Power distribution network fault positioning model optimization method and related device

Abstract

The invention provides a power distribution network fault positioning model optimization method and a related device, and the method comprises the steps: firstly, obtaining a model evaluation data set of a first model; the first model is a power distribution network fault positioning model; then, based on the model evaluation data set, credibility information of the first model is determined; the credibility information comprises process credibility, result credibility and interpretable credibility; and finally, based on the credibility information, performing optimization processing on the first model to obtain a second model. Visibly, by constructing a three-dimensional credibility quantification system of'process + result + interpretability ', namely process credibility, result credibility and interpretability credibility, the system performs three-dimensional credibility evaluation on each power distribution network fault positioning and drives hierarchical decision, so that the accuracy, efficiency and transparency of power distribution network fault positioning are comprehensively improved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method and related apparatus for optimizing a fault location model in a power distribution network. Background Technology

[0002] Due to the complex terrain and wide distribution of power distribution lines in mountainous areas, fault location in power distribution networks relies heavily on manual inspection, which is inefficient and costly. In recent years, neural network technology has been introduced into the field of fault location, but its "black box" nature leads to a lack of user trust in the results. If a fault is located incorrectly, the entire line must be re-inspected, incurring significant trial-and-error costs. Current technologies lack quantitative evaluation methods for assessing the trustworthiness of neural network outputs, resulting in low user acceptance of intelligent fault location results. Summary of the Invention

[0003] This application provides a method and related apparatus for optimizing a distribution network fault location model, aiming to improve the accuracy and efficiency of the financial service system in optimizing the distribution network fault location model.

[0004] In a first aspect, embodiments of this application provide a method for optimizing a distribution network fault location model, including: Obtain the model evaluation dataset for the first model; the first model is a distribution network fault location model. Based on the model evaluation dataset, the trust information of the first model is determined; the trust information includes process trust, outcome trust, and interpretability trust. Based on trust information, the first model is optimized to obtain the second model.

[0005] Secondly, embodiments of this application provide a distribution network fault location model optimization device, the device comprising: The acquisition unit acquires the model evaluation dataset of the first model; the first model is a distribution network fault location model. The determination unit is used to determine the trust information of the first model based on the model evaluation dataset; the trust information includes process trust, outcome trust, and interpretability trust. The optimization unit is used to optimize the first model based on trust information to obtain the second model.

[0006] Thirdly, embodiments of this application provide an electronic device including a processor, a memory, and one or more programs, the one or more programs being stored in the memory and configured to be executed by the processor, the programs including instructions for performing the steps of the method as described in the first aspect of embodiments of this application.

[0007] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program / instructions stored thereon, wherein the computer program / instructions, when executed by a processor, implement the steps of the method of the first aspect of embodiments of this application.

[0008] Fifthly, embodiments of this application provide a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the method of the first aspect of embodiments of this application.

[0009] As can be seen, in this embodiment, a model evaluation dataset of the first model is obtained; the first model is a distribution network fault location model; then, based on the model evaluation dataset, the confidence information of the first model is determined; the confidence information includes process confidence, result confidence, and interpretability confidence; finally, based on the confidence information, the first model is optimized to obtain the second model. It is evident that by constructing a three-dimensional confidence quantification system of "process + result + interpretability," namely process confidence, result confidence, and interpretability confidence, this system drives hierarchical decision-making by conducting a three-dimensional confidence assessment of each distribution network fault location, thereby comprehensively improving the accuracy, efficiency, and transparency of distribution network fault location. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in this application 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart illustrating a method for optimizing a distribution network fault location model, as provided in an embodiment of this application. Figure 2 This is a schematic diagram illustrating an interpretable trust level definition provided in an embodiment of this application; Figure 3 This is a quantification graph of an interpretable trust level definition provided in an embodiment of this application; Figure 4 This is a functional unit block diagram of a power distribution network fault location model optimization device provided in an embodiment of this application; Figure 5 This is a structural block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0012] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0014] It should be noted that the terms "first," "second," "third," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the term "comprising" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or electronic device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0015] This application embodiment constructs a three-dimensional trust quantification system of "process + result + explainability", namely process trust, result trust and explainability trust. This system drives hierarchical decision-making by conducting three-dimensional credibility assessment of each distribution network fault location, thereby comprehensively improving the accuracy, efficiency and transparency of distribution network fault location.

[0016] Please see Figure 1 , Figure 1 This is a flowchart illustrating a method for optimizing a distribution network fault location model, as provided in an embodiment of this application. The method is applied to, for example... Figure 1 As shown, the method includes the following steps S101-S103: S101. Obtain the model evaluation dataset for the first model.

[0017] The first model is a distribution network fault location model; the model evaluation dataset includes: the operation data of the first model, user feedback data, and historical benchmark data; the operation data includes the fault data of the distribution network and the fault measurement results generated by the first model; the user feedback data is the feedback from users on the fault measurement results; the historical benchmark data includes historical fault location data before the introduction of the first model; the model evaluation dataset also includes the actual fault results corresponding to the fault measurement results.

[0018] S102. Based on the model evaluation dataset, determine the trust information of the first model; the trust information includes process trust, outcome trust, and interpretability trust.

[0019] S103. Based on trust information, the first model is optimized to obtain the second model.

[0020] Trust data collection and calculation: User feedback is quantified into interpretable trust level, and the accuracy and timeliness of distribution network fault location are used to obtain performance trust level, thus constructing a trust level assessment dataset. The information gain formula is used to calculate the gain value of each decision layer.

[0021] I. Timeliness (Process Trustworthiness) Among them, timeliness is an evaluation indicator for the entire process, which measures the response time Δt=t2-t1 from the start of fault data processing (t1) to the final location result output (t2).

[0022] 1. The trust calculation of "timeliness" (process trust) is divided into two levels.

[0023] The first level: Before and after introducing the model, compare the timeliness of fault location to obtain the "timeliness rate" (first process trust level).

[0024] The second level: After introducing the model, compare the timeliness of the previous model's fault location judgment with the timeliness of the model's fault location judgment after optimization based on the overall trust level obtained from the previous model, and then obtain the "timeliness rate" (second process trust level).

[0025] In other words, historical baseline data includes the first timeliness rate of the fault location process before the introduction of the first model; operational data includes the second timeliness rate of the fault location process after the introduction of the first model; operational data includes the third timeliness rate of the fault location process before the current optimization of the first model; and the fourth timeliness rate of the fault location process after the current optimization of the first model. Trust information includes the first process trust and / or the second process trust. Based on the model evaluation dataset, the trust information of the first model is determined, including: determining the first-level timeliness rate based on the difference or ratio between the second and first timeliness rates, where the first-level timeliness rate is the first process trust; and / or, determining the second-level timeliness rate based on the difference or ratio between the fourth and third timeliness rates, where the second-level timeliness rate is the second process trust.

[0026] 2. The calculation of response time before introducing the model needs to be adapted to local conditions, as the development of distribution networks varies in different regions. In some less developed areas, traditional methods are still used for fault location. The response time calculation is as follows: from fault detection and alarm (a line fault causes the circuit breaker to trip, and the substation monitoring system issues an alarm) → manual inspection (the specific environment needs to be considered, such as in mountainous areas, forest areas, at night, or in severe weather) → completion of fault location confirmation. In some areas where the distribution network is developing rapidly, the calculation can be based on the time required for fault location based on the equipment already introduced at this stage.

[0027] 3. ① After introducing the model, it is divided into two periods. The first period requires going through the entire process of fault location determination (from real-time data processing, network input, training, output, to the response time of the entire process of determining fault location). During this period, the model needs to be continuously trained and optimized in the field to achieve the effect of adapting to local conditions. In the second period, the "training" step can be skipped and the output can be directly generated in the stage of determining the fault location of the distribution network. However, this does not mean that "training" is not required in this period. Instead, post-training review is conducted to optimize the model and achieve the effect of continuous optimization.

[0028] ② The judgment model for distribution network fault location differs from general models. It requires training not only before deployment but also after deployment. This is because the surrounding environment varies in different regions (specific environments need to be considered, such as mountainous areas, forest areas, nighttime, or severe weather). The model needs to be trained according to local conditions to develop a model suitable for a particular region, thus enabling better distribution network fault location.

[0029] II. Result Trust Measurement Parameters: Accuracy (Performance Trust) The evaluation metrics are results-oriented, considering the accuracy of the final fault location result. Throughout the entire application of fault location in power distribution lines, the accuracy of fault location results before and after the comparison is performed.

[0030] In some possible implementations, the running data includes fault measurement results generated by the first model; the model evaluation dataset also includes the actual fault results corresponding to the fault measurement results; the confidence information includes the result confidence; based on the model evaluation dataset, the confidence information of the first model is determined, including: constructing a confusion matrix based on the matching relationship between the fault measurement results and the corresponding actual fault results to obtain four types of sample data; the four types of sample data include true positives (TP), false negatives (FN), false positives (FP), and true negatives (TN); wherein, true positives (TP) are samples that are actually fault data and predicted as fault data; false negatives (FN) are samples that are actually normal data and predicted as fault data; false positives (FP) are samples that are actually normal data and predicted as normal data; true negatives (TN) are samples that are actually normal data and predicted as normal data; the accuracy is obtained based on the ratio between the first sum and the second sum; the accuracy is the result confidence; the first sum is the sum between the number of true positives and the number of true negatives; the second sum is the sum between the number of true positives, the number of false negatives, the number of false positives, and the number of true negatives.

[0031] First, a confusion matrix is ​​used to assess the accuracy of the final results of fault location and fault prediction.

[0032]

[0033] Note: TP represents a fault data sample predicted as fault data; FN represents a normal data sample predicted as fault data; FP represents a normal data sample predicted as normal data; TN represents a normal data sample predicted as normal data.

[0034] Using accuracy as an outcome-oriented evaluation metric:

[0035] We hope the accuracy rate will be close to 100%, the higher the better.

[0036] In some possible implementations, determining the confidence information of the first model based on the model evaluation dataset further includes: simultaneously calculating the false positive rate and the false negative rate, using them as auxiliary evaluation indicators of the result confidence; the false positive rate is the number of false negatives divided by the total number of false negatives, false positives, and true negatives, representing the proportion of actually normal data that is predicted as faulty data out of all actual normal data; the false negative rate is the number of actually faulty data that is predicted as normal data divided by the number of true positives, representing the proportion of actually faulty data that is predicted as normal out of all actual faulty data; the false positive rate and the false negative rate are used as optimization targets during model training to help improve the result confidence.

[0037] In other words, the false alarm rate and the missed alarm rate can also be used to assess the situation of missed alarms and false alarms: ;

[0038] We hope for the lowest possible false alarm and missed alarm rates. However, considering the special security requirements of the distribution network, it is better to have false alarms than false alarms.

[0039] The false alarm rate and false alarm rate are training indicators for improving accuracy (performance reliability) and are part of the training of the entire distribution network fault location model. Improving the false alarm rate and false alarm rate is also an optimization of the model.

[0040] Fault location accuracy before and after improvement: improved by at least 5%. Final accuracy: not less than 99%.

[0041] 3. Explainability and Trustworthiness: The fault measurement results output by the neural network are hierarchically labeled using a decision tree (with user feedback as input data).

[0042] In some possible implementations, the acquired user feedback data is user feedback on the fault measurement results; the feedback includes the user's positive or negative trust judgment of the results and the reasons supporting that judgment; based on the model evaluation dataset, the trust information of the first model is determined, including: using the user's trust judgment of the results as a first-level decision label; using the reasonableness judgment of the reasons provided by the user as a second-level decision label; classifying feedback labeled as reasonable positive trust and reasonable negative trust as trust, and classifying the remaining combinations as distrust; calculating the information gain value based on the distribution of the trust and distrust classification results; the information gain value is the interpretable trust level.

[0043] refer to Figure 2 and Figure 3 , Figure 2 and Figure 3 An interpretable definition and quantitative illustration of trust level are given for applications in mountainous power distribution networks.

[0044] First, for the interpretable application results of neural networks in mountainous power distribution networks, the decision tree method is used to quantify the aforementioned level of trust: First, the decision branches for the level of trust brought to users are quantified. For each layer of decision, the left branch is marked as "1" and the right branch is marked as "0".

[0045] Secondly, collect the users' explainable decision results and label "reasonable positive trust" and "reasonable negative trust" as "positive trust," and "unreasonable positive trust" and "unreasonable negative trust" as "negative trust," as shown in Table 1 below: Collect users' "positive trust" or "negative trust" for using this model for distribution network fault location, and the corresponding reasons. Specifically: ① If the user chooses positive trust, the first-level decision is 1; if the user's reason for trust is reasonable, the second-level decision is 1; otherwise, if the user does not provide a reasonable reason, or the reason provided is unreasonable, the second-level decision is 0. ② If the user chooses negative trust, the first-level decision is 0; if the user's reason for trust is reasonable, the second-level decision is 1; otherwise, if the user does not provide a reasonable reason, or the reason provided is unreasonable, the second-level decision is 0.

[0046] Table 1 shows an example of users' interpretable decision results:

[0047] Then, using the information gain formula in the decision tree method, the information gain value of each layer is calculated sequentially until the root node is reached. , , , : ; in, The entropy represents the information entropy of a random variable X, which measures the degree of uncertainty or disorder of the random variable X; the summation symbol ensures that the calculation of entropy takes all possibilities into account. It represents the probability that a random variable X will take a specific value x; Calculate probability The base-2 logarithm of x represents the amount of information carried when event x occurs.

[0048] This refers to reasonable positive trust. This refers to unreasonable positive trust. This refers to reasonable negative trust. It refers to unreasonable negative trust; trust is positive trust, and negative trust is distrust.

[0049] Overall understanding: Information entropy It is the expected value (weighted average) of the information content of all possible events of the random variable X.

[0050] Finally, the sum of the information gain values ​​corresponding to "reasonable positive trust" and "reasonable negative trust" is used as a trust index to measure the degree of trust brought about by explainability: This result is the sum of the information gain values ​​corresponding to "reasonable positive trust" and "reasonable negative trust", which is the trust index of "interpretable trust" and is part of the training of the entire distribution network fault location model.

[0051] expect Nearly 100%, The higher the better.

[0052] In this application, the first model is optimized based on trust information to obtain a second model, including: calculating a first information gain value based on interpretability trust; performing preliminary optimization on the first model based on the first information gain value to obtain a third model; calculating a second information gain value based on the output, process trust, and result trust of the third model; and further optimizing the third model based on the second information gain value to obtain the second model.

[0053] Consistent with the embodiments shown above, such as Figure 4 As shown, Figure 4 This is a functional unit block diagram of a power distribution network fault location model optimization device provided in an embodiment of this application. Figure 4 In this context, the distribution network fault location model optimization device 400 is applied to electronic equipment, and the distribution network fault location model optimization device 410 includes: Unit 411 acquires the model evaluation dataset of the first model; the first model is a distribution network fault location model. The determining unit 412 is used to determine the trust information of the first model based on the model evaluation dataset; the trust information includes process trust, result trust, and interpretability trust. The optimization unit 413 is used to optimize the first model based on the trust information to obtain the second model.

[0054] Based on the same inventive concept, the principle and beneficial effects of the power grid fault location model optimization device provided in the embodiments of this application can be found in the principle and beneficial effects of the power grid fault location model optimization method, which will not be repeated here for the sake of brevity.

[0055] Figure 5 This is a structural block diagram of an electronic device provided in an embodiment of this application. For example... Figure 5As shown, the electronic device 500 may include one or more of the following components: a processor 501 and a memory 502 coupled to the processor 501, wherein the memory 502 may store one or more computer programs, which may be configured to perform the following operations when executed by one or more processors 501: acquiring a model evaluation dataset of a first model; the first model is a distribution network fault location model; determining the confidence information of the first model based on the model evaluation dataset; the confidence information includes process confidence, result confidence, and interpretability confidence; and optimizing the first model based on the confidence information to obtain a second model.

[0056] Based on the same inventive concept, the principle and beneficial effects of the electronic equipment provided in the embodiments of this application in solving the problem can be found in the principle and beneficial effects of the power distribution network fault location model optimization method, which will not be repeated here for the sake of brevity.

[0057] This application also provides a computer-readable storage medium (Memory), which is a memory device of an electronic device used to store programs and data. It is understood that the computer-readable storage medium here may include the electronic device or its built-in storage medium, or it may include the electronic device or its supported extended storage medium. The computer-readable storage medium provides storage space that stores the electronic device or its operating system. Furthermore, the storage space also stores one or more computer programs suitable for being loaded and executed by a processor. It should be noted that the computer storage medium here may be high-speed RAM or non-volatile memory, such as at least one disk storage device; optionally, it may also be at least one computer-readable storage medium located remotely from the aforementioned processor.

[0058] Based on the same inventive concept, the principle and beneficial effects of the computer-readable storage medium provided in the embodiments of this application in solving the problem can be found in the principle and beneficial effects of the distribution network fault location model optimization method, which will not be repeated here for the sake of brevity.

[0059] The aforementioned computer-readable storage medium can be an electronic device provided in any of the foregoing embodiments or an internal storage unit of the aforementioned computer device, such as a hard disk or memory of the computer device. The computer-readable storage medium can also be an external storage device of the computer device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on the computer device. Furthermore, the computer-readable storage medium may include both internal storage units and external storage devices of the computer device. The computer-readable storage medium is used to store the computer program and other programs and data required by the computer device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0060] This application provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the aforementioned... Figure 3 The description of the microcontroller unit (MCU) program control method in the corresponding embodiments will not be repeated here. Furthermore, the beneficial effects of using the same method will also not be repeated. For technical details not disclosed in the computer-readable storage medium embodiments related to this application, please refer to the description of the method embodiments of this application.

[0061] The terms "first," "second," etc., in the specification, claims, and drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the term "comprising," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or modules, but may optionally include steps or modules not listed, or may optionally include other step units inherent to these processes, methods, apparatuses, products, or devices.

[0062] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0063] The methods and related apparatus provided in this application are described with reference to the method flowcharts and / or structural diagrams provided in this application. Specifically, each block of the method flowchart and / or structural diagram, as well as combinations of blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable electronic device to produce a machine, such that the instructions executable by the processor of the computer or other programmable electronic device generate instructions for implementing the process. Figure 1 A schematic diagram of one or more processes and / or structures. Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable electronic device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 A schematic diagram of one or more processes and / or structures. Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable electronic device to cause a series of operational steps to be performed on the computer or other programmable electronic device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable device for implementing the process. Figure 1 A process or multiple processes and / or structures illustrate the steps of the functions specified in one or more boxes.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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 or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for optimizing a fault location model in a distribution network, characterized in that, include: Obtain the model evaluation dataset for the first model; the first model is a distribution network fault location model. Based on the model evaluation dataset, the trust information of the first model is determined; The trust information includes process trust, outcome trust, and explainability trust; Based on the trust information, the first model is optimized to obtain the second model.

2. The method according to claim 1, characterized in that, The model evaluation dataset includes: The first model's runtime data, user feedback data, and historical benchmark data; The operational data includes fault data of the distribution network and fault measurement results generated by the first model; The user feedback data refers to the user's feedback regarding the fault measurement results; The historical baseline data includes historical fault location data prior to the introduction of the first model; The model evaluation dataset also includes the actual fault results corresponding to the fault measurement results.

3. The method according to claim 2, characterized in that, The historical baseline data includes the first timeliness rate of the fault location process before the introduction of the first model; the operational data includes the second timeliness rate of the fault location process after the introduction of the first model; the operational data includes the third timeliness rate of the fault location process before the current optimization of the first model; and the fourth timeliness rate of the fault location process after the current optimization of the first model. The trust information includes the first process trust and / or the second process trust. The step of determining the trust information of the first model based on the model evaluation dataset includes: Based on the difference or ratio between the second timeliness rate and the first timeliness rate, the first level timeliness rate is determined, and the first level timeliness rate is the first process trust level. And / or, Based on the difference or ratio between the fourth timeliness rate and the third timeliness rate, a second-level timeliness rate is determined, which is the second process trust level.

4. The method according to claim 2, characterized in that, The operational data includes fault measurement results generated by the first model; the model evaluation dataset also includes the actual fault results corresponding to the fault measurement results. Trust information includes the trustworthiness of the results; The step of determining the trust information of the first model based on the model evaluation dataset includes: Based on the matching relationship between the fault measurement results and the corresponding actual fault results, a confusion matrix is ​​constructed to obtain four types of sample data. The four types of sample data include true positives (TP), false negatives (FN), false positives (FP), and true negatives (TN). Among them, true positives (TP) are samples that are actually fault data and are predicted as fault data; false negatives (FN) are samples that are actually normal data and are predicted as fault data; false positives (FP) are samples that are actually normal data and are predicted as normal data; and true negatives (TN) are samples that are actually normal data and are predicted as normal data. The accuracy is obtained based on the ratio between the first sum and the second sum; the accuracy is the confidence level of the result; the first sum is the sum between the number of true positives and the number of true negatives; the second sum is the sum between the number of true positives, the number of false negatives, the number of false positives, and the number of true negatives.

5. The method according to claim 4, characterized in that, The step of determining the trust information of the first model based on the model evaluation dataset further includes: False alarm rate and false negative rate are calculated simultaneously and used as auxiliary evaluation indicators of result reliability; The false alarm rate is the number of false negatives divided by the total number of false negatives, false positives, and true negatives. The false alarm rate represents the proportion of all actual normal data that is predicted as faulty data. The false negative rate is the number of data that is actually faulty but is predicted as normal, divided by the number of true cases. The false negative rate represents the proportion of data that is actually faulty but is predicted as normal out of all actual faulty data. The false positive rate and false negative rate are used as optimization targets during model training to help improve the reliability of the results.

6. The method according to claim 2, characterized in that, The user feedback data refers to the user's feedback on the fault measurement results; the feedback includes the user's positive or negative trust judgment of the results and the reasons supporting the judgment; The step of determining the trust information of the first model based on the model evaluation dataset includes: Use the user's trust judgment of the result as the first-level decision marker; The reasonableness of the reasons provided by the user is used as the second-level decision marker; Feedback that is labeled as reasonable positive trust and reasonable negative trust is classified as trust, and the rest is classified as distrust; Based on the distribution of trust and distrust classification results, the information gain value is calculated; the information gain value is the interpretable trust level.

7. The method according to any one of claims 1-6, characterized in that, The optimization of the first model based on the trust information to obtain the second model includes: Based on the interpretability trust level, calculate the first information gain value; Based on the first information gain value, the first model is initially optimized to obtain the third model; Based on the output of the third model, the process confidence level, and the result confidence level, calculate the second information gain value; The third model is further optimized based on the second information gain value to obtain the second model.

8. A fault location model optimization device for a power distribution network, characterized in that, The device includes: The acquisition unit acquires the model evaluation dataset of the first model; the first model is a distribution network fault location model. The determining unit is used to determine the trust information of the first model based on the model evaluation dataset; the trust information includes process trust, result trust, and interpretability trust. An optimization unit is used to optimize the first model based on the trust information to obtain a second model.

9. An electronic device, characterized in that, It includes a processor and a memory, the memory being used to store one or more programs and configured to be executed by the processor, the programs including instructions for performing the steps of the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, A computer program for storing electronic data interchange, wherein the computer program causes a computer to perform the method as described in any one of claims 1-7.

Images