An intelligent fault diagnosis and monitoring system for primary and secondary integrated pole-mounted circuit breakers

Abstract

The present invention discloses an intelligent fault diagnosis and monitoring system for a primary-secondary integrated pole-mounted circuit breaker, which relates to the technical field of electrical fault diagnosis and analysis, and includes an information acquisition module, a perception and monitoring module, a diagnosis and prediction module, an evaluation module, and an optimization module. The advantages of the present invention are: achieving unified data formats for heterogeneous devices, breaking down communication barriers between different devices, making the system compatible with multiple devices, improving the comprehensiveness and compatibility of data acquisition, ensuring that data from various types of devices can be effectively collected and integrated, ensuring data accuracy and reliability, reducing data transmission volume through compressed transmission, reducing network bandwidth pressure, ensuring that key fault information can be transmitted in a timely manner, facilitating rapid response and processing by operation and maintenance personnel, reducing the scope and time of fault impact, and utilizing a feature extraction module to accurately extract key information according to actual needs, thereby improving the pertinence and accuracy of fault diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical fault diagnosis and analysis, and in particular to an intelligent fault diagnosis and monitoring system for a primary-secondary integrated pole-mounted circuit breaker. Background Art

[0002] A circuit breaker is a switching device that can close, carry, and interrupt current under normal circuit conditions, and can close, carry, and interrupt current under abnormal circuit conditions within a specified time. Circuit breakers are divided into high-voltage circuit breakers and low-voltage circuit breakers according to their scope of use. The distinction between high and low voltage is relatively vague. Generally, voltages above 3kV are called high voltage. Circuit breakers can be used to distribute electrical energy, start asynchronous motors infrequently, and protect power lines and motors. When they experience severe overload, short circuit, undervoltage, etc., they can automatically cut off the circuit. Their function is equivalent to a combination of a fuse switch and an over-undertemperature relay. In addition, after breaking the fault current, no component changes are generally required. Therefore, they have been widely used.

[0003] Common intelligent fault diagnosis and monitoring systems have difficulty achieving direct communication and data interaction during use. As a result, when integrating multiple devices, it is necessary to develop a separate communication interface for each device, which increases development costs and time. Compatibility issues are prone to occur, hindering comprehensive data collection and system integration. There is a lack of effective outlier filtering methods, and key fault information cannot be transmitted to operation and maintenance personnel in a timely manner, delaying the timing of fault handling. To this end, we propose an intelligent fault diagnosis and monitoring system for pole-mounted circuit breakers that integrates primary and secondary systems. Summary of the Invention

[0004] The object of the present invention is to provide an intelligent fault diagnosis and monitoring system for a primary-secondary integrated pole-mounted circuit breaker.

[0005] To achieve the above objectives, the present invention provides the following technical solutions: an intelligent fault diagnosis and monitoring system for a primary / secondary integrated pole-mounted circuit breaker, the intelligent fault diagnosis and monitoring system comprising:

[0006] The information collection module collects primary and secondary fused pole-mounted circuit breaker fault information and corresponding environmental information from the internet and treats it as case information. The perception monitoring module collects the circuit breaker's electrical parameters, mechanical parameters, and environmental indexes to obtain circuit breaker parameter information. An integrated intelligent monitoring terminal is designed, integrating an edge computing unit and a multi-protocol communication module.

[0007] Obtain Modbus-TCP, IEC61850 and MQTT protocol stacks, and use Modbus-TCP, IEC61850 and MQTT protocol stacks to enable the circuit breaker to have multi-protocol conversion capabilities, automatically match communication protocols according to device types, unify the data formats of heterogeneous devices, perform real-time verification through CRC checksum verification during the data acquisition phase, and obtain a sliding window algorithm. Use the sliding window algorithm to filter outliers. The window size in the sliding window algorithm is set to 10 sampling points. After removing noise data, it is compressed and transmitted. Obtain priority queue management data flow, use priority queue management data flow and time-sensitive network technology to reduce data delay. 50ms, and set the fault alarm data priority to the highest level;

[0008] The feature extraction module analyzes the information types in the case information and arranges them in order. Users have the authority to edit the order of information types. The module obtains a standard sequence table, outputs standardized data based on the dynamic protocol adaptation layer, and sorts the data by features according to the standard sequence table. The module then uses edge computing algorithms to extract parameter features and fault features.

[0009] The diagnosis and prediction module collects data from a deep learning framework to train a time series diagnosis model, obtains a basic model, and uses case information and circuit breaker parameter information to form a training set. A subset of the training set is randomly selected as a test set, and the remaining subsets are used as training sets to train the basic model and obtain a training result set.

[0010] Evaluation module, which evaluates the circuit breaker effect based on different case information and parameter characteristics.

[0011] As a further solution of the present invention: the intelligent fault diagnosis and monitoring system further includes:

[0012] The optimization module uses the training results to optimize the basic model, obtain the optimized model, and build a hierarchical computing architecture, where the hierarchical computing architecture includes the edge and cloud platforms. It uses real-time data analysis and primary decision-making on the edge to analyze different case information and record it on the cloud platform. It iterates and optimizes the optimization model through the cloud platform, develops a visual monitoring interface, integrates equipment status display, fault alarm, and maintenance suggestion generation functions, supports multi-terminal access, connects to the power grid management system, realizes automatic push of fault information, work order distribution and resource scheduling, and forms an intelligent operation and maintenance closed loop.

[0013] As a further solution of the present invention, when sorting the circuit breaker parameter information, the feature extraction module establishes a blank processing unit. When the standard sequence table records an information type but the circuit breaker parameter information does not, the corresponding position in the parameter sequence table is blanked. Blank processing refers to leaving the position where the information type is entered with a numerical value blank. When collecting the circuit breaker parameter information, the circuit breaker parameter information is sorted according to the order of the standard sequence table to obtain the parameter sequence table.

[0014] As a further solution of the present invention: When evaluating the effectiveness of a circuit breaker, the evaluation module converts fault features of different case information into retrieval features, extracts parameter features, analyzes retrieval feature values ​​and parameter feature values, calculates the ratio and total ratio of the parameter feature values ​​to the retrieval feature values ​​of all case information, obtains feature ratios and total feature values, sorts the total feature values ​​in ascending order, collects the position of the circuit breaker, establishes a three-dimensional visualization model, and enters the position of the circuit breaker into the three-dimensional visualization model to obtain a circuit breaker model.

[0015] As a further solution of the present invention, when extracting the parameter features corresponding to the retrieval features in the evaluation module, the retrieval features corresponding to the parameter features are simultaneously screened so that the information types corresponding to the retrieval feature values ​​and the parameter feature values ​​are the same. At the same time, the fluctuation range of the parameter feature values ​​in historical fault cases of similar circuit breakers is counted, and the standard deviation of the parameter feature values ​​is calculated. Let the parameter feature value be , let the number of parameter characteristic values ​​be , let the mean of the parameter characteristic value be :

[0016] ;

[0017] The mean of the parameter characteristic values ​​is calculated using the above formula.

[0018] As a further solution of the present invention: after the mean of the parameter characteristic value is calculated, the standard deviation of the parameter characteristic value is set to :

[0019] ;

[0020] The standard deviation of the parameter characteristic values ​​is calculated using the above formula.

[0021] As a further solution of the present invention: after the standard deviation of the parameter characteristic value is calculated, the standard deviation of one tenth of the parameter characteristic value is extracted as the characteristic threshold, and the retrieval characteristic value is set to , let the characteristic ratio of different case information be , and set weight coefficients for different parameter features. Let the weight coefficients for different parameter features be ;

[0022] ;

[0023] The characteristic ratios of different case information are calculated using the above formula.

[0024] As a further solution of the present invention: After the characteristic ratios in the evaluation module are calculated, all characteristic ratios are added together, and the total characteristic value is set to :

[0025] ;

[0026] The characteristic total values ​​of the parameter information of different circuit breakers are calculated by the above formula. After the characteristic total values ​​are sorted in ascending order, a characteristic index sorting table of all circuit breakers is obtained.

[0027] As a further solution of the present invention: after the three-dimensional visualization model in the evaluation module is obtained, the circuit breaker status is divided into 6 levels according to the result of sorting the total feature values, and is represented by red, yellow, green, dark red, orange, and light green in sequence. The brightness increases with the level, and the upper limit of the color value is 255.

[0028] Also includes fault location capabilities:

[0029] Adopting an integrated design within the FTU, the secondary acquisition and processing of current and voltage signals, communication unit and safety protection adopt an integrated design;

[0030] Support Beidou satellite synchronous positioning;

[0031] It can automatically identify fault traveling wave current, power frequency current and power frequency voltage signals, has data acquisition and waveform storage functions, supports fault, load and other information calls, and sends the information to the master station;

[0032] It has the function of fault recording. When a fault occurs, it will automatically record the fault for fault diagnosis. The recording start conditions may include current mutation, voltage mutation, etc., and the threshold can be set.

[0033] The national grid encryption chip is used for hard encryption. The data is connected to the main station through two-way encryption authentication and an APN intranet channel independent of the FTU. The data is transmitted using MQTT, 101 and 104 protocols.

[0034] For low-current grounding systems, such as ungrounded and arc suppression coil grounded, it can achieve fault distance measurement under ground faults such as metallic grounding, arc grounding, low-resistance grounding, and high-resistance grounding;

[0035] It has remote maintenance functions, including: self-protection and self-checking functions, active alarm in abnormal situations, automatic reset and recovery for possible crashes, local and remote maintenance functions, and support for remote program download and upgrade;

[0036] It has anti-false alarm logic and will not give false alarms in situations such as load fluctuation, heavy load switching, inrush current during closing (including reclosing), and live removal of FTU.

[0037] By adopting the above technical solution, compared with the prior art, the beneficial effects of the present invention are:

[0038] 1. The present invention can automatically match the communication protocol according to the device type through the information collection module, and realize the unification of the data format of heterogeneous devices. This breaks the communication barriers between different devices, makes the system compatible with multiple devices, improves the comprehensiveness and compatibility of data collection, ensures that data from various types of devices can be effectively collected and integrated, and guarantees the accuracy and reliability of the data. The compressed transmission reduces the data transmission volume, improves the transmission efficiency, reduces the pressure on the network bandwidth, ensures that the key fault information can be transmitted in time, facilitates the operation and maintenance personnel to respond quickly, reduces the scope and time of the fault impact, and uses the feature extraction module to accurately extract key information according to actual needs, thereby improving the pertinence and accuracy of fault diagnosis.

[0039] 2. The present invention uses a feature extraction module to process data according to standard sequence table specifications, avoiding data processing errors caused by missing information. This enables the system to better handle incomplete data and ensures more reliable subsequent analysis and diagnosis results based on this data. The evaluation module can accurately quantify the degree of difference between the current circuit breaker parameter characteristics and the search characteristics in the case information. This provides a precise numerical basis for evaluating the operating status of the circuit breaker, helping operation and maintenance personnel quickly determine the similarity between the current circuit breaker and historical fault cases, quickly identify key monitoring targets, and further refine the evaluation process.

[0040] 3. The present invention can reflect the actual operating status of the circuit breaker through the evaluation module, providing more accurate data support for fault diagnosis and prediction. It can intuitively compare the operating status of different circuit breakers and quickly identify circuit breakers with poor operating status and high fault risk. It provides a clear basis for operation and maintenance decision-making, facilitates the reasonable arrangement of operation and maintenance resources and maintenance plans, and is easy for users to understand and identify, avoiding information misunderstanding due to confusing color settings, and ensuring that the display color can accurately reflect the relative status of each circuit breaker. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 4 is a flowchart of the working process of the system of the present invention. DETAILED DESCRIPTION

[0042] The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0043] In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0044] Example 1:

[0045] Please see the attached Figure 1 The present invention provides an intelligent fault diagnosis and monitoring system for a primary-secondary integrated pole-mounted circuit breaker. In a city, the distribution network is like the "electrical blood vessel" of the city, supplying electric energy to various users. Pole-mounted circuit breakers are widely distributed in the streets and alleys of the city, connecting power lines in different areas. Taking a prosperous commercial district as an example, there are dense commercial facilities such as shopping malls, office buildings, and hotels, which have extremely high requirements for the stability and reliability of the power supply. Once a pole-mounted circuit breaker fails, even a short power outage may cause huge economic losses to businesses. Traditional monitoring technology mainly relies on manual regular inspections and discrete secondary devices, which have significant limitations: manual inspections have poor timeliness, high costs, and are difficult to cover hidden faults. The discrete monitoring system is isolated from the primary and secondary equipment data, resulting in information fragmentation, making it impossible to achieve multi-source data fusion analysis, and the fault misjudgment rate remains high.

[0046] Therefore, in order to effectively solve the above problems, the present application proposes a primary and secondary integrated pole mounted circuit breaker intelligent fault diagnosis and monitoring system, as shown in the accompanying drawings of the specification. Figure 1 As shown, the intelligent fault diagnosis and monitoring system includes:

[0047] The information collection module collects primary and secondary fused pole-mounted circuit breaker fault information and corresponding environmental information from the internet and treats it as case information. The perception monitoring module collects the circuit breaker's electrical parameters, mechanical parameters, and environmental indexes to obtain circuit breaker parameter information. An integrated intelligent monitoring terminal is designed, integrating an edge computing unit and a multi-protocol communication module.

[0048] Table 1 Communication protocol table

[0049]

[0050] Design an integrated intelligent monitoring terminal, integrate edge computing units and multi-protocol communication modules, support localized data processing and multi-protocol compatibility. The communication protocol table is suitable for designing an integrated intelligent monitoring terminal, integrating edge computing units and multi-protocol communication modules;

[0051] Obtain Modbus-TCP, IEC61850 and MQTT protocol stacks, and use Modbus-TCP, IEC61850 and MQTT protocol stacks to enable the circuit breaker to have multi-protocol conversion capabilities, automatically match communication protocols according to device types, unify the data formats of heterogeneous devices, perform real-time verification through CRC checksum verification during the data acquisition phase, and obtain a sliding window algorithm. Use the sliding window algorithm to filter outliers. The window size in the sliding window algorithm is set to 10 sampling points. After removing noise data, it is compressed and transmitted. Obtain priority queue management data flow, use priority queue management data flow and time-sensitive network technology to reduce data delay. 50ms, and set the fault alarm data priority to the highest level;

[0052] The feature extraction module analyzes the information types in the case information and arranges them in order. Users have the authority to edit the order of information types. The module obtains a standard sequence table, outputs standardized data based on the dynamic protocol adaptation layer, and sorts the data by features according to the standard sequence table. The module then uses edge computing algorithms to extract parameter features and fault features.

[0053] The diagnosis and prediction module collects data from a deep learning framework to train a time series diagnosis model, obtains a basic model, and uses case information and circuit breaker parameter information to form a training set. A subset of the training set is randomly selected as a test set, and the remaining subsets are used as training sets to train the basic model and obtain a training result set.

[0054] Evaluation module, which evaluates the circuit breaker effect based on different case information and parameter characteristics;

[0055] The intelligent fault diagnosis and monitoring system also includes:

[0056] The optimization module uses the training results to optimize the basic model, obtain the optimized model, and build a layered computing architecture. The layered computing architecture includes edge and cloud platforms. It uses real-time data analysis and primary decision-making on the edge to analyze different case information and record it on the cloud platform. The optimization model is iterated and optimized through the cloud platform. A visual monitoring interface is developed, integrating equipment status display, fault alarm, and maintenance suggestion generation functions. It supports multi-terminal access and connects to the power grid management system to realize automatic push of fault information, dispatch of work orders, and resource scheduling, forming an intelligent operation and maintenance closed loop.

[0057] The standard sequence table is preset by the system for common information types (such as current, voltage, and temperature). Users can drag and drop to adjust the order through the visual interface to generate a custom sequence table;

[0058] Electrical parameters refer to current and voltage, mechanical parameters refer to vibration and displacement, and environmental parameters refer to temperature and humidity;

[0059] Designing an all-in-one intelligent monitoring terminal with integrated edge computing units and multi-protocol communication modules enables local data processing and remote interaction, eliminating the physical isolation and communication barriers of traditional discrete devices.

[0060] Edge computing units are located at the edge of the network, close to the data source or where the data is generated, such as IoT devices, sensors, and cameras. Unlike traditional cloud computing, which transmits data to a remote data center for processing, edge computing units move some or all data processing functions to the edge of the network, reducing the data transmission distance and latency in the network.

[0061] Multi-protocol communication is designed to enable devices or systems with different protocols to communicate and work together, breaking down barriers between protocols and enabling wider device interconnection and data sharing.

[0062] Edge computing algorithms refer to various data processing and analysis algorithms that run in edge computing scenarios. These algorithms are designed to fully utilize the computing power of edge computing units to perform real-time data processing and analysis close to the data source to meet the needs of specific applications. The following are some common types and characteristics of edge computing algorithms.

[0063] The training results in the training result set include diagnostic information and prediction information;

[0064] When training the time series diagnostic model using a deep learning framework, high-precision classification and pattern recognition are achieved for typical fault types such as short circuits, overloads, and mechanical jams. Historical fault data and simulation scenarios are also used to optimize the model's generalization capabilities, adapting to complex operating conditions and environmental changes, and improving diagnostic robustness.

[0065] Deep learning frameworks provide a range of tools and algorithms for building and training neural network models. Time series diagnostic models are specifically designed for time series data. They can capture the characteristics and patterns of data in the temporal dimension. Their basic principle is to automatically extract features from a large amount of time series data and establish a mapping relationship between the input sequence and the output results. For example, in a power system, power equipment operating data collected at different times is used as input. Through model learning, the model outputs diagnostic results such as whether the equipment is faulty and the fault type.

[0066] The optimization module also supports fault self-healing functions, establishes a communication unit, and uses the communication unit to link with the power grid dispatching system to achieve automated operations such as fault isolation and backup line switching, reducing manual intervention;

[0067] Fault self-healing functions and automated decision-making reduce power outage time and energy loss, in line with the development goals of a green grid. Predictive maintenance mechanisms reduce the frequency of manual inspections, optimize resource allocation, and significantly save on operation and maintenance investment.

[0068] Edge refers to computing devices or nodes located at the edge of the network, close to data sources or users. The cloud platform is a centralized computing and storage environment based on cloud computing technology, consisting of a large number of servers, storage devices, and network resources, usually located in a data center.

[0069] Multi-terminal access means that users can access the same application, service or data resource through multiple different types of terminal devices;

[0070] Modbus-TCP encapsulates the Modbus protocol in TCP data packets and uses IP networks for data transmission, enabling devices to communicate and control remotely over the network;

[0071] IEC61850 is a communication standard developed by the International Electrotechnical Commission for substation automation systems;

[0072] The MQTT protocol stack uses asynchronous communication and implements message transmission through message queues, allowing loosely coupled communication between devices;

[0073] CRC check code is a check code that adds a certain number of bits after the data to be transmitted, so that the entire data frame can be divided by a specific generating polynomial.

[0074] The sliding window algorithm maintains a fixed or dynamic size window and slides it over the data sequence. It also performs calculations and judgments based on the elements in the window. During the sliding process, the position and size of the window are continuously updated to find the optimal subsequence or substring that meets specific conditions.

[0075] Priority queue management data flow is a special data structure that can sort elements according to their priority. It has many application scenarios in managing data flow, such as task scheduling and event processing.

[0076] Time-Sensitive Networking (TSN) is a network technology that supports real-time communications and time-sensitive applications. It is designed to meet the needs of industrial automation, in-vehicle communications, smart grids, and other fields for precise synchronization and low latency.

[0077] Specifically, the fault information and corresponding environmental information of the primary and secondary fusion pole-mounted circuit breaker are collected, the electrical parameters, mechanical parameters and environmental index of the circuit breaker are collected, an integrated intelligent monitoring terminal is designed, the edge computing unit and the multi-protocol communication module are integrated, Modbus-TCP, IEC61850 and MQTT protocol stacks are obtained, and Modbus-TCP, IEC61850 and MQTT protocol stacks are used to enable the circuit breaker to have the function of multi-protocol conversion, and the communication protocol is automatically matched according to the device type to unify the data format of heterogeneous devices. Real-time verification is performed through CRC checksum verification during the data collection stage, and a sliding window algorithm is obtained. The sliding window algorithm is used to filter outliers. The window size in the sliding window algorithm is set to 10 sampling points. After removing the noise data, it is compressed and transmitted. The priority queue management data stream is obtained, and the priority queue management data stream and time-sensitive network technology are used to reduce data delay. 50ms, and set the fault alarm data priority to the highest level. The information types in the case information are analyzed and arranged in order. Standardized data is output according to the dynamic protocol adaptation layer and feature-sorted according to the standard sequence table. The edge computing algorithm is used to extract parameter features and fault features. The deep learning framework is collected to train the time series diagnosis model. The case information and circuit breaker parameter information form a training set. A subset of the training set is randomly selected as the test set, and the remaining subset is used as the training set to train the basic model. The circuit breaker effect is evaluated. The basic model is optimized using the training results to obtain an optimized model. A hierarchical computing architecture is constructed, which includes the edge and cloud platforms. Real-time data analysis and primary decision-making at the edge are used to analyze different case information and record it on the cloud platform. The optimization model is iterated and optimized through the cloud platform. A visual monitoring interface is developed, integrating equipment status display, fault alarm, and maintenance suggestion generation functions, supporting multi-terminal access, and connecting to the power grid management system to realize automatic push of fault information, work order distribution and resource scheduling, forming an intelligent operation and maintenance closed loop.

[0078] Example 2:

[0079] When sorting the circuit breaker parameter information in the feature extraction module, a blank processing unit is established. When the standard sequence table records an information type that does not contain circuit breaker parameter information, the corresponding position in the parameter sequence table is blanked. Blank processing means leaving the position where the information type is filled with a value blank. When collecting circuit breaker parameter information, the circuit breaker parameter information is sorted according to the order of the standard sequence table to obtain the parameter sequence table.

[0080] When evaluating the circuit breaker effect in the evaluation module, the fault features of different case information are converted into retrieval features, parameter features are extracted, the retrieval feature values ​​and parameter feature values ​​are analyzed, the ratio and total ratio of the parameter feature values ​​to the retrieval feature values ​​of all case information are calculated, the feature ratio and total feature value are obtained, the total feature value is sorted in ascending order, the position of the circuit breaker is collected, a three-dimensional visualization model is established, and the position of the circuit breaker is entered into the three-dimensional visualization model to obtain the circuit breaker model;

[0081] When extracting the parameter features corresponding to the retrieval features in the evaluation module, the retrieval features corresponding to the parameter features are screened synchronously to make the information types corresponding to the retrieval feature values ​​and the parameter feature values ​​the same. At the same time, the fluctuation range of the parameter feature values ​​of the same type of circuit breakers in historical fault cases is counted, and the standard deviation of the parameter feature values ​​is calculated. Suppose the parameter feature value is , let the number of parameter characteristic values ​​be , let the mean of the parameter characteristic value be :

[0082] ;

[0083] The mean of the parameter characteristic values ​​is calculated by the above formula;

[0084] After the mean of the parameter characteristic value is calculated, the standard deviation of the parameter characteristic value is set to :

[0085] ;

[0086] The standard deviation of the parameter characteristic value is calculated by the above formula;

[0087] Specifically, a blank processing unit is established. When the information type recorded in the standard sequence table is not available in the circuit breaker parameter information, the corresponding position in the parameter sequence table is blanked. When the circuit breaker parameter information is collected, the circuit breaker parameter information is sorted in the order of the standard sequence table to obtain the parameter sequence table. The fault characteristics of different case information are converted into retrieval characteristics, the parameter characteristics are extracted, the retrieval characteristic values ​​and the parameter characteristic values ​​are analyzed, the ratio and the total ratio of the parameter characteristic values ​​to the retrieval characteristic values ​​of all case information are calculated, the characteristic ratio and the total characteristic value are obtained, the total characteristic value is sorted in ascending order, the position of the circuit breaker is collected, a three-dimensional visualization model is established, the position of the circuit breaker is entered into the three-dimensional visualization model, the circuit breaker model is obtained, the retrieval characteristics corresponding to the parameter characteristics are screened, so that the information types corresponding to the retrieval characteristic values ​​and the parameter characteristic values ​​are the same, and at the same time, the fluctuation range of the parameter characteristic values ​​of the same type of circuit breakers in historical fault cases is counted, and the standard deviation of the parameter characteristic values ​​is calculated.

[0088] Example 3:

[0089] After the standard deviation of the parameter characteristic value is calculated, one tenth of the standard deviation of the parameter characteristic value is extracted as the characteristic threshold, and the retrieval characteristic value is set to , let the characteristic ratio of different case information be , and set weight coefficients for different parameter features. Let the weight coefficients for different parameter features be ;

[0090] ;

[0091] The characteristic ratios of different case information are calculated using the above formula;

[0092] After the characteristic ratios in the evaluation module are calculated, all characteristic ratios are added together and the total characteristic value is set to :

[0093] ;

[0094] The characteristic total values ​​of different circuit breaker parameter information are calculated using the above formula, and after the characteristic total values ​​are sorted in ascending order, a characteristic index ranking table of all circuit breakers is obtained;

[0095] After obtaining the 3D visualization model in the evaluation module, the circuit breaker status is divided into 6 levels according to the total feature value sorting results. The levels are represented by red, yellow, green, dark red, orange, and light green. The brightness increases with the level, and the upper limit of the color value is 255.

[0096] Filter the retrieval features corresponding to the parameter features so that the types of information corresponding to the retrieval feature values ​​and the parameter feature values ​​are the same. At the same time, calculate the fluctuation range of the parameter feature values ​​in historical fault cases of similar circuit breakers and calculate the standard deviation of the parameter feature values.

[0097] Specifically, after the characteristic ratio is calculated, all characteristic ratios are added together to calculate the total characteristic value of different circuit breaker parameter information. After the total characteristic values ​​are sorted in ascending order, a characteristic index ranking table of all circuit breakers is obtained. According to the sorting results of the total characteristic values, the circuit breaker status is divided into 6 levels, represented by red, yellow, green, dark red, orange, and light green, respectively. The brightness increases with the level, and the upper limit of the color value is 255.

[0098] Table 2 Fault distance table

[0099]

[0100] The fault distance table is suitable for collecting the electrical parameters, mechanical parameters and environmental indexes of the circuit breaker in the diagnosis prediction module and the acquisition module.

[0101] Working principle:

[0102] First, collect the fault information and corresponding environmental information of the primary and secondary fusion pole-mounted circuit breaker, collect the electrical parameters, mechanical parameters and environmental index of the circuit breaker, design an integrated intelligent monitoring terminal, integrate the edge computing unit and the multi-protocol communication module, obtain Modbus-TCP, IEC61850 and MQTT protocol stacks, and use Modbus-TCP, IEC61850 and MQTT protocol stacks to enable the circuit breaker to have the function of multi-protocol conversion, automatically match the communication protocol according to the device type, unify the data format of heterogeneous devices, perform real-time verification through CRC checksum verification in the data collection stage, and obtain a sliding window algorithm, use the sliding window algorithm to filter outliers, and set the window size in the sliding window algorithm to 10 sampling points. After removing the noise data, it is compressed and transmitted, obtain the priority queue management data stream, and use the priority queue management data stream and time-sensitive network technology to reduce data delay. 50ms, and set the fault alarm data priority to the highest level, analyze the information types in the case information, and arrange the information types in order, output standardized data according to the dynamic protocol adaptation layer, and sort it according to the standard sequence table. Use the edge computing algorithm to extract parameter features and fault features, establish a blank processing unit, and when the information type recorded in the standard sequence table is not available for the circuit breaker parameter information, the corresponding position in the parameter sequence table is blanked. When collecting the circuit breaker parameter information, sort the circuit breaker parameter information according to the order of the standard sequence table to obtain the parameter sequence table, collect the deep learning framework training data, and then analyze the circuit breaker parameter information. The time series diagnosis model is trained. The case information and circuit breaker parameter information form a training set. A subset of the training set is randomly selected as a test set. The remaining subsets are used as training sets to train the basic model. The circuit breaker effect is evaluated. The fault features of different case information are converted into retrieval features. Parameter features are extracted. The retrieval feature values ​​and parameter feature values ​​are analyzed. The ratio and total ratio of the parameter feature value to the retrieval feature value of all case information are calculated to obtain the feature ratio and the feature total value. The feature total value is sorted in ascending order. The position of the circuit breaker is collected, a three-dimensional visualization model is established, and the position of the circuit breaker is entered into the three-dimensional visualization model to obtain the circuit breaker model. Type, filter the retrieval features corresponding to the parameter features, make the retrieval feature values ​​and the parameter feature values ​​correspond to the same type of information, at the same time, count the fluctuation range of the parameter feature values ​​of the same type of circuit breakers in historical fault cases, calculate the standard deviation of the parameter feature values, after the feature ratio is calculated, add all the feature ratios, calculate the total feature value of the parameter information of different circuit breakers, sort the total feature values ​​in ascending order, and obtain the feature index ranking table of all circuit breakers. According to the feature total value ranking result, the circuit breaker status is divided into 6 levels, represented by red, yellow, green, dark red, orange, and light green, with the brightness increasing with the level, and the upper limit of the color value. The value is 255. The basic model is optimized using the training results to obtain the optimized model and build a hierarchical computing architecture. The hierarchical computing architecture includes edge and cloud platforms. Real-time data analysis and primary decision-making at the edge are used to analyze different case information and record it on the cloud platform. The optimization model is iterated and optimized through the cloud platform. A visual monitoring interface is developed to integrate equipment status display, fault alarm, and maintenance suggestion generation functions. It supports multi-terminal access and connects to the power grid management system to realize automatic push of fault information, work order distribution, and resource scheduling, forming an intelligent operation and maintenance closed loop. At this point, the entire workflow is completed.

[0103] Although the present invention is disclosed above with reference to preferred embodiments, this is not intended to limit the present invention. Any person skilled in the art may make possible changes and modifications without departing from the spirit and scope of the present invention. Therefore, any modifications, equivalent variations, and modifications made to the above embodiments in accordance with the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the scope of protection defined by the claims of the present invention.

Claims

1. An intelligent fault diagnosis and monitoring system for primary and secondary integrated pole-mounted circuit breakers, characterized in that: The intelligent fault diagnosis and monitoring system includes: The information collection module collects primary and secondary fused pole-mounted circuit breaker fault information and corresponding environmental information from the internet and treats it as case information. The perception monitoring module collects the circuit breaker's electrical parameters, mechanical parameters, and environmental indexes to obtain circuit breaker parameter information. An integrated intelligent monitoring terminal is designed, integrating an edge computing unit and a multi-protocol communication module. Adopt Modbus-TCP, IEC61850 and MQTT protocol stacks, and use Modbus-TCP, IEC61850 and MQTT protocol stacks to enable the circuit breaker to have the function of multi-protocol conversion, automatically match the communication protocol according to the device type, unify the data format of heterogeneous devices, perform real-time verification through CRC checksum verification during the data acquisition phase, and adopt a sliding window algorithm to filter outliers. The window size in the sliding window algorithm is set to 10 sampling points. After eliminating noise data, it is compressed and transmitted. Priority queues are used to manage data flow. Priority queues are used to manage data flow and time-sensitive network technology to reduce data delay. 50ms, and set the fault alarm data priority to the highest level; The feature extraction module collects information types from case information and arranges them in order. Users have the authority to edit the order of information types. A standard sequence table is obtained, and standardized data is output according to the dynamic protocol adaptation layer. The data is then sorted by features according to the standard sequence table, and edge computing algorithms are used to extract parameter features and fault features. The diagnosis and prediction module uses a deep learning framework to train a time series diagnosis model to obtain a basic model. Case information and circuit breaker parameter information form a training set. A subset of the training set is randomly selected as a test set, and the remaining subsets are used as training sets to train the basic model and obtain a training result set. Evaluation module, which evaluates the circuit breaker effect based on different case information and parameter characteristics; When evaluating the effectiveness of a circuit breaker, the evaluation module converts fault features of different case information into retrieval features, extracts parameter features, analyzes retrieval feature values ​​and parameter feature values, calculates the ratio of the parameter feature values ​​to the retrieval feature values ​​of all case information and the total ratio, obtains feature ratios and total feature values, sorts the total feature values ​​in ascending order, collects the position of the circuit breaker, establishes a three-dimensional visualization model, and enters the position of the circuit breaker into the three-dimensional visualization model to obtain a circuit breaker model; When extracting the parameter features corresponding to the retrieval features in the evaluation module, the retrieval features corresponding to the parameter features are simultaneously screened so that the information types corresponding to the retrieval feature values ​​and the parameter feature values ​​are the same. At the same time, the fluctuation range of the parameter feature values ​​of the same type of circuit breakers in historical fault cases is counted, and the standard deviation of the parameter feature values ​​is calculated. Let the parameter feature value be , let the number of parameter characteristic values ​​be , let the mean of the parameter characteristic value be : ； The mean of the parameter characteristic values ​​is calculated by the above formula; After the mean of the characteristic value of the parameter is calculated, the standard deviation of the characteristic value of the parameter is set to : ； The standard deviation of the parameter characteristic value is calculated by the above formula; After the standard deviation of the parameter characteristic value is calculated, one tenth of the standard deviation of the parameter characteristic value is extracted as the characteristic threshold, and the retrieval characteristic value is set to , let the characteristic ratio of different case information be , and set weight coefficients for different parameter features. Let the weight coefficients for different parameter features be ; ； The characteristic ratios of different case information are calculated using the above formula; After the characteristic ratios in the evaluation module are calculated, all characteristic ratios are added together and the total characteristic value is set to : ； The characteristic total values ​​of the parameter information of different circuit breakers are calculated by the above formula. After the characteristic total values ​​are sorted in ascending order, a characteristic index sorting table of all circuit breakers is obtained.

2. The intelligent fault diagnosis and monitoring system for primary and secondary integrated pole mounted circuit breakers according to claim 1 is characterized in that: The intelligent fault diagnosis and monitoring system also includes: The optimization module uses the training results to optimize the basic model, obtain the optimized model, and build a hierarchical computing architecture, where the hierarchical computing architecture includes the edge and cloud platforms. It uses real-time data analysis and primary decision-making on the edge to analyze different case information and record it on the cloud platform. It iterates and optimizes the optimization model through the cloud platform, develops a visual monitoring interface, integrates equipment status display, fault alarm, and maintenance suggestion generation functions, supports multi-terminal access, connects to the power grid management system, realizes automatic push of fault information, work order distribution and resource scheduling, and forms an intelligent operation and maintenance closed loop.

3. The intelligent fault diagnosis and monitoring system for a primary / secondary integrated pole-mounted circuit breaker according to claim 1 is characterized by: When sorting the circuit breaker parameter information, the feature extraction module establishes a blank processing unit. When the standard sequence table records the information type but the circuit breaker parameter information does not, the corresponding position in the parameter sequence table is blanked. The blank processing refers to leaving the position where the information type is filled with a numerical value blank. When collecting the circuit breaker parameter information, the circuit breaker parameter information is sorted according to the order of the standard sequence table to obtain the parameter sequence table.

4. The intelligent fault diagnosis and monitoring system for a primary / secondary integrated pole-mounted circuit breaker according to claim 1 is characterized by: After the three-dimensional visualization model in the evaluation module is obtained, the circuit breaker status is divided into 6 levels according to the result of the total feature value sorting, and is represented by red, yellow, green, dark red, orange, and light green in sequence. The brightness increases with the level, and the upper limit of the color value is 255.

Images