A Fault Diagnosis Method for Phase-Shifting Transformers in Flexible Interconnected Distribution Networks Based on Bayesian Inference

Abstract

This invention proposes a fault diagnosis method for flexible interconnection phase-shifting transformers in distribution networks based on Bayesian inference. Based on the analysis of the principles of Bayesian inference and combined with on-site operation and maintenance conditions and expert experience, the calculation of prior probability, likelihood function, and marginal likelihood rate in the Bayesian algorithm is simplified. The posterior probability in Bayesian inference is directly given, and the posterior probability is continuously iterated and corrected through a fault diagnosis model. Then, a fault diagnosis model is constructed by combining the calculation rules of the Bayesian algorithm with industry knowledge. Given a fault mode, the fault mode is determined according to the relationship between the fault mode and the membership degree of the representative event, and the membership degree is calculated. The membership degree concept is similar to the posterior probability, used to quantify the probability of a fault mode. Through the equipment fault diagnosis results, the equipment capability status can be further accurately assessed, and it can be used to evaluate the safety risks of equipment operation.

Description

Technical Field

[0001] This invention relates to the field of distribution transformer fault diagnosis technology, and in particular to a fault diagnosis method for flexible interconnection phase-shifting transformers in distribution networks based on Bayesian inference. Background Technology

[0002] The power system is a vital public service project affecting people's basic needs, including food, clothing, housing, and transportation. It is a crucial guarantee for ensuring safe electricity use and improving people's quality of life. The power system is complex in composition and structure, involving multiple fields such as electrical engineering, communications, and infrastructure. It is a complex system that requires cooperation among these fields. The electrical component of the power system is a comprehensive system involving five parts: power generation, transmission, transformation, distribution, and consumption. Each part works together and is indispensable.

[0003] Distribution is a crucial bridge for power transmission between the power system and users. The section of the power system from the output of a step-down or high-voltage distribution substation to the user end is called the distribution system. A distribution system is a power network system composed of various distribution equipment or components and facilities that transforms voltage and directly distributes electrical energy to end users. The distribution network consists of overhead lines, towers, cables, distribution transformers, switchgear, reactive power compensation capacitors, and other distribution equipment and auxiliary facilities. Its main function in the power grid is to distribute electrical energy.

[0004] Distribution transformers are key components of power distribution networks, serving as electrical devices that convert voltage and adjust current. A distribution transformer, often simply called a "transformer," is a static electrical appliance in an electrical system that transmits AC power by transforming AC voltage and current according to the law of electromagnetic induction. Assessing the operational status of distribution transformers is crucial for ensuring the stable and safe operation of the power distribution network. Against this backdrop, power grid companies are required to adopt reasonable and feasible methods to assume responsibility for ensuring the good and stable operation of all equipment in the power system and guaranteeing a safe, reliable, and continuous supply of electricity. As an indispensable piece of electrical equipment in power distribution, maintaining the normal and healthy operating condition of distribution transformers is a necessary condition for the stability of the power system. If a distribution transformer fails, it can cause minor power outages or, in severe cases, jeopardize the stability of the power system and lead to widespread grid collapse. Therefore, it is essential to detect and resolve distribution transformer faults in their early stages. Addressing defects in distribution transformers before they cause failures reduces the probability of failure and is more beneficial for the stable operation of the power system.

[0005] Phase-shifting transformers are a crucial type of distribution transformer. Fault diagnosis and maintenance decisions are key to analyzing and repairing equipment defects, and are also important means to extend their service life and improve reliability. For a long time, periodic maintenance of phase-shifting transformers has been the primary method for obtaining their health status and operational condition in China. However, with my country's rapid economic development, the number of phase-shifting transformers has increased year by year, rendering previous maintenance methods inapplicable. If the previous periodic maintenance approach continues, the workload will be extremely heavy; furthermore, it may lead to excessively long maintenance cycles, where faults occurring during operation may not be detected in time due to the extended maintenance period, potentially escalating into serious malfunctions with disastrous consequences. Moreover, most equipment is in good condition, and indiscriminate maintenance would result in over-maintenance and wasted resources. In recent years, condition-based maintenance methods have been proposed, which, based on understanding the operating status of distribution transformers, consider various factors to rationally plan equipment maintenance. Targeted maintenance helps improve maintenance quality and efficiency, ensuring the stable and reliable operation of the power grid. Condition-based maintenance, also known as predictive maintenance, determines the operating status, fault type, and severity of equipment based on online monitoring and condition assessment. This scientific assessment allows for more rational equipment maintenance scheduling. Health status assessments of distribution phase-shifting transformers reveal their health condition, while fault diagnosis enables further analysis of transformers operating under adverse conditions, revealing the relationship between "fault symptoms" and "fault modes." Fault diagnosis allows for the timely detection of faults and potential risks in distribution phase-shifting transformers, preventing the escalation and spread of fault hazards. Summary of the Invention

[0006] To address the shortcomings and deficiencies of existing technologies, this invention proposes a fault diagnosis method for distribution network flexible interconnection phase-shifting transformers based on Bayesian inference. Under the premise of meeting the actual production needs, this invention establishes an effective fault diagnosis model for distribution network phase-shifting transformers, thereby further accurately assessing the equipment's capacity status and can be used to assess the equipment's operational safety risks.

[0007] This method, based on the analysis of the principles of Bayesian inference and combined with on-site operation and maintenance conditions and expert experience, simplifies the calculation of prior probability, likelihood function, and marginal likelihood rate in the Bayesian algorithm. It directly provides the posterior probability in Bayesian inference and continuously iterates through a fault diagnosis model to correct the posterior probability. Then, it combines the calculation rules of the Bayesian algorithm with industry knowledge to construct an equipment fault diagnosis model. Given a fault mode, it determines the fault mode based on the relationship between the fault mode and the membership degree of the representative event, and calculates the membership degree. The membership degree concept is similar to the posterior probability, used to quantify the likelihood of a fault mode. Through the equipment fault diagnosis results, the equipment capability status can be further accurately assessed, and it can be used to evaluate the safety risks of equipment operation.

[0008] The present invention specifically adopts the following technical solution:

[0009] A fault diagnosis method for phase-shifting transformers in distribution networks based on Bayesian inference, characterized in that:

[0010] The Bayesian inference in phase-shifting transformer fault diagnosis is simplified as follows:

[0011] (1) After exhausting all abnormal events, assume that each abnormal event is independent of the others;

[0012] (2) Assume that the mapping relationships between all abnormal events and faults are independent of each other;

[0013] (3) Based on the on-site operation and maintenance situation and expert experience, the calculation of prior probability, likelihood function and marginal likelihood rate in Bayesian algorithm is simplified, the posterior probability in Bayesian inference is directly given, and the posterior probability is corrected by iterating through the fault diagnosis model.

[0014] The process of constructing the fault diagnosis model is as follows:

[0015] Step S1: Construct the equipment fault tree and abnormal event set;

[0016] Step S2: Using Bayesian inference, by taking the set of anomalous events as the observation data of each failure mode, the failure inference is abstracted into: after giving E as evidence of the anomalous event, solve the conditional probability of the parameter F that characterizes the failure mode, that is, solve P(F|E).

[0017] Step S3: For a single event (E) i When the failure mode (F) occurs j The probability of P(F) is obtained by continuously accumulating event and fault data to obtain frequency relationships, gradually approaching its P(F) value. i |E i );

[0018] Step S4: Calculate the associated event (E)i E j Failure mode (F) occurs k When calculating the probability of E, after simplification, we can... i E j As a random independent event, i.e., P(F) k |E i &E j )=P(F k |E i )+P(F k |E j )-P(F k |E i )P(F k |E j The physical meaning of this formula is: when two or more abnormal events related to a failure mode occur, the judgment on the probability of the failure mode will be further strengthened.

[0019] Step S5: By constructing a data relationship chain between the abnormal event set and the failure mode, a data-driven equipment fault diagnosis model is realized.

[0020] Furthermore, in step S3, when there is insufficient data sample, an initial value is assigned based on expert experience.

[0021] Furthermore, according to the established failure mode, the failure mode is determined based on the relationship between the failure mode and the membership degree of the representative event, and the membership degree is calculated. The membership degree is based on posterior probability and is used to quantify the likelihood of representing the failure mode. The determination process includes the following steps:

[0022] (1) Set the membership degree of an event to a fault mode according to the causal relationship between the state variable deduction and the fault mode;

[0023] (2) The membership degree of the evidence event shall be in the range of (0,1), the membership degree of the main evidence event shall not be less than 0.5, the membership degree of the secondary evidence event shall not be greater than 0.4; the membership degree of the veto evidence event shall be (-0.1,-1).

[0024] (3) The membership degree of the risk probability of a risk warning event ranges from (0, 0.5), and the risk probability is [0.1, 0.5] depending on the importance of the risk event;

[0025] (4) When multiple label data are accumulated, the membership degree or risk probability is automatically corrected to the membership degree of the event to the failure mode, i.e., the risk probability, based on the statistical results of the "single event-failure mode".

[0026] (5) When multiple events representing equipment failure modes occur, the membership degree of the failure modes needs to be fused and calculated.

[0027] (6) When the membership degree of the fault mode is positive, the diagnostic mode outputs; when it is negative, the fault diagnostic mode does not output.

[0028] Compared with the prior art, the present invention and its preferred embodiments have at least the following beneficial effects:

[0029] 1. By adopting Bayesian inference, the probability of equipment failure can be calculated under certain assumptions. Through the abstract summary of the mechanism and component-level failure modes of the distribution phase-shifting transformer and the simplification of Bayesian inference, an effective fault diagnosis model for the distribution phase-shifting transformer can be established to meet the actual business needs of production.

[0030] 2. The equipment fault diagnosis results can be used to further accurately assess the equipment's capability status and to evaluate the safety risks of equipment operation.

[0031] 3. Case studies show that the constructed fault diagnosis system for distribution phase-shifting transformers can effectively realize equipment fault analysis, and the fault diagnosis results have a high degree of consistency with the actual situation of distribution phase-shifting transformers.

[0032] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0033] Figure 1 This is a framework diagram of the method implementation of an embodiment of the present invention.

[0034] Figure 2 This is the Venn diagram involved in Bayes' theorem in this invention. Detailed Implementation

[0035] In the following, specific embodiments of this application will be described in detail with reference to the accompanying drawings. Based on these detailed descriptions, those skilled in the art will be able to clearly understand and implement this application. Without departing from the principles of this application, features from various embodiments can be combined to obtain new implementations, or certain features from some embodiments can be substituted to obtain other preferred implementations.

[0036] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0037] To make the features and advantages of this patent more apparent and understandable, specific embodiments are provided below for detailed explanation:

[0038] like Figure 1 As shown, the specific construction process of the fault diagnosis method for flexible interconnection phase-shifting transformers in distribution networks based on Bayesian inference provided by this embodiment of the invention can be summarized as follows:

[0039] Step 1: Analyze the principles of Bayesian inference.

[0040] Bayesian inference, based on Bayes' principle in statistics, is a knowledge reasoning method that obtains the probability of events by analyzing the relationships between phenomena. Bayes' principle has a simple structure and clear logic, and has been widely used in big data analysis and reasoning. The Naive Bayes algorithm is a simplified algorithm based on Bayes' principle, which makes an independence assumption between the attributes of phenomena. This reduces the classification effect of the algorithm to some extent, but also greatly reduces the complexity of the algorithm, making it better suited for logical reasoning. The derivation of Bayes' theorem is done using a Venn diagram; details of the Venn diagram can be found in [link to Venn diagram]. Figure 2 .

[0041] Since the features are independent of each other, the expression for P(X|Y) for category y is: In the formula: d is the number of samples in X. Therefore, Since the value of P(X) is relatively fixed, only the numerators need to be compared. Therefore, the category y i The Naive Bayes expression is: Furthermore, as shown in the Venn diagram, the probability of events X and Y occurring simultaneously is: P(X∪Y)=P(X)+P(Y)-P(X∩Y). When X and Y are independent, P(X∩Y)=P(X)P(Y), that is: P(X∪Y)=P(X)+P(Y)-P(X)P(Y).

[0042] Based on Bayesian inference, due to limited sample data, in order to meet actual production needs, and in combination with on-site operation and maintenance conditions and expert experience, the calculation of prior probability, likelihood function and marginal likelihood rate in Bayesian algorithm is simplified, the posterior probability in Bayesian inference is directly given, and the posterior probability is continuously corrected through the fault diagnosis model iteratively.

[0043] include:

[0044] (1) After exhausting all abnormal events, assume that each abnormal event is independent of the others;

[0045] (2) Assume that the mapping relationships between all abnormal events and faults are independent of each other;

[0046] (3) Based on the on-site operation and maintenance situation and expert experience, the calculation of prior probability, likelihood function and marginal likelihood rate in Bayesian algorithm is simplified, the posterior probability in Bayesian inference is directly given, and the posterior probability is corrected by continuous iteration through the fault diagnosis model.

[0047] Step 2: Based on on-site operation and maintenance conditions and expert experience, simplify the calculation of prior probability, likelihood function and marginal likelihood rate in Bayesian algorithm, directly give the posterior probability in Bayesian inference, and continuously iterate and correct the posterior probability through fault diagnosis model.

[0048] The Bayesian algorithm's calculation rules are as follows:

[0049] (1) If only a single state variable anomaly occurs, then from the formula we can obtain: P(F)=P(F|e i )

[0050] (2) If two events occur, assume that the mapping relationship between the abnormal events and the fault modes of each state variable is independent, that is: P(F)=P(F|e1)+P(F|e2)-P(F|e1)P(F|e2)

[0051] (3) If a third event occurs, then by generalizing the above formula, we can obtain: P(F|e1e2)=P(F|e1)+P(F|e2)-P(F|e1)P(F|e2), P(F)=P(F|e1e2)+P(F|e3)-P(F|e1e2)P(F|e3)

[0052] (4) The formula for calculating the comprehensive membership degree of multiple events can be deduced in turn.

[0053] Step 3: Construct an equipment fault diagnosis model by combining Bayesian algorithm calculation rules with industry knowledge. Based on the aforementioned industry knowledge and mathematical methods, the equipment fault diagnosis model achieves a data-driven equipment fault diagnosis model by constructing a data relationship chain between abnormal event sets and fault modes. The fault diagnosis model involves the following ideas:

[0054] Based on industry knowledge such as repair and testing procedures and fault analysis guidelines, a fault tree and anomaly event set are constructed for the equipment. Drawing on Bayesian inference methods, the fault reasoning is abstracted into: given evidence (anomaly events, denoted by E), the conditional probability of the parameter (fault mode, denoted by F) is solved, i.e., P(F|E).

[0055] Single event (E) i When the failure mode (F) occurs j The probability of P(F) can be obtained by continuously accumulating event and fault data, eventually obtaining the frequency relationship, and gradually approaching its P(F) value. i |Ei If the initial data sample is small, initial values ​​can be manually assigned based on expert experience. Related events (E) i E j When the failure mode (F) occurs k To calculate the probability of ), in principle, its marginal distribution must be obtained. k |E i &E j Considering practical engineering conditions, and to simplify the analysis, E will be omitted in this model. i E j As a random independent event, i.e., P(F) k |E i &E j )=P(F k |E i )+P(F k |E j )-P(F k |E i )P(F k |E j The physical meaning of this formula is: when two or more abnormal events related to a failure mode occur, the judgment of the likelihood of the failure mode will be further strengthened.

[0056] Based on the aforementioned industry knowledge and mathematical methods, a data-driven equipment fault diagnosis model can be realized by constructing a data relationship chain between abnormal event sets and fault modes. In the future, with the continuous accumulation of sample data, the diagnostic model parameters can be optimized through mathematical statistics, machine learning, and other methods, thereby improving the accuracy of model diagnosis.

[0057] Step 4: Based on the established failure modes and the relationship between failure modes and their membership degrees, determine the failure modes. The failure modes and their related attributes are shown in Table 1. Calculate the membership degrees. The membership degree concept is similar to posterior probability and is used to quantify the likelihood of representing a failure mode.

[0058] Table 1 Fault Modes and Related Attributes of Distribution Phase-Shifting Transformers

[0059]

[0060]

[0061] Specifically, the membership degree of an event to a failure mode is set according to the causal relationship between the state quantity deduction and the failure mode. The membership degree of evidential events ranges from (0,1), with the membership degree of primary evidential events not less than 0.5 (default 0.6), and the membership degree of secondary evidential events not greater than 0.4 (typical values: 0.4, 0.3, 0.2, 0.1); the membership degree of veto evidential events is (-0.1, -1). The risk probability (membership degree) of risk warning events ranges from (0,0.5], and the risk probability is generally [0.1,0.5] depending on the importance of the risk event, with typical values ​​of 0.1, 0.2, 0.3, 0.4, 0.5.

[0062] After multiple sets of labeled data accumulate, the membership degree or risk probability can be automatically adjusted based on the statistical results of the "single event-failure mode" to correct the membership degree (risk probability) of the event to the failure mode. When multiple events representing equipment failure modes occur, the membership degree (risk probability) of the failure modes needs to be calculated by fusion. When the membership degree (risk probability) of the failure mode is positive, the diagnostic mode outputs it; when it is negative, the fault diagnosis mode does not output it.

[0063] Step 5: The equipment fault diagnosis results can be used to further accurately assess the equipment's capability status and to evaluate the safety risks of equipment operation.

[0064] Taking a 10kV distribution flexible interconnection phase-shifting transformer as an example, the beneficial effects of the present invention are further illustrated.

[0065] The distribution phase-shifting transformer was put into operation on October 13, 2008. In the latest numerical test, its no-load loss was 2kW, its load loss was 8kW, its high-voltage side DC resistance was higher than the other two phases at 16.7MΩ, and its high-voltage to ground insulation resistance was lower than the measured value at 55.2kΩ. Based on comprehensive analysis, the overall rating of this phase-shifting transformer is B, with a score of 87.33. The equipment fault diagnosis results are as follows:

[0066] (1) The equipment has faults such as multiple grounding points in the core or short circuits between segments.

[0067] (2) The equipment is at risk of failure such as internal insulation dampness or process defects.

[0068] (3) There is a short circuit between turns or between layers in the winding.

[0069] Based on expert analysis and on-site inspection by maintenance personnel, the distribution phase-shifting transformer equipment has been in operation for a long time and has potential problems with bushings and windings. The DC resistance on the high-voltage side, the insulation resistance of the high-voltage side to ground, and the dielectric loss of the windings and bushings are abnormal. The fault diagnosis results of the model are basically consistent with the actual situation.

[0070] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0071] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, 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 data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0072] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing 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 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0073] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0074] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0075] This patent is not limited to the above-described preferred embodiment. Anyone can derive other forms of fault diagnosis method for flexible interconnection phase-shifting transformers in distribution networks based on Bayesian inference under the guidance of this patent. All equivalent changes and modifications made within the scope of this patent application shall fall within the scope of this patent.

Claims

1. A fault diagnosis method for phase-shifting transformers in distribution networks based on Bayesian inference, characterized in that: The Bayesian inference in phase-shifting transformer fault diagnosis is simplified as follows: (1) After exhausting all abnormal events, assume that each abnormal event is independent of the others; (2) Assume that the mapping relationships between all abnormal events and faults are independent of each other; (3) Based on the on-site operation and maintenance situation and expert experience, the calculation of prior probability, likelihood function and marginal likelihood rate in Bayesian algorithm is simplified, the posterior probability in Bayesian inference is directly given, and the posterior probability is corrected by iterating through the fault diagnosis model. The process of constructing the fault diagnosis model is as follows: Step S1: Construct the equipment fault tree and abnormal event set; Step S2: Using Bayesian inference, by taking the set of anomalous events as the observation data of each failure mode, the failure inference is abstracted into: after giving E as evidence of the anomalous event, solve the conditional probability of the parameter F that characterizes the failure mode, that is, solve P(F|E). Step S3: For a single event (E) i When the failure mode (F) occurs j The probability of P(F) is obtained by continuously accumulating event and fault data to obtain frequency relationships, gradually approaching its P(F) value. i |E i ); Step S4: Calculate the associated event (E) i E j Failure mode (F) occurs k When calculating the probability of E, after simplification, we can... i E j As a random independent event, i.e., P(F) k |E i &E j )=P(F k |E i )+P(F k |E j )-P(F k |E i )P(F k |E j The physical meaning of this formula is: when two or more abnormal events related to a failure mode occur, the judgment on the probability of the failure mode will be further strengthened. Step S5: By constructing a data relationship chain between the abnormal event set and the failure mode, a data-driven equipment fault diagnosis model is realized.

2. The method for fault diagnosis of phase-shifting transformers in flexible interconnection distribution networks based on Bayesian inference according to claim 1, characterized in that: In step S3, when there is insufficient data sample, an initial value is assigned based on expert experience.

3. The method for fault diagnosis of phase-shifting transformers in flexible interconnection distribution networks based on Bayesian inference as described in claim 1, characterized in that: According to the established failure modes, the failure modes are determined based on the relationship between the failure modes and the membership degrees of the representative events, and the membership degrees are calculated. The membership degrees are based on posterior probabilities and are used to quantify the likelihood of representing the failure modes. The determination process includes the following steps: (1) Set the membership degree of an event to a fault mode according to the causal relationship between the state variable deduction and the fault mode; (2) The membership degree of the evidence event shall be in the range of (0,1), the membership degree of the main evidence event shall not be less than 0.5, the membership degree of the secondary evidence event shall not be greater than 0.4; the membership degree of the veto evidence event shall be (-0.1,-1). (3) The membership degree of the risk probability of a risk warning event ranges from (0, 0.5), and the risk probability is [0.1, 0.5] depending on the importance of the risk event; (4) When multiple label data are accumulated, the membership degree or risk probability is automatically corrected to the membership degree of the event to the failure mode, i.e. risk probability, based on the statistical results of "single event-failure mode". (5) When multiple events representing equipment failure modes occur, the membership degree of the failure modes needs to be fused and calculated. (6) When the membership degree of the fault mode is positive, the diagnostic mode outputs; when it is negative, the fault diagnosis mode does not output.

Images