An adaptive multi-stage power station misoperation prevention locking system debugging system
By introducing a dynamic evaluation threshold mechanism and a neural network model, the shortcomings of fixed thresholds in the evaluation and debugging of the anti-misoperation interlocking system are solved, enabling adaptive evaluation and debugging of the real-time state of the power grid and improving the accuracy and timeliness of the system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUADIAN YUNNAN POWER CO LTD
- Filing Date
- 2026-01-23
- Publication Date
- 2026-06-05
AI Technical Summary
The evaluation and commissioning of existing anti-misoperation interlocking systems rely on preset fixed threshold standards, which cannot detect changes in the real-time operating status of the power grid. This leads to a disconnect between the evaluation results and actual risks, affecting the accuracy of status judgment and the timeliness of early warning.
A dynamic evaluation threshold mechanism is introduced, and a health assessment model is constructed through a neural network convolutional structure. By combining the real-time load rate and fault frequency of the power grid, the evaluation criteria are dynamically adjusted to match the risk level of the power grid, so as to realize the real-time assessment of the system health and the debugging suggestions.
It significantly improves the accuracy of status judgment and the timeliness of risk warning in complex operating environments of the anti-misoperation interlocking system, and enhances the system's adaptive protection capability.
Smart Images
Figure CN122159491A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power system automation and safety protection technology, specifically to a commissioning system for an adaptive multi-level power station anti-misoperation interlocking system. Background Technology
[0002] The anti-misoperation interlocking system is a core automation and safety protection device for ensuring the safe and stable operation of power systems, especially multi-level power plants. Its core function is to prevent operational errors by maintenance personnel through a series of electrical and logical interlocking rules, thereby avoiding serious consequences such as grid failures, equipment damage, and even large-scale power outages caused by human error. Therefore, the reliability, logical correctness, and adaptability to complex operating environments of the anti-misoperation interlocking system directly affect the overall safety protection level of the power grid. Regularly assessing and debugging the anti-misoperation interlocking system is a crucial step in ensuring its continued effective protective function.
[0003] According to existing technology, such as Chinese patent document CN108549650A, a method and system for configuring source-end configuration of anti-misoperation interlocking logic rules in intelligent substations includes: constructing an anti-misoperation interlocking logic rule source-end information database based on an intelligent substation system configuration information model, and the association relationships and indexes of each element in the anti-misoperation interlocking logic rule source-end information database; automatically identifying anti-misoperation devices in the anti-misoperation interlocking logic rule source-end information database based on element characteristics; and searching for primary and secondary equipment information participating in the interlocking logic rule calculation of the anti-misoperation devices based on the association relationships and indexes of each element.
[0004] However, in the aforementioned existing technologies, the evaluation and commissioning of the anti-misoperation interlocking system rely on preset fixed threshold standards. This static evaluation method cannot perceive and respond to changes in the real-time operating status of the power grid, resulting in the evaluation results being out of sync with the actual risk level of the power grid. During high-risk periods, the threshold may be too low to provide timely warnings, while during stable periods, the threshold may be too high, leading to unnecessary interventions. This restricts the accuracy of state judgment, the timeliness of warnings, and the pertinence of commissioning measures, affecting the system's adaptive protection capability against complex operating environments.
[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention
[0006] The purpose of this invention is to provide a commissioning system for an adaptive multi-level power plant anti-misoperation interlocking system, thereby solving the problems mentioned in the background art. This invention overcomes the limitation of traditional anti-misoperation interlocking systems, which use fixed thresholds for evaluation and are disconnected from the actual operating state of the power grid, by introducing a dynamic evaluation threshold mechanism. The thresholds of this invention can be dynamically adjusted according to the real-time load rate and fault frequency of the power grid, ensuring that the system health assessment standard is always matched with the current power grid risk level in real time. This significantly improves the accuracy of state judgment and the timeliness of risk warning, fundamentally enhancing the reliability and adaptive protection of the system in complex operating environments.
[0007] To achieve the above objectives, the present invention provides the following technical solution: A commissioning system for an adaptive multi-level power plant's anti-misoperation interlocking system includes the following functional modules: Feature acquisition and preprocessing module: Collects feature data of the anti-misoperation interlocking system of the power station to be commissioned. The feature data includes the number of cross-station rules, the maximum number of devices associated with a single rule, and the number of historical malfunction events. The feature data is preprocessed, including feature data cleaning and feature data standardization. Health assessment module: Constructs a system health assessment model using a neural network convolutional structure, including an input layer, a hidden layer, and an output layer. The input layer receives preprocessed feature data, the hidden layer is responsible for performing nonlinear transformations and deep feature learning on the preprocessed feature data, and the output layer is used to output the system health index. The preprocessed feature data is input into the system health assessment model to obtain the system health index. The power grid situation awareness module collects real-time operating data of the power grid to which the power station to be tested is connected. The operating data includes the current total active load of the power grid and the total number of power grid fault trips within a fixed period. The collected operating data is cleaned, and based on the cleaned operating data, the load factor and fault frequency factor are calculated respectively. Based on the load factor and fault frequency factor, the real-time risk situation index of the power grid is obtained. Adaptive Judgment and Debugging Module: Based on the real-time risk status index of the power grid, a dynamic evaluation threshold is calculated. The system health index is compared with the dynamic evaluation threshold. Based on the comparison results, the risk level status of the anti-misoperation interlocking system of the power station to be debugged is determined and output, and targeted debugging suggestions are generated.
[0008] Furthermore, the feature data is preprocessed, including feature data cleaning and feature data standardization, based on the following methods: Data cleaning of feature data includes the detection and removal of outliers and duplicate data, as well as the handling of missing values. Box-line analysis is used to identify and remove outliers from feature data, consistency comparison is used to identify and remove duplicate data from feature data, and the mean, median, or mode of feature data is used to fill in missing values in feature data.
[0009] Furthermore, the method for standardizing the feature data is as follows: The min-max normalization method is used to scale the collected feature data to the [0,1] interval, based on the following formula:
[0010] in, It is the feature data after normalization. For the original feature data, It is the minimum value among the same type of feature data in the dataset. It represents the maximum value among the same type of feature data in the dataset.
[0011] Furthermore, the method used to construct the system health assessment model is as follows: The system health assessment model adopts a neural network convolutional structure, including an input layer, a hidden layer, and an output layer. The input layer is used to receive preprocessed feature data. The hidden layer consists of convolutional layers and pooling layers, which are used to perform nonlinear transformation and deep feature learning on the preprocessed feature data. The output layer is used to output the system health index. The convolutional layers are configured with convolutional kernels to extract local correlation patterns in the feature data. The pooling layers use average pooling operations. The system health index is limited to a continuous value between 0 and 1. The mean squared error function is selected as the loss function. The loss function value is calculated based on the output and the true label. The gradient is calculated through the backpropagation algorithm to update the weights and biases of the neural network. The update operation is repeated until the model reaches the predetermined number of training rounds.
[0012] Furthermore, the method used for cleaning the collected operational data is as follows: The sliding window statistical method is used to identify and remove outliers in the running data, and duplicate running data is deduplicated. The formula for calculating the load factor is as follows:
[0013] in, Load factor; This represents the preprocessed total active power load of the power grid. This is the preset maximum safe operating load value for the power grid.
[0014] Furthermore, the formula for calculating the fault frequency factor is as follows:
[0015] in, This is the failure frequency factor; This represents the total number of power grid fault trips within a fixed period after preprocessing. The fixed period is a preset statistical time window.
[0016] Furthermore, the formula used to derive the real-time power grid risk situation index based on the load factor and fault frequency factor is as follows:
[0017] in, This is a real-time risk situation index for the power grid. and The preset weighting coefficients are used, and they satisfy the following conditions: , > .
[0018] Furthermore, the formula used to calculate the dynamic assessment threshold based on the real-time power grid risk situation index is as follows:
[0019] in, For dynamic evaluation thresholds; The preset basic evaluation threshold; To standardize the real-time risk situation index, a linear scaling method is used to adjust the real-time risk situation index of the power grid. Transform to the [0,1] interval to form a standardized real-time risk situation index. .
[0020] Furthermore, the system health index is compared with the dynamic evaluation threshold, and based on the comparison result, the risk level status of the anti-misoperation interlocking system of the power station to be commissioned is determined and output. The execution logic is as follows: when When the time is right, the anti-misoperation interlocking system of the power station to be tested is determined and output as being in a safe state; when At that time, it is determined and outputs that the anti-misoperation interlocking system of the power station to be commissioned is in a risky state; in, The system health index output by the system health assessment model; Based on the system health index of the anti-misoperation interlocking system of the power plant under commissioning, which is in a risky state, it is further divided into early warning state and alarm state, and the execution logic is as follows: when When the system is in a state of alarm, it determines and outputs that the anti-misoperation interlocking system of the power station to be commissioned is in an alarm state. when When the alarm is triggered, the system will determine and output that the anti-misoperation interlocking system of the power station to be tested is in an alarm state. in, The preset warning coefficient, the warning coefficient It is set based on the experience of experts in the field of power system protection.
[0021] Furthermore, the output of targeted debugging suggestions follows the following execution logic: When the anti-misoperation interlocking system of the power station to be commissioned is in a safe state, it outputs a prompt message indicating that the system is in a healthy state and recommends performing routine inspections. When the anti-misoperation interlocking system of the power station to be commissioned is in an early warning state, the output system shows signs of potential risks and suggests checking the cross-station interlocking rule logic and the status of its associated equipment. When the anti-misoperation interlocking system of the power plant to be commissioned is in alarm state, it outputs an alarm message indicating that the system is at high risk and recommends performing a comprehensive verification and simulation test of the interlocking logic.
[0022] Compared with the prior art, the beneficial effects of the present invention are: This invention introduces a dynamic assessment threshold mechanism based on the real-time risk status of the power grid, enabling the health index assessment standard of the anti-maloperation interlocking system to adaptively adjust with the power grid's operating state: when the power grid load increases or the frequency of faults rises, the dynamic threshold increases accordingly, thereby imposing higher requirements on the system's health and ensuring sufficient reliability of the interlocking system under high-risk environments; conversely, when the power grid is operating smoothly, the threshold is appropriately lowered to avoid imposing unnecessary stringent constraints on the system. This dynamic threshold design overcomes the problem of traditional fixed assessment standards being disconnected from the actual risks of the power grid, achieving real-time linkage between assessment conditions and the operating environment. This significantly improves the accuracy of risk status identification and the timeliness of early warning, making commissioning recommendations more targeted and effectively enhancing the adaptive protection capability of the anti-maloperation interlocking system against complex and ever-changing power grid environments. Attached Figure Description
[0023] Figure 1 A block diagram of a commissioning system for an adaptive multi-level power station anti-misoperation interlocking system; Figure 2 This is a schematic diagram of the commissioning process for an adaptive multi-level power station anti-misoperation interlocking system. Detailed Implementation
[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.
[0025] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0026] Example: Please see Figures 1-2 The present invention provides a technical solution: A commissioning system for an adaptive multi-level power plant's anti-misoperation interlocking system includes the following functional modules: Feature Acquisition and Preprocessing Module: This module collects feature data from the anti-misoperation interlocking system of the power station to be commissioned. The feature data includes the number of cross-station rules, which measures the cross-station coordination scope and complexity of the system's rule system; the maximum number of devices associated with a single rule, which characterizes the scale of devices involved in key rules in the system and their potential impact range; and the number of historical malfunction events, which directly relates to the reliability and stability performance of the system in actual operation. The feature data is preprocessed to ensure the accuracy and consistency of subsequent analysis. The preprocessing includes feature data cleaning and feature data standardization. The feature data is preprocessed, including feature data cleaning and feature data standardization, based on the following methods: Data cleaning of feature data includes the detection and removal of outliers and duplicate data, as well as the handling of missing values. Box-line analysis is used to identify and remove outliers from the feature data. By calculating the quartiles and interquartile ranges of the data, the normal fluctuation range of the data is defined, and outlier data points that deviate from this range are screened out, thereby reducing the interference of extreme values on the overall data distribution and model training. By verifying the complete consistency of each data record on key fields, duplicate entries that may have occurred during collection or transmission are effectively identified and removed, ensuring the uniqueness and accuracy of the dataset. Finally, the mean, median, or mode of the feature data are used to fill in missing values in the feature data.
[0027] The method for standardizing the feature data is as follows: The min-max normalization method is used to scale the collected feature data to the [0,1] interval, based on the following formula:
[0028] in, It is the feature data after normalization. For the original feature data, It is the minimum value among the same type of feature data in the dataset. It represents the maximum value among the same type of feature data in the dataset.
[0029] Health Assessment Module: Constructs a system health assessment model using a neural network convolutional structure, including an input layer, hidden layers, and an output layer. The input layer receives preprocessed standardized feature data and transforms it into a numerical representation that can be processed by subsequent layers. The hidden layer performs deep feature learning on the input data through a series of nonlinear transformations, gradually extracting and fusing the implicit correlation patterns and state information in the data, thereby forming a multi-level, abstract representation of the system's operating status. The output layer generates a continuous system health index based on the high-order features extracted by the hidden layer through a mapping function. This index comprehensively reflects the overall operating status and reliability level of the anti-misoperation interlocking system. The preprocessed feature data is input into the system health assessment model to obtain the system health index. The method used to construct the system health assessment model is as follows: The system health assessment model employs a neural network convolutional structure, comprising an input layer, hidden layers, and an output layer. The input layer receives preprocessed standardized feature data and transforms it into a numerical representation suitable for subsequent layers. The hidden layer, composed of convolutional and pooling layers, performs nonlinear transformations and deep feature learning on the input feature data. The convolutional layers, through configured kernels, slide through the data to extract localized patterns and dependencies, thereby capturing spatial or structural characteristics between features. The pooling layers downsample the features output by the convolutional layers, using average pooling to retain key information while reducing data dimensionality, thus enhancing feature robustness and generalization ability. The output layer, based on the deep abstract features extracted by the hidden layer, generates a system health index through mapping and aggregation. This system health index is defined as a continuous value between 0 and 1. During model training, the mean squared error function (MSE) is selected as the loss function to quantify the difference between the model's output and the true labels. This loss function is calculated based on the model's forward propagation output of the input data. By comparing the output with the actual labeled data point by point, the mean squared error is obtained, thus forming an optimizable scalar loss value. Subsequently, using the backpropagation algorithm, the system calculates the gradient information of each parameter in the network layer by layer based on the loss value. These gradients reflect the contribution of each weight and bias to the overall error and their adjustment direction. Based on the obtained gradients, the weights and biases in the neural network are iteratively updated using an optimization algorithm to gradually reduce the loss function value and improve the model's prediction accuracy. This process is repeated after each training iteration, continuously performing forward propagation, loss calculation, backpropagation, and parameter updates until the model training reaches the preset total number of rounds. This ensures that the model can fully learn the features and patterns in the data and possess stable generalization performance.
[0030] The power grid situation awareness module collects real-time operational data from the power grid connected to the power station under commissioning. This data includes the current total active power load and the total number of fault trips within a fixed period. This data reflects the real-time operating status and safety level of the power grid. The total active power load reflects the grid's carrying capacity and load pressure, while the total number of fault trips within a fixed period directly relates to the grid's stability and reliability. The module cleans the collected operational data and calculates a load factor and a fault frequency factor based on this data. The load factor quantifies the ratio between the current load and its safe operating limit, reflecting the load pressure on the grid. The fault frequency factor characterizes the frequency of faults occurring in the grid per unit time, reflecting the density of abnormal situations during grid operation. Based on the load factor and fault frequency factor, a real-time power grid risk situation index is derived. This index is a comprehensive assessment indicator that uses a weighted fusion of load pressure and fault frequency to form a quantitative description of the overall risk level of the power grid. The method used to clean the collected operational data is as follows: The sliding window statistical method is used to identify and remove outliers in the running data. This method sets a data window of fixed length, slides it sequentially on the time series, and calculates the statistical characteristics of the data within the window in real time. Then, based on the preset deviation threshold, it judges and removes abnormal data points that deviate significantly from the normal fluctuation range. In response to the problem of duplicate data that may be caused by the collection or transmission process, the system compares the timestamps and numerical content of the data records to identify and merge duplicate entries that are completely identical or highly similar, retaining only the unique valid record, thereby ensuring the uniqueness and consistency of the data used in subsequent analysis. The formula for calculating the load factor is as follows:
[0031] in, The load factor is a dimensionless ratio used to quantify the relative relationship between the current load of the power grid and its safe carrying capacity. This represents the preprocessed total active power load of the power grid. This is the preset maximum safe operating load value for the power grid, representing the upper limit of the load that the power grid can withstand under long-term stable operating conditions; when When it increases, A corresponding increase indicates that the current load on the power grid is approaching or exceeding the safety limit, increasing the operational pressure on the power grid and raising the potential risk of overload; when When decreasing, The corresponding decrease reflects that the power grid load is relatively light, the operating margin is sufficient, and the system is in a relatively relaxed load state.
[0032] The formula for calculating the fault frequency factor is:
[0033] in, This is the fault frequency factor, used to quantify the frequency of fault tripping events occurring in the power grid per unit time. This represents the total number of power grid fault trips within a fixed period after preprocessing, reflecting the degree of anomaly in the power grid during that time period. The fixed period is a preset statistical time window, and the range of the fixed period is set based on the power grid dispatch cycle, monitoring real-time requirements, and fault data statistical characteristics. when When it increases, The corresponding increase indicates that the power grid experiences more frequent fault events within the statistical period, the system is in an unstable or highly disturbed operating state, and the safety risks are significantly increased; when When decreasing, The corresponding decrease reflects that the power grid is operating relatively smoothly, with few fault events and high system reliability.
[0034] The formula used to derive the real-time risk status index of the power grid based on the load factor and the fault frequency factor is as follows:
[0035] in, The Real-Time Risk Situation Index of the Power Grid is a dimensionless quantitative indicator that comprehensively reflects the current operational risk level of the power grid. and These are preset weighting coefficients used to adjust the relative importance of the load rate factor and the failure frequency factor in the comprehensive evaluation, and satisfy the following conditions: , > ; when Increase, meaning the power grid is approaching full load. This increase also indicates a decrease in the power grid stability margin and a high-load risk state for the system. At this time, the requirements for the correctness and real-time performance of the anti-misoperation interlocking system are correspondingly increased; when Increase, meaning an increase in the number of failures per unit time. The increase indicates that the uncertainty of the power grid operating environment is increasing, and the system needs to have stronger anti-interference and anti-malfunction capabilities.
[0036] Adaptive Judgment and Debugging Module: Based on the real-time risk status index of the power grid, a dynamic evaluation threshold is calculated. This threshold is not fixed, but is dynamically adjusted in real time according to factors such as the power grid load level and fault frequency, so as to more realistically reflect the current reliability requirements of the power grid for the anti-misoperation interlocking system. The system health index is compared with the dynamic evaluation threshold. Based on the comparison results, the system determines the risk level of the anti-misoperation interlocking system in the current power grid environment. Based on the specific risk level determined, the system automatically generates and outputs debugging and maintenance suggestions with clear and targeted suggestions. The formula used to calculate the dynamic assessment threshold based on the real-time risk situation index of the power grid is as follows:
[0037] in, For dynamic evaluation thresholds; The preset basic evaluation threshold is... It is set based on historical operating data, system design specifications, and the experience of domain experts. To standardize the real-time risk situation index, a linear scaling method is used to adjust the real-time risk situation index of the power grid. Transform to the [0,1] interval to form a standardized real-time risk situation index. ; when When it increases, The corresponding increase indicates that the current risk level of the power grid has risen, and the system health must reach a higher standard to be considered safe, thus dynamically increasing the reliability requirements for the anti-misoperation interlocking system; when When decreasing, The corresponding reduction indicates that the power grid is operating relatively smoothly, and the requirements for system health are moderately relaxed to avoid overly restricting system operation in a low-risk environment.
[0038] The process of comparing the system health index with the dynamic evaluation threshold, determining and outputting the risk level of the anti-misoperation interlocking system of the power station under commissioning based on the comparison result, executes the following logic: when When the system meets the safety requirements of the current power grid environment in terms of structural integrity, rule rationality, and historical performance, it is determined and outputs that the anti-misoperation interlocking system of the power station to be debugged is in a safe state. when When this occurs, it indicates that the system has failed to meet the safety level required by the current power grid risk environment and cannot fully adapt to the current power grid load and fault frequency pressure. The system is judged and output that the anti-misoperation interlocking system of the power station to be tested is in a risky state. in, The system health index output by the system health assessment model; Based on the system health index of the anti-misoperation interlocking system of the power plant under commissioning, which is in a risky state, it is further divided into early warning state and alarm state, and the execution logic is as follows: when When this occurs, it indicates that the system has deviated from the ideal operating range and the health status has initially declined. Although it has not reached the level of immediate danger, it has already indicated that there may be potential hidden dangers or performance degradation in its internal rules and logic or equipment associations. Attention should be paid and preventive checks should be prepared. The system should be judged and output as being in a warning state for the anti-misoperation interlocking system of the power station to be commissioned. when When this occurs, it indicates that the system health has dropped to a significantly low level, and there is a high risk to the reliability of its interlocking logic, the correctness of rule execution, or the coordination between devices. Intervention measures need to be taken to prevent safety events such as malfunction or refusal to operate. The system is then determined to be in an alarm state. in, The preset warning coefficient, the warning coefficient It is set based on the experience of experts in the field of power system protection.
[0039] The output provides targeted debugging suggestions, and the execution logic is as follows: When the anti-misoperation interlocking system of the power station to be commissioned is in a safe state, the output system is in a healthy state and it is recommended to perform the regular routine inspection process, including periodically checking each functional module of the system, recording operating parameters, confirming the integrity of the rule base, and checking whether the communication link is normal. When the anti-misoperation interlocking system of the power station to be commissioned is in an early warning state, there are potential risk signs in the output system and it is recommended to start a special verification procedure. The focus is on a comprehensive review of the logical structure of the cross-station interlocking rules, verifying the coordination and consistency between the rules, and checking the real-time status and historical action records of each related device to analyze whether there are abnormal behaviors or response delays, so as to adjust the rule configuration or equipment maintenance strategy in a timely manner. When the anti-misoperation interlocking system of the power plant to be commissioned is in an alarm state, the output system is at high risk and it is recommended to immediately perform a comprehensive verification and in-depth simulation test of the interlocking logic. The comprehensive verification includes verifying the logic of all interlocking rules one by one, conflict detection and coverage analysis; constructing a simulation scenario that closely resembles the actual operation of the power grid for simulation testing, testing the system's response behavior under various fault conditions and extreme load conditions, evaluating its accuracy and timeliness, and ensuring that the system can correctly and reliably execute the interlocking function in real faults.
[0040] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.
[0041] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.
[0042] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.
[0043] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.
Claims
1. A commissioning system for an adaptive multi-level power station's anti-misoperation interlocking system, characterized in that, Includes the following functional modules: Feature acquisition and preprocessing module: Collects feature data of the anti-misoperation interlocking system of the power station to be commissioned. The feature data includes the number of cross-station rules, the maximum number of devices associated with a single rule, and the number of historical malfunction events. The feature data is preprocessed, including feature data cleaning and feature data standardization. Health assessment module: Constructs a system health assessment model using a neural network convolutional structure, including an input layer, a hidden layer, and an output layer. The input layer receives preprocessed feature data, the hidden layer is responsible for performing nonlinear transformations and deep feature learning on the preprocessed feature data, and the output layer is used to output the system health index. The preprocessed feature data is input into the system health assessment model to obtain the system health index. The power grid situation awareness module collects real-time operating data of the power grid to which the power station to be tested is connected. The operating data includes the current total active load of the power grid and the total number of power grid fault trips within a fixed period. The collected operating data is cleaned, and based on the cleaned operating data, the load factor and fault frequency factor are calculated respectively. Based on the load factor and fault frequency factor, the real-time risk situation index of the power grid is obtained. Adaptive Judgment and Debugging Module: Based on the real-time risk status index of the power grid, a dynamic evaluation threshold is calculated. The system health index is compared with the dynamic evaluation threshold. Based on the comparison results, the risk level status of the anti-misoperation interlocking system of the power station to be debugged is determined and output, and targeted debugging suggestions are generated.
2. The commissioning system for an adaptive multi-level power station anti-misoperation interlocking system according to claim 1, characterized in that: The feature data is preprocessed, including feature data cleaning and feature data standardization, based on the following methods: Data cleaning of feature data includes the detection and removal of outliers and duplicate data, as well as the handling of missing values. Box-line analysis is used to identify and remove outliers from feature data, consistency comparison is used to identify and remove duplicate data from feature data, and the mean, median, or mode of feature data is used to fill in missing values in feature data.
3. The commissioning system for an adaptive multi-level power station anti-misoperation interlocking system according to claim 2, characterized in that: The method for standardizing the feature data is as follows: The min-max normalization method is used to scale the collected feature data to the [0,1] interval, based on the following formula: in, It is the feature data after normalization. For the original feature data, It is the minimum value among the same type of feature data in the dataset. It represents the maximum value among the same type of feature data in the dataset.
4. The commissioning system for an adaptive multi-level power station anti-misoperation interlocking system according to claim 1, characterized in that: The method used to construct the system health assessment model is as follows: The system health assessment model adopts a neural network convolutional structure, including an input layer, a hidden layer, and an output layer. The input layer is used to receive preprocessed feature data. The hidden layer consists of convolutional layers and pooling layers, which are used to perform nonlinear transformation and deep feature learning on the preprocessed feature data. The output layer is used to output the system health index. The convolutional layers are configured with convolutional kernels to extract local correlation patterns in the feature data. The pooling layers use average pooling operations. The system health index is limited to a continuous value between 0 and 1. The mean squared error function is selected as the loss function. The loss function value is calculated based on the output and the true label. The gradient is calculated through the backpropagation algorithm to update the weights and biases of the neural network. The update operation is repeated until the model reaches the predetermined number of training rounds.
5. The commissioning system for an adaptive multi-level power station anti-misoperation interlocking system according to claim 1, characterized in that: The method used to clean the collected operational data is as follows: The sliding window statistical method is used to identify and remove outliers in the running data, and duplicate running data is deduplicated. The formula for calculating the load factor is as follows: in, Load factor; This represents the preprocessed total active power load of the power grid. This is the preset maximum safe operating load value for the power grid.
6. The commissioning system for an adaptive multi-level power station anti-misoperation interlocking system according to claim 5, characterized in that: The formula for calculating the fault frequency factor is: in, This is the failure frequency factor; This represents the total number of power grid fault trips within a fixed period after preprocessing. The fixed period is a preset statistical time window.
7. The commissioning system for an adaptive multi-level power station anti-misoperation interlocking system according to claim 6, characterized in that: The formula used to derive the real-time risk status index of the power grid based on the load factor and the fault frequency factor is as follows: in, This is a real-time risk situation index for the power grid. and The preset weighting coefficients are used, and they satisfy the following conditions: , > .
8. The commissioning system for an adaptive multi-level power station anti-misoperation interlocking system according to claim 1, characterized in that: The formula used to calculate the dynamic assessment threshold based on the real-time risk situation index of the power grid is as follows: in, For dynamic evaluation thresholds; The preset basic evaluation threshold; To standardize the real-time risk situation index, a linear scaling method is used to adjust the real-time risk situation index of the power grid. Transform to the [0,1] interval to form a standardized real-time risk situation index. .
9. The commissioning system for an adaptive multi-level power station anti-misoperation interlocking system according to claim 8, characterized in that: The process of comparing the system health index with the dynamic evaluation threshold, determining and outputting the risk level of the anti-misoperation interlocking system of the power station under commissioning based on the comparison result, executes the following logic: when When the time is right, the anti-misoperation interlocking system of the power station to be tested is determined and output as being in a safe state; when At that time, it is determined and outputs that the anti-misoperation interlocking system of the power station to be commissioned is in a risky state; in, The system health index output by the system health assessment model; Based on the system health index of the anti-misoperation interlocking system of the power plant under commissioning, which is in a risky state, it is further divided into early warning state and alarm state, and the execution logic is as follows: when When the system is in a state of alarm, it determines and outputs that the anti-misoperation interlocking system of the power station to be commissioned is in an alarm state. when When the alarm is triggered, the system will determine and output that the anti-misoperation interlocking system of the power station to be tested is in an alarm state. in, The preset warning coefficient, the warning coefficient It is set based on the experience of experts in the field of power system protection.
10. The commissioning system for an adaptive multi-level power station anti-misoperation interlocking system according to claim 9, characterized in that: The output provides targeted debugging suggestions, and the execution logic is as follows: When the anti-misoperation interlocking system of the power station to be commissioned is in a safe state, it outputs a prompt message indicating that the system is in a healthy state and recommends performing routine inspections. When the anti-misoperation interlocking system of the power station to be commissioned is in an early warning state, the output system shows signs of potential risks and suggests checking the cross-station interlocking rule logic and the status of its associated equipment. When the anti-misoperation interlocking system of the power plant to be commissioned is in alarm state, it outputs an alarm message indicating that the system is at high risk and recommends performing a comprehensive verification and simulation test of the interlocking logic.