A power grid visual diagnosis system and method based on multi-source data fusion

Abstract

The application discloses a kind of power grid visual diagnostic system based on multi-source data fusion, it is related to power grid diagnostic system field.The system includes information acquisition module, electrical fault alarm module, non-electrical quantity fault alarm module, fault discrimination module, communication module, fault recording module;Firstly, information acquisition module is carried out to power grid information acquisition, and then information is transported to electrical fault alarm module and non-electrical quantity fault alarm module and is monitored, when abnormality occurs, corresponding action is made, subsequently information is transmitted to fault discrimination module, analysis input information obtains fault type, while fault recording module is recorded to broadcast to power grid;Communication module collects the information of other modules in system, and carries out information exchange with monitoring system.The application combines multi-source data to monitor and alarm the abnormal information of power grid, and discriminates fault type by data model, realizes the all-round monitoring of power grid and provides data support for subsequent maintenance plan.

Description

Technical Field

[0001] This invention relates to a power grid diagnostic system, and more particularly to a power grid visualization diagnostic system and method based on multi-source data fusion. Background Technology

[0002] With economic development, the scale of power systems is constantly expanding, and the power grid structure is becoming increasingly complex. Traditional manual inspections and simple monitoring methods are no longer sufficient to meet the needs of comprehensive and real-time monitoring of the power grid's operating status. Advanced diagnostic systems are needed to improve the efficiency and accuracy of fault detection and diagnosis. Today, the reliability requirements for power supply are extremely high; power outages can cause huge losses to production and daily life. Power grid diagnostic systems can promptly detect potential faults, quickly locate fault positions, and shorten fault repair time, thereby improving the reliability and continuity of power supply. Similarly, the construction of smart grids is a development trend in the power industry, emphasizing the intelligence, automation, and informatization of the power grid. As an important component of smart grids, power grid diagnostic systems can realize the real-time acquisition, analysis, and processing of power grid operating data, providing technical support for the optimized operation and self-healing control of smart grids.

[0003] Chinese patent CN114252716A discloses a power grid fault diagnosis method and device. First, alarm information from the power grid is acquired. The alarm information is processed to obtain a wavelet packet time-spectrum grayscale image. Then, a preset CNN-SVM fusion model is used as the fault diagnosis model to perform feature recognition on the wavelet packet time-spectrum grayscale image to obtain the power grid fault type. Finally, the power grid fault type is pushed to a maintenance terminal for visualization. This invention solves the technical problems of low efficiency and inability to promptly handle power grid faults in traditional power grid fault diagnosis methods. Summary of the Invention

[0004] The technical problem to be solved by this invention is how to provide a power grid diagnostic system and method with more comprehensive monitoring data and more intelligent fault type identification.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by this utility model is as follows:

[0006] A power grid visualization diagnostic system based on multi-source data fusion includes the following components: an information acquisition module, an electrical fault alarm module, a non-electrical quantity fault alarm module, a fault discrimination module, a communication module, and a fault recording module.

[0007] The information acquisition module performs electrical quantity acquisition and non-electrical quantity acquisition.

[0008] 1) Electrical quantity acquisition: The information acquisition module collects voltage and current signals from the power grid in real time, converts the signals into weak current signals through the AC plug-in, and outputs electrical quantity signals after processing by the internal CPU of the module;

[0009] 2) Non-electrical quantity acquisition: The information acquisition module collects physical information from various sensors and input plug-ins in the system in real time, and outputs non-electrical quantity signals after processing by the internal CPU of the module.

[0010] The electrical fault alarm module receives electrical quantity signals output by the information acquisition module and judges the electrical quantities through preset logic, and issues an alarm when an abnormality occurs.

[0011] The non-electrical fault alarm module receives the non-electrical quantity signal output by the information acquisition module and judges it according to the preset control word and preset logic. When an abnormality occurs, an alarm is triggered.

[0012] When an alarm occurs, the fault identification module receives fault information, identifies the fault type, and outputs the fault type.

[0013] The communication module exchanges information in the system with the power grid monitoring system in real time, and displays it visually through the human-machine interface and the power grid monitoring system.

[0014] The fault recording module records the fault electrical quantity data after determining that an abnormality in the electrical quantity has occurred.

[0015] The information acquisition module collects electrical and non-electrical quantities of the monitored power grid;

[0016] The acquisition of electrical quantity information involves receiving the raw analog signals from the monitoring points through measuring equipment and converting them into low-current and low-voltage AC signals. The measuring equipment includes voltage transformers and current transformers. The signals are then converted into unidirectional sinusoidal signals through a signal conditioning circuit and further processed by an analog-to-digital conversion module. The analog-to-digital conversion module acquires the electrical signals output from the signal conditioning circuit in real time, performs analog-to-digital conversion, and outputs the electrical signals to the CPU module. The CPU module processes the electrical signals according to preset logic and outputs electrical quantity parameter signals. The electrical quantity signals include voltage, current, and frequency, power factor, active power, reactive power signals, differential current, zero-sequence current, and negative-sequence current obtained after calculation and processing by the information acquisition system.

[0017] The acquisition of electrical quantity information involves receiving physical information from sensors and input modules, then converting the analog quantity information received from the sensors into digital signals through an analog-to-digital converter module. The digital signals, along with the signals received from the input modules, are transmitted to the CPU module. After processing, non-electrical parameter signals are output, including transformer temperature, oil level, gas concentration, and internal air pressure of the transformer.

[0018] The electrical fault alarm module provides alarms for various electrical quantity anomalies, including but not limited to differential instantaneous trip alarm protection, ratio differential alarm protection, differential current overrun alarm protection, overcurrent stage one, overcurrent stage two, and overcurrent stage three alarm protection, overload protection function PT disconnection alarm protection, overvoltage alarm protection, undervoltage alarm protection, CT disconnection alarm, single-phase grounding alarm, zero-sequence current alarm protection, and negative-sequence current alarm protection.

[0019] The non-electrical quantity fault alarm module selects the control function activation / deactivation, control output delay, and trip alarm status based on the preset control word; it provides alarm protection for various non-electrical quantity anomalies, including but not limited to heavy gas trip, light gas alarm, SF6 gas pressure anomaly alarm, transformer high temperature alarm, transformer over-temperature trip, and transformer low oil level alarm.

[0020] The fault identification module analyzes and identifies fault information using a fault identification data model. The specific steps are as follows:

[0021] 1) Define the parameter attribute set and membership function for the input parameters;

[0022] 2) Calculate the membership values ​​of the input variables;

[0023] 3) Construct a rule base based on fault mechanisms;

[0024] 4) Fault type identification;

[0025] 5) Output fault type.

[0026] The communication module is connected to each module in the system to collect information, including digital signals processed by the information acquisition system. The electrical quantity information in the digital signals includes voltage, current, and frequency, power factor, active power, reactive power, differential current, zero-sequence current, and negative-sequence current information calculated and processed by the information acquisition system. The non-electrical quantity signals include switching signals collected by the input plug-in and physical information collected by the sensors. The module also collects the operating status information of the electrical fault alarm module, the non-electrical quantity fault alarm module, and the fault discrimination module, including the parameters of the protection action, the operating status of the system's own plug-in, and the fault type output by the fault alarm module when the system is abnormal.

[0027] The communication module is equipped with a human-machine interface, which provides a user interface for control. The communication module supports multiple communication protocols, including IEC60870-5-101, IEC60870-5-104, Modbus, CDT, 9702 and proprietary protocols. It connects to the power plant monitoring system through network transmission to complete the functions of sending system information and receiving external commands.

[0028] The fault recording module is activated under the following conditions: overcurrent activation, undervoltage activation, and differential trip activation. When these conditions are met, the device records the corresponding electrical quantities and records the changes in power grid system parameters before and after the fault, which facilitates fault analysis and accident recall.

[0029] A power grid visualization diagnostic method based on multi-source data fusion, the method using the aforementioned power grid visualization diagnostic method, includes the following steps:

[0030] 1) In the information acquisition phase, electrical and non-electrical signals of the power system are acquired and processed into digital signals by the CPU module.

[0031] 2) During the fault alarm phase, the electrical fault alarm module and the non-electrical quantity fault alarm module receive the information output by the information acquisition module. Through preset control words and logic, they execute the corresponding alarm protection or trip protection operation when an abnormality occurs.

[0032] 3) In the fault identification stage, after a fault alarm occurs, the fault information is input into the fault identification module. Through the fault identification data model, the membership degree of the fault parameters is calculated. Combined with the preset rule base, the possible field faults are judged. The correlation between the fault parameters and the field faults is sorted and output through the rule strength.

[0033] 4) During the fault recording phase, when a fault occurs and the conditions for overcurrent start-up, undervoltage start-up, and differential trip start-up are met, the fault recording module records the changes in the power grid system parameters monitored by the system before and after the fault occurs.

[0034] The beneficial effects of adopting the above technical solution are as follows:

[0035] This invention uses sensors and input modules to collect comprehensive information on both electrical and non-electrical parameters of the power grid's operating status, providing a rich data foundation for subsequent fault diagnosis.

[0036] This invention is equipped with an alarm module based on electrical and non-electrical parameters. When an abnormality occurs, it can perform alarm and trip actions, which can effectively prevent the fault range from expanding.

[0037] The fault discrimination data model proposed in this invention defines membership degrees for the range of parameters and then performs fault discrimination based on the membership degrees. Compared with traditional threshold discrimination, it is more detailed and intelligent in critical data discrimination, reducing the risk of misjudgment and missed judgment. Attached Figure Description

[0038] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0039] Figure 1 This is a system diagram of a power grid visualization diagnostic system based on multi-source data fusion proposed in this invention, excluding the communication module.

[0040] Figure 2 This is a system diagram of the communication module collecting and visualizing information in the power grid visualization diagnostic system based on multi-source data fusion proposed in this invention. Detailed Implementation

[0041] To make the above-mentioned objectives, features, and advantages of the present invention more apparent and understandable, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] This invention discloses a power grid visualization diagnostic system based on multi-source data fusion, such as... Figure 1 It includes the following components: information acquisition module, electrical fault alarm module, non-electrical quantity fault alarm module, fault discrimination module, communication module, and fault recording module.

[0043] (a) Information Collection Module

[0044] 1. Electrical quantity acquisition

[0045] The information acquisition module achieves accurate acquisition of electrical quantities through circuit design and processing flow. At the power grid monitoring point, voltage transformers (PTs) and current transformers (CTs) work closely together, using high-precision electromagnetic induction to linearly convert the original high-voltage, high-current analog power grid signals into low-current and low-voltage AC signals suitable for subsequent processing. In practical applications, common 110kV voltage level lines can be converted into a standard 100V secondary voltage signal using PTs, and the CTs can convert thousands of amperes of current into 5A or 1A secondary current signals according to the turns ratio. Subsequently, the signal conditioning circuit optimizes these AC signals. First, the filtering function effectively filters out complex high-frequency noise interference and low-frequency fluctuation components in the power grid, retaining only pure sinusoidal waveform signals that accurately reflect the electrical characteristics of the power grid. At the same time, the amplification or attenuation function precisely adjusts the signal amplitude to the optimal input range of the analog-to-digital converter (ADC). In practical applications, the signal conditioning circuit can amplify weak AC signals with amplitudes fluctuating from a few millivolts to several volts to the standard range of 0V-5V, ensuring high accuracy and stability of the ADC conversion.

[0046] As a crucial component in the digitization of electrical quantities, the analog-to-digital conversion module employs a high-speed, high-resolution ADC chip. It utilizes a sampling frequency of up to several MSPS (millions of samples per second) to acquire the unidirectional sinusoidal signal output from the signal conditioning circuit in real time. Based on a preset quantization precision, it converts the continuous analog electrical signal into a discrete digital coded signal. The specific quantization precision is determined according to the specific grid parameters and the required accuracy. The discrete digital signal is then transmitted to a high-performance CPU module within the module. The CPU module performs in-depth processing of the digital electrical signal according to a preset logic program. In practical applications, it typically uses the Fast Fourier Transform (FFT) algorithm to accurately calculate the signal's frequency components, amplitude, and phase information. Based on circuit principles such as Ohm's law and Kirchhoff's laws, combined with grid topology parameters, it calculates key electrical quantity parameters such as the effective voltage, effective current, active power, reactive power, and power factor at the monitoring point in real time. The calculated parameter signals provide a solid data foundation for subsequent detection and fault diagnosis.

[0047] 2. Non-electrical quantity acquisition

[0048] The information acquisition module acquires non-electrical quantities by collecting analog information from various sensors in the system (such as temperature sensors, pressure sensors, liquid level sensors, etc.) and digital signals from input plug-ins, and then performs targeted processing on each.

[0049] For analog signals from sensors, such as transformer temperature information and the 4-20mA current signal or 0-5V voltage signal output by the transformer oil temperature sensor (which is linearly related to the oil temperature), these analog signals are converted into digital signals via an analog-to-digital converter. Simultaneously, for switch signals acquired by the input modules, they are transmitted directly in digital format. The converted digital signals from the sensors and the switch signals from the input modules are transmitted together to the CPU module. The CPU module analyzes, integrates, and processes these non-electrical signals according to preset program logic, ultimately outputting non-electrical parameter signals such as transformer oil temperature, SF6 gas pressure, and equipment switching status, providing data support for subsequent power grid monitoring and fault diagnosis.

[0050] (ii) Electrical fault alarm module

[0051] The electrical fault alarm module establishes a complete and sophisticated alarm mechanism for abnormal electrical quantities. Its core lies in setting rigorous thresholds and logical judgment criteria for various electrical quantity parameters, covering core electrical quantity indicators for power grid operation.

[0052] 1. Differential instantaneous trip alarm protection: Based on key parameters such as power grid equipment capacity and short-circuit impedance, and combined with expert experience, a reasonable differential current instantaneous trip threshold is set. The measured differential current is compared with this threshold, and the differential instantaneous trip protection is activated by AND operation. This enables the system to immediately trigger the differential instantaneous trip alarm and execute the protection trip action when the measured differential current exceeds the differential current instantaneous trip threshold. This ensures a millisecond-level rapid response to serious internal short-circuit faults and effectively prevents the fault range from expanding.

[0053] 2. The ratio differential alarm protection is based on the second harmonic braking ratio differential principle. Through dynamic correlation analysis of differential current and braking current, it accurately defines the operating boundaries under different operating conditions. When the ratio of differential current to braking current exceeds the preset curve boundary, the system quickly issues a ratio differential alarm, accurately identifying potential faults such as minor inter-turn short circuits and ground short circuits inside transformers, lines, and other equipment, ensuring the safe and stable operation of the equipment. Specifically, it involves several situations, including: when the differential current is greater than the differential protection operating setting, and the braking current is less than the braking current setting, and the ratio of the two is greater than the proportional braking coefficient setting, an alarm operation is performed; when the CT disconnection alarm is activated and the CT disconnection duration exceeds 400ms, the system protection alarm operation is triggered; when the differential current of any phase is in the differential operating zone, the CT disconnection blocking differential protection is not activated, the CT disconnection alarm is activated, and the ratio of the second harmonic current of any phase differential current to the differential current is greater than the harmonic coefficient, an alarm is also triggered and a protection trip operation is performed.

[0054] 3. At the overcurrent protection level, for the first, second, and third stage overcurrent alarm protection, the thermal stability limit of the line, the coordination margin between upper and lower level protection, and the load fluctuation characteristics are fully considered, and multi-stage overcurrent thresholds and corresponding operating delays are set respectively. The first stage of overcurrent protection is used as instantaneous overcurrent protection, and the threshold setting is generally much higher than the rated current, several to more than ten times the rated current of the line. The operating delay is extremely short, generally set to 0 seconds or only tens of milliseconds, mainly used for quickly clearing severe short-circuit faults near the end; the second stage of overcurrent protection has a moderate threshold and a delay set between 0.3s and 0.5s, protecting most areas of the line from faults; the third stage of overcurrent protection is used as backup protection, with a threshold slightly higher than the rated current and a delay set to 1s, ensuring reliable protection of the line and equipment and preventing overload damage even in extremely complex fault scenarios, such as when the upper level protection fails to operate or the fault is not completely cleared.

[0055] 4. Overload alarm protection: When the protection current of any phase exceeds the overload protection current setting and the overload protection is activated, the system will execute the protection alarm action after a time limit of the load. This protection is for short-term overload conditions of the transformer and prevents overheating damage.

[0056] 5. PT disconnection alarm protection: When the PT disconnection control word is activated, if the positive sequence voltage is less than 6V and the protection current of any phase is less than 0.5A, or if the negative sequence voltage is greater than 6V, the system will execute the protection alarm action after the state lasts for 3 seconds. This protection monitors the abnormal voltage of the secondary circuit to ensure accurate voltage measurement and reliable protection function.

[0057] 6. Overvoltage alarm protection: When any line voltage exceeds the overvoltage protection voltage setting and the overvoltage protection is activated, after a time delay of the voltage protection time limit, the system will perform a protection trip and issue a protection alarm signal. This protection defends against overvoltage impacts on the transformer caused by grid voltage fluctuations and prevents insulation damage.

[0058] 7. Undervoltage alarm protection: When any line voltage is less than the undervoltage protection voltage setting and the undervoltage protection is engaged and the switch is in the closed position, after the undervoltage protection time limit delay, the system will perform a protection trip action and issue a protection alarm signal. This protection prevents the low grid voltage from affecting the normal operation of transformers and loads and prevents equipment malfunctions.

[0059] 8. CT disconnection protection: When the CT disconnection protection is activated, if the current change of one phase is less than 0.2A, and the current of the other five phases is greater than 0.3A for 0.5 seconds without any change, the system will execute the protection alarm action. This protection monitors the abnormal current in the monitoring circuit, ensuring that the current sampling of the protection device is accurate and the action is reliable.

[0060] 9. Single-phase grounding alarm protection: When the zero-sequence voltage is greater than the zero-sequence voltage setting value of the single-phase grounding protection and the single-phase grounding protection is activated, the system will execute the protection alarm action after the single-phase grounding time limit delay. This protection detects single-phase grounding faults in the three-phase four-wire wiring method on the low-voltage side to prevent the fault from escalating.

[0061] 10. Negative sequence current alarm protection: If the negative sequence current is greater than the current setting of the negative sequence stage 1 protection and the negative sequence stage 1 protection is activated, the system will execute a protection alarm action after a time limit delay of the negative sequence stage 1 protection. If the negative sequence current is greater than the current setting of the negative sequence stage 2 protection and the negative sequence stage 2 protection is activated, the system will execute a protection trip action after a time limit delay of the negative sequence stage 2 protection. The negative sequence current is calculated based on the phase current on the high-voltage side and is for transformer ungrounded asymmetrical faults.

[0062] This module addresses numerous critical electrical quantity anomalies by establishing strict judgment logic and threshold standards based on power grid operation procedures, equipment tolerance, and extensive practical operating experience data. Once an electrical quantity parameter deviates from its normal operating range and triggers an alarm threshold or violates preset logic rules, the module immediately and accurately triggers the corresponding alarm signal. Through multiple methods, including audible and visual alarms and alarm information uploads via communication interfaces, it promptly informs maintenance personnel of the potential power grid fault, saving valuable time for fault diagnosis and repair, and significantly reducing the risk of power outages and equipment damage.

[0063] (III) Non-electrical quantity fault alarm module

[0064] The non-electrical quantity fault alarm module achieves precise function control and intelligent fault alarm based on the preset control word. The preset control word comprehensively manages function activation / deactivation, output delay, and trip alarm status selection.

[0065] Regarding function activation / deactivation control, maintenance personnel can flexibly set control words through communication modules on the host computer interface or on-site commissioning tools, based on factors such as the operating conditions of power grid equipment, maintenance plans, and seasonal characteristics, to selectively enable or disable specific non-electrical quantity protection functions. In actual operation, during the no-load trial operation phase after transformer maintenance, the heavy gas tripping function can be temporarily disabled, while the light gas alarm function can be enabled to closely monitor potential minor fault signs inside the transformer and avoid false tripping interfering with the test process. After the transformer is officially put into operation and has been running stably for a certain period of time, the control words can be precisely switched according to the equipment status and power grid needs to restore key protection functions such as heavy gas tripping to normal operation, comprehensively ensuring the safety and reliability of the transformer at different operating stages.

[0066] The output delay setting is optimized in conjunction with equipment characteristics, fault development patterns, and power grid fault handling procedures. For non-urgent but timely fault types such as transformer high-temperature alarms, a relatively moderate output delay of around 10-30 seconds is set. This delay provides a buffer period for the equipment to recover on its own or for maintenance personnel to confirm the fault on-site, avoiding false alarms caused by instantaneous interference. It also ensures that when the equipment temperature continues to rise abnormally and the fault situation worsens, an alarm signal is issued in a timely manner, allowing maintenance personnel sufficient time to take countermeasures such as checking the cooling system and adjusting load distribution. For emergency and severe fault scenarios such as transformer overheating tripping, the output delay is strictly controlled to a very short time, ensuring that the system quickly cuts off the power supply at the critical moment when the equipment is about to suffer irreversible damage. This sacrifices local power supply to protect the overall equipment assets of the power grid and prevents catastrophic chain reactions.

[0067] Meanwhile, the trip alarm status selection provides the system with diverse and flexible configuration modes. Based on the importance classification of power grid equipment, fault consequence assessment, and differences in operation and maintenance strategies, operation and maintenance personnel can precisely set certain non-electrical quantity anomalies to alarm-only mode, which facilitates continuous monitoring of fault development and the scheduling of planned maintenance at appropriate times. For fault situations that seriously endanger equipment and power grid safety, such as heavy gas tripping and ultra-high temperature tripping, the system is configured to trip + alarm mode. This ensures that the faulty equipment is isolated by tripping at the moment of the fault, while sending detailed alarm information to the operation and maintenance monitoring center to assist in rapid fault location and emergency repair and restoration operations. This minimizes the power outage time, reduces the scope of power outage impact, improves the power grid's fault response and self-healing capabilities, and ensures that power supply continuity and reliability indicators are achieved.

[0068] Regarding specific alarm actions, this module addresses a series of typical non-electrical quantity anomalies, such as heavy gas tripping, light gas alarms, SF6 gas pressure anomaly alarms, transformer high temperature alarms, transformer over-temperature tripping, and low transformer oil level alarms. Based on feedback signals from various sensors, equipment physical characteristics, and operational experience threshold standards, this module constructs accurate and reliable judgment logic. This includes, but is not limited to: For heavy gas protection, based on the gas rate, gas accumulation, and oil flow characteristics generated by internal transformer faults, real-time monitoring is conducted using high-precision gas detection and oil flow sensing elements installed in the gas relay. When the gas volume or oil flow velocity exceeds the preset severe fault threshold, a tripping command is immediately triggered. For light gas alarms, when a trace amount of gas is detected (such as a small amount of gas escaping due to the slow aging and decomposition of insulation paper) before the heavy gas tripping condition is met, a timely warning signal is issued to alert maintenance personnel to potential health hazards in the equipment. This provides crucial decision-making basis for the formulation and implementation of preventative maintenance strategies, optimizing the balance between maintenance costs and efficiency throughout the equipment's lifecycle.

[0069] (iv) Fault diagnosis module

[0070] This module proposes a novel fault identification data model that analyzes power grid parameters at the time of a fault and identifies the fault type. The improved data model specifically includes:

[0071] 1. Define the parameter attribute set and membership function.

[0072] Based on extensive historical fault data statistical analysis, equipment fault physical model research, and accumulated field operation experience, this system constructs a scientifically sound parameter attribute set and membership function system for key fault parameters. When analyzing line short-circuit faults, the system categorizes short-circuit current amplitude into three parameter attributes: "high," "medium," and "low," establishing parameter attribute sets for each. The membership function adopts a triangular function form, precisely determining the function parameters based on the line's rated current, the short-circuit current calculation boundary, and the differences in current characteristics for different short-circuit types. For some 10kV lines, if the three-phase short-circuit current amplitude exceeds 5kA, it is defined as a "high" parameter attribute set with a membership degree of 1. In the 2-5kA range, the membership degree linearly increases from 0 to 1, belonging to the "medium" parameter attribute set. If it is less than 2kA, the membership degree linearly decreases from 1 to 0, falling into the "low" parameter attribute set. The function is specifically expressed as follows:

[0073] The membership function for a small current is:

[0074]

[0075] The membership function of the medium current is:

[0076]

[0077] The membership function for high current is:

[0078]

[0079] Where I is the input current, I1 is the current near 2kA and slightly less than 2kA, I2 is the current near 2kA and slightly greater than 2kA, I3 is the intermediate current between 2-5kA, I4 is the current near 5kA and slightly greater than 5kA, and I5 is the current near 5kA and slightly less than 5kA. The specific parameters have some overlap in the three parameter attribute sets. Compared with the simple threshold method, this method of dividing parameters has more accurate and intelligent performance. It is more detailed when processing critical data and reduces the risk of misjudgment and missed judgment.

[0080] Similarly, for key parameters such as voltage drop amplitude, power fluctuation, and fault duration, dedicated parameter datasets and membership functions are constructed according to their physical meaning and importance for fault diagnosis. The cross boundaries and degree changes of fault characteristic parameters are adjusted to lay a solid foundation for subsequent intelligent reasoning calculations.

[0081] 2. Calculate the membership values ​​of the input variables.

[0082] At the moment a power grid fault occurs, real-time collected electrical and non-electrical quantity data are input into the system as input variables. Based on a predefined membership function, the membership value of each input variable relative to different parameter datasets is calculated precisely point by point using linear interpolation. When the measured short-circuit current of the line is 3.5kA at a certain moment, according to the above short-circuit current membership function, its membership degree is about 1 in the "medium" parameter dataset, while its membership degree is close to 0 in the "high" and "low" parameter datasets. The same calculation process is performed simultaneously on other input variables such as voltage and power to obtain the complete membership value distribution vector of each variable. This provides a multi-dimensional quantitative indicator system for a comprehensive assessment of the fault status, effectively mining the deep-seated fault feature information hidden in the data, overcoming the defects of traditional binary logic judgment that are prone to omission and misjudgment of fault feature details, and improving the completeness and accuracy of fault diagnosis information.

[0083] 3. Build a rule base

[0084] This study delves into the mechanisms of power grid faults, exploring the causal relationships between fault parameter changes and fault types from multiple dimensions, including circuit theory, equipment structural characteristics, electromagnetic transient processes, and the interaction of operating conditions. It further integrates long-term field operation and maintenance experience in fault case diagnosis, expert knowledge, and simulation test results to construct a rich and logically rigorous rule base. This includes, but is not limited to: if the short-circuit current amplitude has a "high" membership degree, the voltage drop amplitude has a "high" membership degree, and the fault duration has a "short" membership degree, then it is determined to be a near-end metallic three-phase short-circuit fault; when the transformer oil temperature has a "high" membership degree, the oil level has a "low" membership degree, and a light gas alarm signal is triggered, then it is determined to be an internal overheating fault in the transformer accompanied by minor insulation damage and gas accumulation. The rule base, composed of numerous such rule statements, forms a complete knowledge base, presented in an "IF-THEN" logical paradigm. It covers common fault scenarios and complex fault combination patterns of various power grid equipment, providing accurate and reliable knowledge rule support for fault reasoning and ensuring the scientific and authoritative nature of the diagnostic decision-making process.

[0085] 4. Fault Type Identification

[0086] Based on the calculated membership values ​​of the input variables and the rule base, fault type discrimination is performed. The system uses the Mamdani inference method to sequentially traverse each rule in the rule base using the membership values ​​of the input variables. For the current rule, the minimum membership value of the input variable under the corresponding rule's preconditions is first taken as the rule's activation strength. Subsequently, based on the fault type dataset defined by the rule's subsequent events, the subsequent event dataset is filtered using the activation strength as the weight to generate an intermediate result set. After traversing all rules and completing the above operations, the fault types in the intermediate result set are sorted and output in ascending order of rule strength. This order can be seen as arranging the correlation coefficients with the fault cause from smallest to largest. The output results combine location and specific type, such as accurately determining the specific fault type as a single-phase ground fault in phase A of the line, a phase-to-phase short-circuit fault in busbar B, or a short-circuit fault in the internal winding of a certain transformer. This provides precise targeting basis for subsequent fault isolation, repair, and power grid restoration strategy formulation, significantly improving the accuracy of fault diagnosis and decision-making efficiency, and reducing the risk of misjudgment and omission.

[0087] (v) Communication Module

[0088] 1. Data collection and integration

[0089] like Figure 2The communication module collects various data processed by the information acquisition system, constructing a panoramic view of the power grid's operating status. Electrical signals include real-time monitoring values ​​of voltage and current, as well as derived key parameters such as frequency, power factor, active power, and reactive power, obtained through complex calculations by the information acquisition system. This electrical data originates from monitoring points at various levels of the power grid and is collected here after rigorous acquisition, conversion, and calculation processes, comprehensively presenting the power grid's power quality, power flow distribution, and dynamic load changes.

[0090] At the non-electrical signal level, the switching signals collected by the input plug record the instantaneous changes in the mechanical state of power grid equipment, such as discrete events like circuit breaker opening and closing actions and changes in disconnector switching states. The level transition edges accurately mark the moment of equipment state transition, providing key clues for dynamic perception of power grid topology and fault event location. The physical information collected by the sensors deeply reflects the internal operating conditions of power grid equipment. From multi-dimensional physical quantity data such as transformer oil temperature, winding temperature gradient changes, slight changes in SF6 insulating gas pressure, and vibration amplitude spectrum characteristics of equipment casing, the health status of equipment is monitored in a comprehensive and real-time manner, laying a solid foundation for early warning of potential fault hazards.

[0091] Similarly, the acquisition of the system's own protection and operational status information is equally indispensable. Detailed records of protection action parameters are kept of each protection device's activation time and action type, including key elements such as differential protection actions, overcurrent protection actions and their corresponding action thresholds, and action time sequences. This provides accurate first-hand data for analyzing fault characteristics and assessing the performance and reliability of protection devices. Real-time feedback of system plug-in operational status data is provided to various functional plug-ins, such as the CPU plug-in, analog-to-digital converter plug-in, and communication plug-in, including core indicators such as operating voltage, temperature, and data transmission error rate, ensuring complete transparency and monitorability of the system's hardware health status. When the system malfunctions, the fault type information output by the fault alarm module serves as a key carrier of the system's fault diagnosis decision-making results. This information is uploaded at high speed via the communication module, seamlessly connecting to the power grid monitoring center's fault handling process. This drives the maintenance team to formulate and execute precise and efficient emergency response strategies, ensuring rapid isolation and repair of power grid faults and swift restoration of stable operation.

[0092] 2. Human-Machine Interface Design

[0093] The communication module features a meticulously designed human-machine interface, serving as the core hub for system-user interaction. It creates a simple, intuitive, powerful, and highly user-friendly control interface. The interface layout adheres to the principles of hierarchical information layering and highlighting key points, using visual graphic elements as the main body, supplemented by concise text labels to ensure users can quickly locate crucial information and accurately grasp the core pulse of the power grid's operational status within the complex maze of power grid data.

[0094] The operation logic is deeply optimized, following users' natural operating habits and cognitive psychological models. It widely adopts one-click quick operation, intelligent interactive guidance and multi-channel interactive collaborative design strategies, which greatly reduces the complexity of user operation and learning cost, allowing users to focus on core operation and maintenance tasks, comprehensively improving operation efficiency and user experience, reducing the risk of human error, and safeguarding the safe operation and maintenance of the power grid.

[0095] 3. Communication protocol supports network connectivity.

[0096] With its powerful software protocol stack and hardware adaptation capabilities, the communication module is fully compatible with internationally recognized power industry communication standards and protocols such as IEC60870-5-101, IEC60870-5-104, Modbus, CDT, and 9702, as well as proprietary custom protocols. It seamlessly connects the data exchange needs between equipment from different manufacturers and heterogeneous systems, breaks down information silos, and builds a unified and integrated power grid monitoring ecosystem.

[0097] In terms of network connectivity architecture, the module is compatible with diverse network communication media and topologies such as industrial Ethernet, fiber optic ring networks, and wireless private networks, allowing for flexible networking based on power grid scale, geographical distribution, and communication reliability requirements. At the substation level, a high-bandwidth, low-latency industrial Ethernet network is used to build the substation's communication backbone, enabling high-density, high-speed, real-time interactive sharing of data between substation equipment. At the wide-area level, a fiber optic ring network is used to construct a regional communication trunk line for the power grid, ensuring reliable long-distance transmission of power grid operation data across stations and regions. In remote, distributed distribution substations or mobile maintenance scenarios, 4G / 5G wireless private networks are introduced to enable flexible equipment access and remote monitoring and management, expanding the breadth and depth of power grid communication coverage. This provides a solid communication infrastructure support for comprehensive, all-weather, and all-space data interconnection and interoperability in intelligent power grid operation and maintenance management, ensuring real-time perception of power grid status and reliable issuance of precise control commands, thereby improving the intelligence and precision of power grid operation management.

[0098] (vi) Fault recording module

[0099] The fault recording module's startup mechanism is precisely designed around the abnormal states of key electrical quantities in the power grid. Overcurrent startup is based on the rated current parameters and overload capacity characteristics of the line or equipment, with multiple overcurrent thresholds set in a tiered and graded manner (e.g., 1.2 times the rated current for minor overload, 1.5 times the rated current for moderate overload, and more than 5 times the rated current for severe short-circuit faults) and corresponding action delay logic (delay of several seconds for minor overload, instantaneous startup for short-circuit faults). This ensures accurate capture of the fault spectrum from brief current surges to sustained overloads and even severe short-circuit full-current overloads. Undervoltage startup is based on the power grid voltage level and allowable voltage fluctuation range specifications, combined with... The equipment features undervoltage tolerance characteristics, setting multi-stage voltage drop amplitude thresholds (such as mild undervoltage alarm at 80% of rated voltage, initiating waveform recording at 60% of rated voltage, and emergency handling for severe undervoltage below 30% of rated voltage) and duration discrimination conditions to sensitively respond to grid voltage fluctuation fault events; the differential trip initiation deeply integrates the principles and fault characteristics of transformer and line differential protection, using the differential current calculation value exceeding the action threshold and the protection device output trip signal as a double insurance mechanism to accurately record the panoramic waveform of electrical quantity changes at the moment of differential protection action, providing key clues for analyzing the root cause of internal short-circuit faults in the equipment.

[0100] Upon module startup, high-density sampling of fault-related electrical quantities is performed at a high sampling frequency. The sampling frequency is dynamically adjustable based on the performance of the fault recording device and the needs of power grid fault analysis. It comprehensively covers the instantaneous, effective, and phase information of three-phase voltage and current of the faulty line or equipment, as well as key derived electrical quantity data such as zero-sequence current and negative-sequence current components. Through high-precision analog-to-digital conversion and large-capacity data caching technology, it records the continuous dynamic changes of electrical quantities from several cycles before the fault occurred to the stable operation period after the fault is cleared, forming a complete fault electrical quantity waveform dataset. Its data accuracy can reach microsecond-level time resolution and amplitude resolution above one-thousandth of the rated value, providing a massive, high-precision, and high-fidelity raw data repository for in-depth post-fault analysis, precise location of accident causes, evaluation of protection device operation performance, and formulation of power grid operation optimization strategies.

[0101] III. Overall System Workflow

[0102] like Figure 1 The power grid visualization diagnostic method based on multi-source data fusion proposed in this invention includes four stages: information acquisition stage, fault alarm stage, fault identification stage, and fault recording stage.

[0103] (I) Information Collection Phase

[0104] After the system is powered on and initialized, the information acquisition module immediately enters a full-time, full-domain monitoring state, synchronously driving the high-speed parallel operation of the electrical quantity acquisition and non-electrical quantity acquisition sub-processes. At the electrical quantity acquisition end, PTs and CTs capture the raw voltage and current signals of the power grid in real time. Through signal conditioning, analog-to-digital conversion, and CPU processing in a pipelined manner, accurate electrical quantity parameter signals are output. In the non-electrical quantity acquisition sub-process, sensors and input modules work together. Sensors digitize analog signals of physical quantities such as oil temperature, air pressure, and liquid level, while input modules directly acquire switch signals. Both are then combined and processed by the CPU module to generate non-electrical parameter signals. These two types of data are updated in real time with millisecond-level synchronization accuracy, constructing a snapshot of the instantaneous holographic data of the power grid's operating status. This provides high-timeliness and high-precision data momentum for subsequent fault diagnosis processes, laying the foundation for accurate system diagnosis.

[0105] (II) Fault Alarm Stage

[0106] The electrical fault alarm module and the non-electrical fault alarm module subscribe to data update events from the information acquisition module in real time, and perform deep analysis and judgment of electrical and non-electrical quantity signals according to preset logic with a sub-second response speed. Once an electrical quantity signal exceeds the threshold of differential instantaneous trip, overcurrent, overvoltage, etc., or violates complex logic rules (such as excessive zero-sequence current imbalance, sudden change in power factor), the electrical fault alarm module immediately triggers an audible and visual alarm and uploads an alarm event package through the communication interface. The event package details key information such as fault type, occurrence time, and extreme values ​​of electrical quantity parameters. When a non-electrical quantity signal deviates from the normal range, it triggers an alarm in the non-electrical fault alarm module, such as light gas action of transformer gas relay or low SF6 gas pressure alarm. According to the preset control word, it determines whether to only alarm or trip, and at the same time reports the non-electrical quantity anomaly details to the monitoring system. This dual alarm mechanism ensures that fault information is transmitted to the operation and maintenance monitoring terminal without delay or omission, and activates the operation and maintenance emergency response plan.

[0107] (III) Fault Diagnosis Stage

[0108] The fault alarm triggers and instantly activates the fault diagnosis module. This module takes real-time electrical and non-electrical quantity data as input, driving the core process of the fault diagnosis model to operate at high speed. First, it accurately quantifies fault parameter characteristics such as short-circuit current amplitude, voltage drop depth, and abnormal equipment temperature rise based on predefined parameter attribute set membership functions. Then, it searches the rule base, efficiently matching fault characteristics with rules using the Mamdani inference method. After rule strength calculation, it outputs the fault type in sequence. Finally, the diagnostic results are pushed to the communication module and distributed to the host computer monitoring system through multiple channels, providing crucial decision-making basis for the maintenance team to formulate repair strategies. This achieves a leap from monitoring fault symptoms to accurately identifying the essence of the fault, improving the timeliness and accuracy of fault handling, and reducing the duration and losses of power grid outages due to faults.

[0109] (iv) Fault recording stage

[0110] The fault recording module is activated at the moment of fault electrical quantity determination. It is precisely triggered based on starting conditions such as overcurrent, undervoltage, and differential tripping, instantly increasing the sampling frequency for panoramic, high-density sampling of fault electrical quantities. Tracing back from the baseline of the grid's steady-state electrical quantities before the fault, it completely captures the electrical quantity waveforms of the fault impact transient, the fault duration, and the entire dynamic process of grid recovery after fault clearance, forming a high-fidelity electrical quantity data archive that deeply covers the entire fault lifecycle. The data is stored locally on a large-capacity storage medium in a standard format and simultaneously uploaded to the monitoring center in real time via a communication module. This provides irreplaceable data support for fault retrospective analysis, protection action verification, and grid planning and design optimization, helping to improve the grid's fault prevention and operation optimization capabilities and safeguarding the long-term reliable operation of the grid.

Claims

1. A power grid visualization diagnostic system based on multi-source data fusion, characterized in that, It includes the following components: information acquisition module, electrical fault alarm module, non-electrical quantity fault alarm module, fault discrimination module, communication module, and fault recording module; The information acquisition module performs electrical quantity acquisition and non-electrical quantity acquisition. 1) Electrical quantity acquisition: The information acquisition module collects voltage and current signals from the power grid in real time, converts the signals into weak current signals through the AC plug-in, and outputs electrical quantity signals after processing by the internal CPU of the module; 2) Non-electrical quantity acquisition: The information acquisition module collects physical information from various sensors and input plug-ins in the system in real time, and outputs non-electrical quantity signals after processing by the internal CPU of the module. The electrical fault alarm module receives electrical quantity signals output by the information acquisition module and judges the electrical quantities through preset logic, and issues an alarm when an abnormality occurs. The non-electrical quantity fault alarm module receives the non-electrical quantity signal output by the information acquisition module and judges it according to the preset control word and preset logic. When an abnormality occurs, an alarm is triggered. When an alarm occurs, the fault identification module receives fault information, identifies the fault type, and outputs the fault type. The communication module exchanges information in the system with the power grid monitoring system in real time, and displays it visually through the human-machine interface and the power grid monitoring system. The fault recording module records the fault electrical quantity data after determining that an abnormality in the electrical quantity has occurred. The fault identification module analyzes and identifies fault information using a fault identification data model. The specific steps are as follows: 1) Define the parameter attribute set and membership function for the input parameters; 2) Calculate the membership values ​​of the input variables; 3) Construct a rule base based on fault mechanisms; 4) Fault type identification; 5) Output the fault type; Real-time collected electrical and non-electrical quantity data are input into the system as input variables. Based on predefined membership functions, the membership values ​​of each input variable relative to different parameter datasets are accurately calculated point by point using a triangular membership function combined with linear interpolation. The rule base is constructed by associating multiple parameters and building a rule base for common fault scenarios of power grid equipment through membership relationships. After obtaining the membership degree, the Mamdani fuzzy reasoning method is used to traverse the rule base built based on the fault mechanism. The minimum membership degree of the input variable under the corresponding rule premise is taken as the rule activation strength. Subsequently, based on the fault type dataset defined by the rules of subsequent events, the subsequent event dataset is filtered by activation intensity as the weight to generate an intermediate result set; After traversing all rules and completing the above operations, sort the fault types in the intermediate result set, and output the fault types sorted from smallest to largest according to the rule strength.

2. The power grid visualization diagnostic system based on multi-source data fusion according to claim 1, characterized in that, The information acquisition module collects electrical and non-electrical quantities of the monitored power grid; The acquisition of electrical quantity information involves receiving the raw analog signals from the monitoring points through measuring equipment and converting them into low-current and low-voltage AC signals. The measuring equipment includes voltage transformers and current transformers. The signals are then converted into unidirectional sinusoidal signals through a signal conditioning circuit and further processed by an analog-to-digital conversion module. The analog-to-digital conversion module acquires the electrical signals output from the signal conditioning circuit in real time, performs analog-to-digital conversion, and outputs the electrical signals to the CPU module. The CPU module processes the electrical signals according to preset logic and outputs electrical quantity parameter signals. The electrical quantity signals include voltage, current, and frequency, power factor, active power, reactive power signals, differential current, zero-sequence current, and negative-sequence current obtained after calculation and processing by the information acquisition system. The acquisition of electrical quantity information involves receiving physical information from sensors and input modules, then converting the analog quantity information received from the sensors into digital signals through an analog-to-digital converter module. The digital signals, along with the signals received from the input modules, are transmitted to the CPU module. After processing, non-electrical parameter signals are output, including transformer temperature, oil level, gas concentration, and internal air pressure of the transformer.

3. The power grid visualization diagnostic system based on multi-source data fusion according to claim 1, characterized in that, The electrical fault alarm module provides alarms for various electrical quantity anomalies, including but not limited to differential instantaneous trip alarm protection, ratio differential alarm protection, differential current overrun alarm protection, overcurrent stage one, overcurrent stage two, and overcurrent stage three alarm protection, overload protection function PT disconnection alarm protection, overvoltage alarm protection, undervoltage alarm protection, CT disconnection alarm, single-phase grounding alarm, zero-sequence current alarm protection, and negative-sequence current alarm protection.

4. The power grid visualization diagnostic system based on multi-source data fusion according to claim 1, characterized in that, The non-electrical quantity fault alarm module selects the control function activation / deactivation, control output delay, and trip alarm status based on the preset control word; it provides alarm protection for various non-electrical quantity anomalies, including but not limited to heavy gas trip, light gas alarm, SF6 gas pressure anomaly alarm, transformer high temperature alarm, transformer over-temperature trip, and transformer low oil level alarm.

5. A power grid visualization diagnostic system based on multi-source data fusion according to claim 1, characterized in that, The communication module is connected to each module in the system to collect information. The collected digital signals are processed by the information acquisition system. The electrical quantity information in the digital signals includes voltage, current, and frequency, power factor, active power, reactive power, differential current, zero-sequence current, and negative-sequence current information obtained after calculation and processing by the information acquisition system. Non-electrical signals include switching signals acquired by input modules and physical information collected by sensors; It also collects the operating status information of electrical fault alarm modules, non-electrical quantity fault alarm modules, and fault discrimination modules, including the parameters of protection actions, the operating status of the system's own plug-ins, and the fault types output by the fault alarm modules when the system is operating abnormally.

6. The power grid visualization diagnostic system based on multi-source data fusion according to claim 1, characterized in that, The communication module is equipped with a human-machine interface, which provides a user interface for control. The communication module supports multiple communication protocols, including IEC60870-5-101, IEC60870-5-104, Modbus, CDT, 9702 and proprietary protocols. It connects to the power plant monitoring system through network transmission to complete the functions of sending system information and receiving external commands.

7. A power grid visualization diagnostic system based on multi-source data fusion according to claim 1, characterized in that, The fault recording module is activated under the following conditions: overcurrent activation, undervoltage activation, and differential trip activation. When these conditions are met, the device records the corresponding electrical quantities and records the changes in power grid system parameters before and after the fault, which facilitates fault analysis and accident recall.

8. A power grid visualization diagnostic method based on multi-source data fusion, wherein the method uses the power grid visualization diagnostic system based on multi-source data fusion as described in any one of claims 1-7, characterized in that, Includes the following steps: 1) In the information acquisition phase, electrical and non-electrical signals of the power system are acquired and processed into digital signals by the CPU module. 2) During the fault alarm phase, the electrical fault alarm module and the non-electrical quantity fault alarm module receive the information output by the information acquisition module. Through preset control words and logic, they execute the corresponding alarm protection or trip protection operation when an abnormality occurs. 3) In the fault identification stage, after a fault alarm occurs, the fault information is input into the fault identification module. Through the fault identification data model, the membership degree of the fault parameters is calculated. Combined with the preset rule base, the possible field faults are judged. The correlation between the fault parameters and the field faults is sorted and output through the rule strength. 4) During the fault recording phase, when a fault occurs and the conditions for overcurrent start-up, undervoltage start-up, and differential trip start-up are met, the fault recording module records the changes in the power grid system parameters monitored by the system before and after the fault occurs.

Images