimination model based on decision tree to obtain the fused circuit breaker discrimination model. The fusion effect based on DS evidence theory depends on the basic probability distribution. The global optimal basic probability distribution scheme is automatically found through multiple iterations of the ant colony algorithm. The ant colony algorithm can find the best probability value distribution method in the search space by simulating the process of ants finding the optimal path. Through multiple rounds of iterations, the ant colony algorithm gradually adjusts and optimizes the probability distribution, making the final fusion result more reliable. In addition, it has the characteristics of parallel search. Multiple ants can explore different solutions at the same time, which provides strong support for the search of the global optimal solution. The basic probability distribution of DS evidence theory is optimized by the ant colony algorithm, which effectively improves the accuracy, robustness and adaptability of evidence fusion.

[0067] S5. Combine the two models with DS evidence theory:

[0068] According to the basic probability distribution scheme generated by each iteration of the ant colony algorithm, multi-source data fusion is performed on the circuit breaker state discrimination model based on the support vector machine and the circuit breaker state discrimination model based on the decision tree. The comprehensive probability value is calculated according to the synthesis rule of the DS evidence theory, and the event or category with the maximum confidence is selected as the final decision result. The DS evidence theory effectively handles these uncertainties and improves the robustness of diagnosis by calculating the confidence and evidence fusion. In addition, since different algorithms (such as support vector machines and decision trees) perform differently under different data features or conditions, the DS evidence theory is used to reasonably fuse the outputs of different models, and generate more reliable comprehensive results based on their accuracy and confidence, avoiding the limitations of relying on a single model. In addition, in actual circuit breaker state monitoring, sensors may have noise, errors or failures. The DS evidence theory has strong fault tolerance and reduces the impact of a single erroneous judgment on the overall system through redundant information from different sources.

[0069] S6. Test the fused model and verify the effect:

[0070] After the previous five steps, the circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory is obtained. The circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory is tested by collecting the real-time current signal and vibration signal of the circuit breaker and extracting relevant features. After passing the test, the circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory is applied to the real environment for use;

[0071] The DS evidence theory includes the following parts:

[0072] a1. Identification framework:

[0073] First, define the set of objects or events S = {a, b, c...}, a total of n mutually exclusive events, then the identification framework Total 2 n elements;

[0074] a2. Basic probability distribution:

[0075] DS evidence theory uses basic probability distribution to express the degree of trust in different hypotheses. Basic probability distribution is to distribute a total probability value to all possible hypotheses. Function m is used to represent the basic probability distribution at each time, and the relationship is satisfied:

[0076]

[0077] That is, the sum of the basic probability distributions of all events is 1;

[0078] a3. Synthesis rules:

[0079] The probability distribution value of the combined set is obtained by combining different events, satisfying the relationship:

[0080]

[0081] Among them, m A (a) and m B (b) is the probability value from different sources, and K is the conflict degree, which is used to measure the inconsistency or conflict in the synthesis process. Specifically, it is calculated based on the degree of conflict between evidence from different sources, satisfying the relationship:

[0082]

[0083] That is, it represents the sum of the combined probability values ​​of all mutually exclusive event sets;

[0084] The process of determining the basic probability distribution of DS evidence theory by the ant colony algorithm includes the following steps:

[0085] b1. Initialization:

[0086] Set parameters such as the number of ants, maximum number of iterations, number of events, and initialize the probability value array to store the probability value of each event;

[0087] b2. Iteration process:

[0088] Each iteration is performed by all ants simultaneously. Each ant evaluates the available data, selects the current algorithm model (support vector machine or decision tree) for judgment, and updates the probability value after obtaining the judgment result. According to the judgment result and the true result label, the trust of the corresponding event is increased accordingly.

[0089] b3. Normalization:

[0090] After each iteration, the probability values ​​need to be normalized to ensure that their sum is 1, that is, to meet the definition of basic probability distribution;

[0091] b4. Output:

[0092] Return the optimized basic probability distribution results to the DS evidence theory fusion model to perform multi-source data fusion;

[0093] The method of automatically finding the global optimal basic probability allocation scheme through multiple iterations of the ant colony algorithm in S4 is:

[0094] Each ant individual will find a basic probability allocation scheme, substitute this scheme into the multi-source data fusion of DS evidence theory, and obtain the discrimination result of the fused circuit breaker discrimination model on the training set. According to the discrimination result, the scheme is evaluated and the probability value is updated. Through multiple multi-individual iterations, the global optimal basic probability allocation scheme can be quickly obtained.

[0095] The specific method for testing the fused model and verifying the effect in S6 includes:

[0096] The real-time current signal and vibration signal of the circuit breaker are collected and relevant features are extracted. The fusion model is tested as a test set to obtain the result of the fusion model on the circuit breaker state discrimination. At the same time, the circuit breaker state discrimination model based on support vector machine and the circuit breaker state discrimination model based on decision tree are tested to obtain the circuit breaker discrimination results corresponding to the two models. The results are compared and analyzed with the circuit breaker state discrimination model results based on ant colony algorithm to improve DS evidence theory. After the test, the circuit breaker state discrimination model based on ant colony algorithm to improve DS evidence theory is applied to the real environment for use.

[0097] The current signal is collected from a current sensor, and the vibration signal is collected from an acceleration sensor;

[0098] The current peak value and the current gradient difference are extracted from the current signal;

[0099] include:

[0100] Data acquisition module: collects continuous coil current signals from the current sensor and collects vibration signals generated by mechanical motion from the acceleration sensor;

[0101] Data processing module: extract relevant feature data from the data;

[0102] Data analysis module: The data is analyzed and processed by using the circuit breaker status discrimination model based on the ant colony algorithm to improve the DS evidence theory, and the status of the circuit breaker is discriminated.

[0103] Example

[0104] In order to reflect the superiority of the circuit breaker state discrimination method based on the ant colony algorithm to improve the DS evidence theory, by extracting 15 pieces of current feature data and vibration feature data corresponding to the four states of the circuit breaker as training data sets, the circuit breaker state discrimination model based on support vector machine and the circuit breaker state discrimination model based on decision tree are trained; then the DS evidence theory is improved based on the ant colony algorithm to obtain the circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory; then, 15 pieces of current feature data and vibration feature data corresponding to the four states of the circuit breaker are extracted again as test sets. The results show that the accuracy of the circuit breaker state discrimination model based on support vector machine is 68.3%, the accuracy of the circuit breaker state discrimination model based on decision tree is 55%, and the accuracy of the circuit breaker state discrimination model based on ant colony algorithm to improve the DS evidence theory is 75%, as shown in Table 1:

[0105] Table 1 Accuracy results of each model in the test set

[0106]

[0107] It can be seen that the overall accuracy of the circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory is significantly higher than that of the other two models. This is because through the DS evidence theory, the outputs of different models can be reasonably integrated, and more reliable comprehensive results can be generated based on their accuracy and confidence, which effectively improves the operating stability and maintenance efficiency of electrical equipment.

[0108] like Figure 3 As shown in the figure, from the overall experimental results, it can be seen that the circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory has the highest overall discrimination accuracy under various circuit breaker states. Therefore, it can be concluded that the circuit breaker state discrimination method based on the ant colony algorithm to improve the DS evidence theory has improved the robustness and stability of the circuit breaker state discrimination to a certain extent, and has improved the accuracy of the circuit breaker state discrimination while ensuring its computational efficiency.

[0109] In the present invention, collecting data and extracting features are prerequisites required by the present invention, but are not features of the present application, so they will not be introduced in detail, and the support vector machine model based on current feature data and the decision tree model based on vibration feature data are only used to reflect the performance improvement brought about by the improved DS evidence theory based on the ant colony algorithm. Therefore, the support vector machine model and the decision tree are both basic tools required by the present invention rather than features of the present application. The improved DS evidence theory based on the ant colony algorithm proposed in the present invention can also be applied to other algorithm models.

[0110] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A circuit breaker state determination method based on improved DS evidence theory, characterized in that: The following steps are involved: S1. Collect data and extract features: Collect the continuous coil current signal and the vibration signal generated by the mechanical movement, and extract relevant characteristic data from the current signal and the vibration signal; S2. Use current characteristic data for support vector machine model training: After S1 obtains the current characteristic data, a circuit breaker state discrimination model based on a support vector machine is established, the collected current characteristic data is used as a training set for the model, the circuit breaker state is discriminated according to the current characteristic data, and the result of the model's discrimination of the circuit breaker state on the training set is obtained; S3. Use vibration feature data for decision tree model training: After S1 obtains the vibration feature data, a circuit breaker state discrimination model based on a decision tree is established, the extracted vibration feature data is used as a training set for the model, the circuit breaker state is discriminated according to the vibration feature data, and the result of the circuit breaker state discrimination of the model on the training set is obtained; S4. Ant colony algorithm determines the basic probability distribution of DS evidence theory: The DS evidence theory is used to fuse the circuit breaker state discrimination model based on support vector machine and the circuit breaker state discrimination model based on decision tree to obtain the fused circuit breaker discrimination model. The fusion effect based on DS evidence theory depends on the basic probability distribution. The global optimal basic probability distribution scheme is automatically found through multiple iterations of the ant colony algorithm. The ant colony algorithm can find the best probability value distribution method in the search space by simulating the process of ants finding the optimal path. Through multiple rounds of iterations, the ant colony algorithm gradually adjusts and optimizes the probability distribution, making the final fusion result more reliable. In addition, it has the characteristics of parallel search. Multiple ants can explore different solutions at the same time, which provides strong support for the search of the global optimal solution. The basic probability distribution of DS evidence theory is optimized by the ant colony algorithm, which effectively improves the accuracy, robustness and adaptability of evidence fusion. S5. Combine the two models with DS evidence theory: According to the basic probability distribution scheme generated by each iteration of the ant colony algorithm, multi-source data fusion is performed on the circuit breaker state discrimination model based on the support vector machine and the circuit breaker state discrimination model based on the decision tree. The comprehensive probability value is calculated according to the synthesis rule of the DS evidence theory, and the event or category with the maximum confidence is selected as the final decision result. The DS evidence theory effectively handles these uncertainties and improves the robustness of diagnosis by calculating the confidence and evidence fusion. Since different algorithms perform differently under different data characteristics or conditions, the DS evidence theory can reasonably fuse the outputs of different models to generate more reliable comprehensive results based on their accuracy and confidence, avoiding the limitations of relying on a single model. In addition, in actual circuit breaker state monitoring, sensors may experience noise, errors or failures. The DS evidence theory has strong fault tolerance and reduces the impact of a single erroneous judgment on the overall system through redundant information from different sources. S6. Test the fused model and verify the effect: After the completion of the previous five steps, a circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory was obtained. The circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory was tested by collecting the real-time current signal and vibration signal of the circuit breaker and extracting relevant features. After passing the test, the circuit breaker state discrimination model based on the ant colony algorithm to improve the DS evidence theory was applied to use in a real environment.

2. A circuit breaker state determination method based on improved DS evidence theory as claimed in claim 1, characterized in that: The DS evidence theory includes the following parts: a1. Identification framework: First, define the set of objects or events S = {a, b, c...}, a total of n mutually exclusive events, then the identification framework Total 2 n elements; a2. Basic probability distribution: DS evidence theory uses basic probability distribution to express the degree of trust in different hypotheses. Basic probability distribution is to distribute a total probability value to all possible hypotheses. Function m is used to represent the basic probability distribution at each time, and the relationship is satisfied: That is, the sum of the basic probability distributions of all events is 1; a3. Synthesis rules: The probability distribution value of the combined set is obtained by combining different events, satisfying the relationship: Among them, m A (a) and m B (b) is the probability value from different sources, and K is the conflict degree, which is used to measure the inconsistency or conflict in the synthesis process. Specifically, it is calculated based on the degree of conflict between evidence from different sources, satisfying the relationship: That is, it represents the sum of the combined probability values ​​of all mutually exclusive event sets.

3. A circuit breaker state determination method based on improved DS evidence theory as claimed in claim 1, characterized in that: The process of determining the basic probability distribution of DS evidence theory by the ant colony algorithm includes the following steps: b1. Initialization: Set parameters such as the number of ants, maximum number of iterations, number of events, and initialize the probability value array to store the probability value of each event; b2. Iteration process: Each iteration is performed by all ants simultaneously. Each ant evaluates the available data, selects the current algorithm model for judgment, and updates the probability value after obtaining the judgment result. According to the judgment result and the true result label, the trust of the corresponding event is increased accordingly. b3. Normalization: After each iteration, the probability values ​​need to be normalized to ensure that their sum is 1, that is, to meet the definition of basic probability distribution; b4. Output: The optimized basic probability distribution results are returned to the DS evidence theory fusion model for multi-source data fusion.

4. A circuit breaker state determination method based on improved DS evidence theory as claimed in claim 1, characterized in that: The method of automatically finding the global optimal basic probability allocation scheme through multiple iterations of the ant colony algorithm in S4 is: Each ant individual will find a basic probability allocation scheme, substitute this scheme into the multi-source data fusion of DS evidence theory, and obtain the discrimination result of the fused circuit breaker discrimination model on the training set. The scheme will be evaluated according to the discrimination result, and the probability value will be updated. Through multiple multi-individual iterations, the global optimal basic probability allocation scheme can be quickly obtained.

5. A circuit breaker state determination method based on improved DS evidence theory as claimed in claim 1, characterized in that: The specific method for testing the fused model and verifying the effect in S6 includes: The real-time current signal and vibration signal of the circuit breaker are collected and relevant features are extracted. The fusion model is tested as a test set to obtain the result of the fusion model on the circuit breaker state discrimination. At the same time, the circuit breaker state discrimination model based on support vector machine and the circuit breaker state discrimination model based on decision tree are tested, and the circuit breaker discrimination results corresponding to the two models are obtained respectively. The results are compared and analyzed with the circuit breaker state discrimination model results based on ant colony algorithm to improve DS evidence theory. After the test, the circuit breaker state discrimination model based on ant colony algorithm to improve DS evidence theory is applied to use in real environment.

6. A circuit breaker state determination method based on improved DS evidence theory as claimed in claim 1, characterized in that: The current signal is collected from the current sensor, and the vibration signal is collected from the acceleration sensor; The current peak value and the current gradient difference are extracted from the current signal.

7. A circuit breaker state judgment system based on improved DS evidence theory, characterized in that: include: Data acquisition module: collects continuous coil current signals from the current sensor and collects vibration signals generated by mechanical motion from the acceleration sensor; Data processing module: extract relevant feature data from the data; Data analysis module: The data is analyzed and processed by using the circuit breaker status discrimination model based on the ant colony algorithm to improve the DS evidence theory, and the status of the circuit breaker is discriminated.

Images