An automatic point-to-point method between a substation and a dispatch master station and related equipment

Abstract

This invention discloses an automatic point-to-point method and related equipment for communication between a substation and a dispatch master station. The method includes: pre-storing standard point table information in a verification platform, which is located on the dispatch master station side and communicatively connected to both the substation and the dispatch master station; in verification mode, the verification platform sends a point-to-point trigger command to the substation; the verification platform captures the point-to-point response signal uploaded by the substation in response to the trigger command, and simultaneously captures the alarm signal generated on the dispatch master station side; the point-to-point response signal and alarm signal are compared with the standard point table information to generate a preliminary comparison result; multi-dimensional feature information of this point-to-point process is extracted and input into a pre-trained machine learning model based on a gradient boosting decision tree to calculate the confidence score of this point-to-point result; based on the comparison result of the confidence score and a preset threshold, corresponding point-to-point decision operations are executed. The purpose of this invention is to achieve automated and intelligent point-to-point communication between substations and the dispatch master station, with self-verification capabilities.

Description

Technical Field

[0001] This invention belongs to the field of power system automation technology, specifically relating to an automatic point-to-point method and related equipment between a substation and a dispatching master station. Background Technology

[0002] In power systems, uninterrupted communication and accurate signal transmission between substations and the dispatch master station are fundamental to ensuring the safe and stable operation of the power grid. Point-to-point matching refers to the process of verifying, after the construction, renovation, or maintenance of a substation automation system, that every signal (such as switch position, alarm information, and measured value) sent to the dispatch master station from equipment within the substation (such as relay protection devices and measurement and control devices) correctly corresponds one-to-one with the signals defined in the master station's database.

[0003] The traditional point-to-point method is as follows: Substation personnel manually trigger a signal (e.g., close a switch) at the substation end according to the point-to-point schedule. This signal is transmitted to the dispatch master station via a communication management device. Dispatch master station personnel monitor the master station system interface, and after confirming receipt of the signal, inform the substation personnel by phone. The substation personnel then compare the telephone notification with their own operation to confirm the accuracy of the point. This traditional method has the following drawbacks: High labor costs: It requires the full participation of personnel from both the dispatch and substation ends, making coordination difficult and time-consuming; Low efficiency: Relying on telephone communication, information transmission is slow and easily affected by personnel's work status; Prone to errors: Manual verification may result in mishearing, misremembering, or omissions, leaving safety hazards; Lack of traceability: The point-to-point process lacks structured records, hindering problem analysis and subsequent review. Therefore, there is an urgent need for a point-to-point solution that can achieve automation, intelligence, and self-verification capabilities. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention provides an automatic point-to-point method and related equipment between substations and dispatching master stations. Its purpose is to realize the automation and intelligence of point-to-point pairing between substations and dispatching master stations, with self-verification capability, overcoming the shortcomings of traditional point-to-point pairing methods such as high labor costs, low efficiency, easy error and lack of traceability, ensuring the accuracy and efficiency of the point-to-point pairing process, and ensuring the safe and stable operation of the power grid.

[0005] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: According to a first aspect of the present invention, an automatic point-to-point method for communication between a substation and a dispatch master station is provided, comprising: The verification platform stores standard point table information in advance. The verification platform is set up on the dispatch master station side and is connected to the substation and the dispatch master station respectively. In verification mode, the verification platform sends a point-to-point trigger command to the substation; The verification platform captures the point-to-point response signal uploaded by the substation in response to the trigger command, and simultaneously captures the alarm signal generated by the dispatch master station. The point response signal and the alarm signal are compared with the standard point table information to generate preliminary comparison results; Extract the multidimensional feature information of this pairing process, and input the multidimensional feature information into a pre-trained machine learning model based on gradient boosting decision tree to calculate the confidence score of this pairing result; Based on the comparison result between the confidence score and the preset threshold, the corresponding point-to-point decision operation is performed.

[0006] In one possible implementation of the first aspect, the multidimensional feature information includes at least two of time feature factors, logical feature factors, consistency feature factors, and quality feature factors; The time characteristic factor characterizes the degree of matching between the signal transmission delay and the expected delay; The logical characteristic factor characterizes whether the signal state change conforms to the physical logic rules of the power system; The consistency feature factor characterizes the stability of the results of the same signal in multiple verifications; The quality characteristic factors characterize the bit error rate and / or packet loss rate of communication messages.

[0007] In one possible implementation of the first aspect, the gradient boosting decision tree-based machine learning model is obtained through pre-training via the following steps: Collect historical point-to-point records, extract multi-dimensional feature vectors for each historical point-to-point record, and construct corresponding true confidence labels; The true confidence label is constructed as follows: if the test is successful on the first attempt and no retry is needed, the label value is set to 1; if the test is successful after a certain number of attempts, the label value is set to 1. k If the retries are successful, the label value is set to 1 / ( k +1); if it ultimately fails, the tag value is set to 0; Using the multidimensional feature vector as input and the true confidence label as the prediction target, the mean squared error is used as the loss function. Multiple regression trees are iteratively trained through the gradient boosting algorithm, and the weighted sum of the multiple regression trees constitutes the machine learning model.

[0008] In one possible implementation of the first aspect, the step of performing a corresponding point-to-point decision operation based on the comparison result of the confidence score and a preset threshold specifically includes: If the confidence score is higher than the first threshold, the pairing is considered successful, the pairing result is recorded, and the pairing task for the next signal is triggered. If the confidence score is between the first threshold and the second threshold, it is determined that a retry is required. The current signal point-to-point instruction is automatically resent to the substation, and the number of retries is recorded. If the confidence score is lower than the second threshold, or the number of retries exceeds the preset upper limit, the point is determined to have failed and an alarm message is generated.

[0009] One possible implementation of the first aspect also includes intelligent inspection steps in the operating mode: Under normal system operation, select some signals from the point table to generate an inspection task list; According to the inspection task list, point-to-point verification is automatically performed within a preset time period; The step of selecting some signals includes: constructing a feature vector for each signal to characterize its importance and / or failure risk, calculating a comprehensive risk score, and sorting and filtering the signals from high to low according to the comprehensive risk score.

[0010] In one possible implementation of the first aspect, the step of selecting a portion of the signal further includes: Cluster analysis is performed on the high-priority signal set after sorting and filtering. The cluster features include at least one or more of the following: physical location features, logical function features, and communication path features of the signals. At least one signal is selected from each clustering result and combined to form the inspection task list.

[0011] In one possible implementation of the first aspect, the step of capturing the point response signal and the alarm signal includes: The point response signal includes point table information uploaded from the substation remote control device and alarm window text information uploaded from the substation back-end computer. The alarm signals generated by the dispatch master station include the alarm window information of the dispatch master station itself. The comparison steps specifically involve: performing a multi-source comparison of the standard point table information, the alarm window text information uploaded by the back-end machine, and the alarm window information on the dispatch master station side.

[0012] One possible implementation of the first aspect also includes: Once all point-to-point tasks are completed, a structured point-to-point report is automatically generated based on the point-to-point results and confidence scores for each signal.

[0013] In one possible implementation of the first aspect, the dispatch master station is directly connected to the substation through a first communication channel for real-time data interaction during normal operation; The substation is connected to the verification platform via a second communication channel, which is used to upload point-to-point response signals to the verification platform in verification mode. The first communication channel and the second communication channel are independent physical channels or logical channels.

[0014] According to a second aspect of the present invention, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the aforementioned automatic point-to-point method between a substation and a dispatch master station.

[0015] According to a third aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the aforementioned automatic point-to-point method between a substation and a dispatch master station.

[0016] According to a fourth aspect of the present invention, a computer program product is provided that, when executed by a processor, implements the aforementioned automatic point-to-point method between a substation and a dispatch master station.

[0017] Compared with the prior art, the present invention has at least the following beneficial effects: This invention provides an automatic point-to-point communication method between a substation and a dispatch master station. By adding a verification platform to the dispatch master station, a dedicated point-to-point communication channel is constructed between the substation, the verification platform, and the master station, while retaining the original communication channel between the substation and the master station for real-time data exchange during normal operation. This architecture separates the point-to-point communication work from the daily monitoring operations of the master station, allowing the process to be carried out independently without affecting the normal operation of the master station. This avoids the problems of traditional point-to-point communication, which requires coordination between personnel at both the dispatch and substation ends and consumes a lot of time. The verification platform automatically sends point-to-point trigger commands to the substation, automatically captures the point-to-point response signals uploaded by the substation and the alarm signals generated by the master station, and automatically completes the signal comparison. The entire process requires no manual intervention, changing the traditional operation mode of point-to-point communication, which requires on-site personnel to manually trigger signals, master station personnel to manually monitor, and then verify via telephone communication. This eliminates the information transmission delay and misunderstanding risks caused by telephone communication and improves point-to-point communication efficiency. A multi-source information comparison strategy was adopted in the signal comparison stage, simultaneously capturing point table information uploaded by the substation, alarm text information from the back-end machine, and alarm window information from the master station. This effectively identifies potential issues such as packet loss, misalignment, and delay during signal transmission, avoiding misjudgments that may occur from a single information source. A machine learning model based on gradient boosting decision trees was introduced to evaluate the confidence level of each pairing result. This model automatically learns the correlation between various features and pairing results from historical pairing data, and can calculate an objective and quantifiable confidence score based on multiple dimensions of information such as transmission delay, logical rationality, historical stability, and communication quality of the current pairing process. Compared to fixed rule judgments, the machine learning-based method can more accurately identify pairing results that appear successful but actually have hidden risks, improving the reliability of pairing quality assessment.

[0018] A tiered decision-making process is executed based on the comparison between the confidence score and a preset threshold. Results with high confidence are directly considered successful, results with medium confidence are automatically retried, and results with low confidence or exceeding the retry limit trigger an alarm. This tiered processing mechanism ensures the efficiency of high-confidence results, re-verifies questionable results, and promptly identifies and alerts to recurring anomalies, facilitating focused attention and handling by operations and maintenance personnel.

[0019] In summary, this invention automates the entire point-to-point work process, reducing manual point-to-point work that previously took days or even weeks to be completed within hours, significantly lowering labor and time costs. Simultaneously, the entire process is fully recorded and traceable, facilitating subsequent analysis and review, fundamentally solving the problems of inefficiency, error-proneness, and lack of traceability inherent in traditional point-to-point methods. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the specific embodiments of the present invention, the drawings used in the description of the specific embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a flowchart of an automatic point-to-point method between a substation and a dispatching master station according to the present invention.

[0022] Figure 2 This is a flowchart illustrating the calculation of confidence scores in an embodiment of the present invention.

[0023] Figure 3 This is a flowchart illustrating the operating mode in an embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram of an automatic point-to-point system between a substation and a dispatching master station according to the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, 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.

[0026] like Figure 1 As shown, this invention provides an automatic point-to-point method between a substation and a dispatch master station, specifically including the following steps: S1. Standard point table information is pre-stored in the verification platform. The verification platform is set on the dispatch master station side and is connected to the substation and the dispatch master station respectively.

[0027] Specifically, based on the original substation remote control unit-communication system-dispatch master station, the channel directly connecting the substation to the master station is retained, serving as the information transmission channel between the substation and the master station during normal operation. A verification platform is added at the master station. A new channel is added to the master station side of the existing communication device connecting the master and substations, connecting to this verification platform. The verification platform then connects to the dispatch master station to complete communication. This verification platform is primarily responsible for automatic point-to-point alignment and intelligent evaluation functions.

[0028] The verification platform pre-stores standard point table information. When used for telemetry point matching, the platform stores the point table; when used for information security systems, the platform stores the corresponding ICD, SCD, and other information. Since the difference between the two is not significant, the following description will only take telemetry point matching as an example.

[0029] The verification platform operates in two modes for different substations: verification mode and operation mode. When a substation is newly built or upgraded, the verification platform's point tables related to that substation run in verification mode, automatically checking the accuracy of the point tables in the master and slave stations. For other normally operating substations, the verification platform's point tables operate in operation mode, intelligently verifying important signals based on real-time conditions to check whether hardware or software issues are preventing signal transmission. Both modes can be manually set by the dispatch master station personnel.

[0030] S2. In verification mode, the verification platform sends a point-to-point trigger command to the substation.

[0031] Specifically, in verification mode, the system initializes, at which point the point table information of the master station and substation has been established. The verification platform establishes a communication connection with the substation and master station monitoring systems, and the master station transmits its own point table to the verification platform, that is, the standard point table information is pre-stored in the verification platform.

[0032] The automatic pairing task is initiated, and the verification platform sends the first pairing signal trigger command to the substation.

[0033] S3. The verification platform captures the point-to-point response signal uploaded by the substation in response to the trigger command, and simultaneously captures the alarm signal generated by the dispatch master station.

[0034] Specifically, after receiving the trigger command, the corresponding device at the substation automatically generates and sends the corresponding point-to-point response signal. Simultaneously, the back-end system also sends the alarm window signals it receives to the verification platform.

[0035] The verification platform acquires point-to-point response signals from the substation, alarm window signals from the master station, and alarm window signals from the back-end system. In other words, the verification platform captures point-to-point response signals uploaded by the substation in response to trigger commands, and simultaneously captures alarm signals generated by the dispatch master station.

[0036] S4. Compare the point response signal and the alarm signal with the standard point table information to generate a preliminary comparison result.

[0037] Specifically, during the comparison, machine learning should be performed first to enable the machine to distinguish which words have the same meaning, thus avoiding misjudgments caused by similar but potentially different information, such as "accident total", "total accidents across the entire site", or "accidents at a certain interval".

[0038] S5. Extract the multidimensional feature information of this pairing process, and input the multidimensional feature information into the pre-trained machine learning model based on gradient boosting decision tree to calculate the confidence score of this pairing result.

[0039] S6. Based on the comparison result between the confidence score and the preset threshold, perform the corresponding point-to-point decision operation.

[0040] In one possible implementation, the multidimensional feature information includes at least two of time feature factors, logical feature factors, consistency feature factors, and quality feature factors; wherein, the time feature factor characterizes the degree of matching between the signal transmission delay and the expected delay; the logical feature factor characterizes whether the signal state change conforms to the physical logic rules of the power system; the consistency feature factor characterizes the stability of the result of the same signal in multiple verifications; and the quality feature factor characterizes the bit error rate and / or packet loss rate of the communication message.

[0041] Specifically, the calculation methods for each factor are as follows: Time characteristic factor Used to characterize the degree of matching between signal transmission delay and expected delay, i.e., to assess whether the delay of the signal from the substation to the main station for reception is within a reasonable range. The calculation formula is: .in, The measured signal transmission delay is the difference between the master station receiving time and the substation transmission time. The desired standard transmission delay (can be preset according to network conditions, such as 200ms). This represents the maximum acceptable latency tolerance (e.g., 500ms). When the absolute difference between the actual latency and the expected latency exceeds this threshold, the factor drops to 0.

[0042] Logical feature factor This is used to characterize whether changes in signal state conform to the physical logic rules of the power system; that is, to evaluate whether changes in signal values ​​comply with the physical logic and rules of the power system. The calculation formula is: .in, This is the sum of penalties for all logical rules triggered during this signal verification. The maximum penalty limit set for the system (e.g., 1.0). Specific penalty rules can be created based on the substation conditions and embedded in the verification platform. For example: if a switch changes from "OFF" to "OFF" and back to "OFF" within 1 second, the penalty is +0.5; if an analog quantity (such as voltage) jumps by more than 50% of its rated value, the penalty is +0.3; if a signal value contradicts the status of other related signals (e.g., the switch is in the "OFF" position but there is current), the penalty is +0.8.

[0043] Consistency Feature Factor This is used to characterize the stability of the results of multiple verifications of the same signal, that is, whether the verification results of the same signal are stable and consistent in multiple consecutive verifications. The calculation formula is: .in, This represents the number of times the signal has been successfully verified within the current verification period. This represents the total number of times the signal has been verified within the current verification period. For the first verification of a signal, this factor can be set to 1 or a default value, and it will dynamically change as the number of retries increases.

[0044] Quality characteristic factor The bit error rate and / or packet loss rate are used to characterize communication messages, i.e., to assess the quality of the communication messages themselves. The calculation formula is: .in, Bit error rate, obtained from the communication interface; This represents the packet loss rate during this point-to-point communication process.

[0045] In one feasible approach, the weight coefficients of each factor... , , , satisfy Confidence score The calculation formula is: . The closer the value is to 1, the more reliable the verification result is; The closer the value is to 0, the less reliable the result. These weights can be set using expert experience; for example, they could be set as follows: , , , This demonstrates that logical correctness is of utmost importance.

[0046] In one possible implementation, the gradient boosting decision tree (GBDT)-based machine learning model is pre-trained through the following steps: Historical peer-to-peer records are collected, and a multi-dimensional feature vector is extracted for each historical peer-to-peer record. A corresponding true confidence label is constructed as follows: if the record is successful on the first attempt and no retry is required, the label value is set to 1; if the record is unsuccessful after several attempts... k If the retries are successful, the label value is set to 1 / ( k +1); if it ultimately fails, the tag value is set to 0; Using the multidimensional feature vector as input and the true confidence label as the prediction target, the mean squared error is used as the loss function. Multiple regression trees are iteratively trained through the gradient boosting algorithm, and the weighted sum of the multiple regression trees constitutes the machine learning model.

[0047] Specifically, this method is an advanced implementation of the present invention, which achieves intelligent evaluation through machine learning and can change the weight coefficients in real time.

[0048] The entire training process is based on the training dataset. ,in It is an eigenvector. It is a true confidence label constructed based on historical pairwise results and the number of retries.

[0049] Feature vector The extraction method is as follows: for the first i Each historical verification record is used to extract its corresponding multi-dimensional feature vector, which includes: time-related features such as delay, delay ratio, and jitter; signal attribute features such as signal type and priority; historical performance features such as historical success rate and recent consistency; and communication quality features such as bit error rate, packet loss rate, and channel load. These features are represented as feature vectors. .

[0050] True confidence label It is a continuous value, ranging between 0 (completely untrustworthy) and 1 (completely trustworthy). This label is not directly recorded, but is automatically generated based on the final result and process behavior of the previous verification. The construction method is as follows: for a verification that succeeds once and requires no retries, its result is defined as the most trustworthy. = 1; For a check that ultimately fails regardless of how many retries it is, its result is defined as the least reliable. Let... = 0; for those who have passed k A successful verification after several retries has a reliability between the two, defined as: For example: if the first attempt fails and the second attempt succeeds (k = 1), then... If it succeeds after two retries (k=2), then .

[0051] Reference Figure 2 The confidence score calculation flowchart shown below illustrates the model training process: First, initialize the model by setting it to a simple constant value, which is the mean of the labels of all training samples: This initial model is the starting point for all predictions.

[0052] Next, iterative improvement is performed, repeating the following steps M times, with each iteration generating a new decision tree (m=1 to M, where M is the total number of trees). For the current iteration m, an ensemble model consisting of the first m-1 trees is obtained. For each sample in the training set i Calculate the negative gradient of its loss function, i.e., the pseudo residual. .

[0053] The mean squared error is used as the loss function, which is defined as follows: This loss function was chosen because of its convexity, strong penalty for large errors, and natural fit with the gradient boosting framework. The negative gradient of the mean squared error is calculated as follows: Pseudo-residual The current model is intuitively represented. For the sample i How far are the predicted values ​​from the actual values? The task of the new tree is to learn and close this gap.

[0054] Use the entire training set To create a regression tree whose structure best fits the spurious residuals. The construction process is a recursive binary splitting process: starting from the root node (containing all samples), for the current node, traverse all possible split points for all features. The criterion for the optimal split is: to minimize the sum of squared errors of the two child nodes after the split. Assume a split will result in a node... S Divide into left subset and right subset The reduction in the squared error of the split is evaluated using the following formula: .in It is a node S The variance of all sample spurious residuals. Choose one that makes... The largest feature and the split point. Repeat the splitting step for the generated child nodes until a stopping condition is met (e.g., the tree reaches its maximum depth, the number of samples in the node is less than the minimum, or the gain is less than a threshold). When splitting stops, the node becomes a leaf node. Each leaf node... j This will cover a subset of the training set. For the first... m tree Each leaf node j Calculate its output value This value is determined by minimizing the loss function: This minimizes the overall loss for all samples falling into that leaf node. For the mean squared error loss function, the optimal value is... It is the average of the pseudo residuals of all samples falling into that leaf node: This calculation means that the newly planted tree For any leaf node j Each sample will output a fixed correction value. This correction value is precisely the average error generated by the previous model on the samples covered by this node. By adding this average value, the new ensemble model can systematically correct the prediction bias on these samples.

[0055] The newly trained tree It is added to the ensemble model. To control overfitting, a learning rate is introduced. (generally ), Learning rate The purpose is to reduce the contribution of each tree, allowing the model to move forward with smaller steps, although this requires more trees (number of iterations). M The model was fitted using a different approach, but the final model had stronger generalization ability and more stable predictions.

[0056] After M After rounds of iterations, the final GBDT model was obtained, consisting of the initial model and all... M The decision trees are obtained by weighted summation: The feature vectors are serialized and deployed on a cloud-based intelligent verification platform to receive real-time point-to-point feature vectors online. X and output the confidence score. CS This provides a core basis for intelligent decision-making in automatic point-to-point systems.

[0057] In one possible implementation, the step of performing the corresponding point-to-point decision operation based on the comparison result of the confidence score and the preset threshold specifically includes: If the confidence score is higher than the first threshold, the pairing is considered successful, the pairing result is recorded, and the pairing task for the next signal is triggered. If the confidence score is between the first threshold and the second threshold, it is determined that a retry is required. The current signal point-to-point instruction is automatically resent to the substation, and the number of retries is recorded. If the confidence score is lower than the second threshold, or the number of retries exceeds the preset upper limit, the point is determined to have failed and an alarm message is generated.

[0058] In other words, after calculating the confidence score CS, the following rules apply: If CS is higher than the first threshold (CS≥0.85), the pairing is considered successful. The pairing result is recorded and the pairing task for the next signal is triggered. The process jumps to determine whether all signals in the pairing table have been completed. If not, the next signal is selected and the process jumps to S2. If all signals are completed, the final pairing report is generated and the automatic pairing task ends.

[0059] If CS is between the first threshold and the second threshold, for example, 0.60≤CS<0.85, it is determined that a retry is required. The current signal point command is automatically resent to the substation (i.e., jump back to S2), and the number of retries is recorded.

[0060] If CS is lower than the second threshold, for example, CS<0.60, or the number of retries exceeds the preset limit, such as 3 times, then the pairing is determined to have failed and an alarm message is generated. The pairing failure is recorded and an alarm is triggered. The process jumps to determine whether all signals in the pairing table have been completed. If not, the next signal is selected and the process jumps to S2. If all signals are completed, the final pairing report is generated and the automatic pairing task ends.

[0061] The aforementioned hierarchical decision-making mechanism based on confidence scores enables intelligent judgment and adaptive processing of point processes.

[0062] In one feasible approach, the intelligent inspection steps in the operating mode are as follows: Under normal system operation, select some signals from the point table to generate an inspection task list; According to the inspection task list, point-to-point verification is automatically performed within a preset time period; The step of selecting some signals includes: constructing a feature vector for each signal to characterize its importance and / or failure risk, calculating a comprehensive risk score, and sorting and filtering the signals from high to low according to the comprehensive risk score.

[0063] Specifically, refer to Figure 3 The flowchart shown illustrates the operation mode. In operation mode, the verification platform uses an intelligent algorithm to select some important point table information and issues automatic point-to-point commands based on this information to ensure communication reliability.

[0064] This implementation plan uses a hybrid intelligent algorithm based on multi-factor weighted scoring and cluster analysis to achieve the above strategy. The specific process is as follows: Under normal system operation, select some signals from the point table to generate an inspection task list, and automatically perform point verification according to the inspection task list within a preset time period (such as the low load period at night).

[0065] The steps for selecting a subset of signals include: constructing a feature vector for each signal to characterize its importance and / or failure risk, calculating a comprehensive risk score, and sorting and filtering signals from high to low based on the comprehensive risk score.

[0066] Specifically, for each signal in the point table Construct a feature vector This is used to quantify its performance across various screening dimensions. .in: The influence score is defined in advance by the operation and maintenance experts based on the signal type (e.g., trip signal = 10, important alarm = 8, normal status = 5, measurement value = 3). The duration of inaction is defined as the time difference (in days) between the current time and the last correct change time, and is then normalized. This represents the historical communication quality index, calculated based on the recent signal transmission success rate and average confidence score. It indicates the current channel quality, based on real-time monitoring of bit error rate, packet loss rate, etc. This represents the environmental context coefficient, which is dynamically generated by the system based on the current season, weather, work plan, and other contexts (value range: 0.5 to 1.5).

[0067] Calculate a comprehensive risk score for each signal. A higher score indicates that the signal needs to be verified more. The calculation formula is: .in, , , , These are the weighting coefficients for each dimension, and their sum is 1. These weights can be set through historical data analysis or expert experience.

[0068] Press all signals The scores are sorted from high to low, and the top Z (e.g., 100) are selected as the high-priority candidate set to form the preliminary screening results of the inspection task list.

[0069] In one possible implementation, the step of selecting a portion of the signal further includes: Cluster analysis is performed on the high-priority signal set after sorting and filtering. The cluster features include at least one or more of the following: physical location features, logical function features, and communication path features of the signals. At least one signal is selected from each clustering result and combined to form the inspection task list to ensure the diversity of the signals being checked in terms of physical distribution and logical function.

[0070] Specifically, in order to ensure the diversity and coverage of the verification points and to avoid all the selected points being concentrated in the same protection cabinet, the same communication link, or the same device, this implementation plan performs cluster analysis on the high-priority signal set after sorting and screening.

[0071] Specifically, k-means clustering is implemented using the features of the signal on the high-priority candidate set. The feature vectors... Features include: Physical location characteristics, such as cabinet number codes, protected room / area codes, etc.

[0072] Logical functional characteristics, such as the device type code.

[0073] Signal function categories, such as classifying signals by function (e.g., "protection action", "status indication", "measurement quantity", "alarm") and encoding them.

[0074] Communication path characteristics, such as communication gateway / IP address range, are used to segment the IP address of the communication management unit or gateway where the signal is located and convert it into numerical characteristics for tag encoding.

[0075] VLAN / virtual channel number, such as directly using VLAN ID as a numerical characteristic.

[0076] After standardizing the above features, perform a k-means clustering step. Select at least one signal from each cluster result; specifically, select signals from each cluster. The top 1-2 signals with the highest scores are combined to form the final inspection task list. This ensures that this proactive verification covers different weak points in the system.

[0077] Subsequently, according to the generated inspection task list, the verification platform automatically sends instructions to the substation during preset low-load periods (such as late at night) to initiate a limited-scale automatic point-to-point task. The process is basically the same as the full point-to-point task in the verification mode: automatic signal triggering, automatic comparison, intelligent assessment of confidence level, and decision on whether to retry.

[0078] The system meticulously records the results, confidence scores, and process data for each active verification. If the failure rate of a certain type of signal (such as a long-term inactive signal) in active verification is found to be significantly higher than expected, the system can automatically increase its corresponding weight. Long-term successful verification data can be used to reduce the future failure rate of that signal. Conversely, scores indicate that a low or unsuccessful result will make it more likely to be selected in subsequent cycles.

[0079] In one possible implementation, the step of capturing the point response signal and the alarm signal includes: The point response signal includes point table information uploaded from the substation remote control device and alarm window text information uploaded from the substation back-end computer. The alarm signals generated by the dispatch master station include the alarm window information of the dispatch master station itself.

[0080] In other words, in this implementation plan, the substation includes a remote control unit (RCU) and a back-end unit. Based on the existing RTU-communication system-master station, an additional channel is added: RTU / back-end unit-communication system-verification platform-master station as an automatic point-to-point channel. The back-end unit only transmits text information from its alarm window to the verification platform, while the RTU transmits point table information to both the verification platform and the master station.

[0081] Specifically, in the steps of capturing point response signals and alarm signals: The point-to-point response signal includes point table information uploaded from the substation's remote control unit and alarm window text information uploaded from the substation's back-end computer. After receiving the point-to-point instruction from the verification platform, the corresponding device at the substation automatically generates and sends the corresponding signal, while the back-end computer also sends the alarm window signal it has received to the verification platform.

[0082] Alarm signals generated by the dispatch master station include alarm window information from the dispatch master station itself. The verification platform receives alarm window information and point table database information transmitted from the master station, point table information transmitted from the remote machine, and text information transmitted from the back-end machine.

[0083] The comparison process involves comparing the standard point table information, the alarm window text information uploaded by the backend machine, and the alarm window information from the dispatch master station. The verification platform uses machine learning to automatically align the points using these three pieces of information and outputs the results to the master station when needed. In addition to its original functions, the master station adds a function to receive the verification platform results and display them to the dispatcher.

[0084] In one feasible approach, after all peering tasks are completed, a structured peering report is automatically generated based on the peering results and confidence scores for each signal.

[0085] Specifically, in verification mode, when it is determined that all signals in the pairing table have been completed, the system generates a final pairing report and ends the automatic pairing task.

[0086] This point-to-point report includes the verification result (success / failure / success after retry) for each signal, the confidence score for each verification, the number of retries, and alarm information for abnormal situations. The report is stored in a structured format, which facilitates problem analysis and subsequent review, and solves the problem of lack of traceability in traditional manual point-to-point methods.

[0087] In one possible implementation, the dispatch master station is directly connected to the substation via a first communication channel for real-time data interaction during normal operation; this channel is originally the channel directly connecting the substation and the master station, and is retained as a channel for transmitting information between the substation and the master station during normal operation.

[0088] The substation is connected to the verification platform via a second communication channel, which is used to upload point-to-point response signals to the verification platform in verification mode. This channel is an additional channel added to the main station side of the communication device that originally connected the main and substations after the verification platform was added at the main station, and is connected to the verification platform.

[0089] The first and second communication channels are independent physical or logical channels. This independent channel setup ensures that point-to-point operation does not affect real-time data exchange between the substation and the master station during normal operation, guaranteeing the continuity of power grid monitoring.

[0090] In summary, through the above series of implementation schemes, the present invention forms a complete intelligent automatic point-to-point method, which can complete the automatic point-to-point connection of newly built substations and verify the communication status during normal operation. This greatly improves the efficiency and reliability of point-to-point connection work, increases the transparency of substations to the master station, and does not affect the normal operation of the original master station.

[0091] In one embodiment, the operation flow of the verification platform in verification mode is as follows: Step 1: System initialization (at this time, the point table information of the master station and substation has been established). The verification platform establishes a communication connection with the substation and the master station monitoring system, and the master station transmits its own point table to the verification platform.

[0092] Step 2: Start the automatic pairing task and verify that the platform sends the first pairing signal trigger command to the substation.

[0093] Step 3: After receiving the instruction, the substation automatically generates and sends the corresponding point-to-point response signal. At the same time, the back-end computer also sends the alarm window signal it has received to the verification platform.

[0094] Step 4: Verify that the platform obtains the signals sent from the substation, the alarm window signals from the master station, and the alarm window signals sent from the back-end computer.

[0095] Step 5: The verification platform compares the alarm signals from the main station with the alarm signals from the backend machine, as well as its own signal comparison engine based on the point table, to generate preliminary verification results. (During the comparison, machine learning should be performed first to enable the machine to distinguish which texts have the same meaning, avoiding misjudgments caused by similar terms such as "accident total," "total accident for the entire station," or "accident total for a certain interval," which may represent the same or different information.) Step 6: The verification platform calls the intelligent algorithm to comprehensively evaluate the verification results based on multi-dimensional data such as signal timing, logical relationship, historical consistency, and signal quality, and calculates the confidence score CS.

[0096] Step 7: The decision and control module determines the confidence score: After calculating CS, the following rules will be followed: If the value is higher than the success threshold, such as CS≥0.85, then the pair is recorded as successful and the process jumps to step 9.

[0097] If the value is below the success threshold but above the failure threshold (e.g., 0.60≤CS<0.85), it is determined that a retry is required, and the process jumps to step 8.

[0098] If the failure threshold is lower than CS < 0.60, or the number of retries exceeds the limit (e.g., 3 times), then the point-to-point failure is recorded and an alarm is triggered, and the process jumps to step 9.

[0099] Step 8: The system automatically resends the point-to-point command to the substation (i.e., jumps back to step 2) and counts the number of retries.

[0100] Step 9: Determine whether all signals in the point table have been completed. If not, select the next signal and proceed to Step 2. If all signals have been completed, proceed to Step 10.

[0101] Step 10: Generate the final point-to-point report and end this automatic point-to-point task.

[0102] In one possible implementation, such as Figure 4 As shown, the present invention provides an automatic point-to-point system between a substation and a dispatch master station, specifically including: The dispatching master station, the substation, and the verification platform set up on the side of the dispatching master station; The dispatch master station is directly connected to the substation through the first communication channel for real-time data interaction during normal operation. The substation is connected to the verification platform via a second communication channel, which is used to upload point-to-point response signals to the verification platform in verification mode. The verification platform is communicatively connected to both the dispatch master station and the substation, and is used to store standard point table information, including: The signal receiving module is used to capture the point-to-point response signal uploaded by the substation and the alarm signal generated by the dispatch master station. The verification module is used to compare the point response signal and the alarm signal with the standard point table information to generate a preliminary comparison result; The intelligent evaluation module incorporates a machine learning model based on gradient boosting decision trees to extract multi-dimensional feature information from the current pairing process and input it into the machine learning model to calculate the confidence score of the current pairing result; and The instruction issuing module is used to perform corresponding point-to-point decision operations based on the comparison result between the confidence score and the preset threshold.

[0103] All relevant content regarding the steps involved in the aforementioned embodiment of the automatic point-to-point method between a substation and a dispatching master station can be referenced from the functional description of the corresponding functional module of the automatic point-to-point device between a substation and a dispatching master station in this embodiment of the invention, and will not be repeated here. The module division in this embodiment of the invention is illustrative and is merely a logical functional division. In actual implementation, there may be other division methods. Furthermore, the functional modules in the various embodiments of the invention can be integrated into a processor, exist as separate physical entities, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or as a software functional module.

[0104] In another embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used for the operation of an automatic point-to-point method between a substation and a dispatch master station.

[0105] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be Random Access Memory (RAM) or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the automatic point-to-point method between a substation and a dispatch master station in the above embodiments.

[0106] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, etc.) containing computer-usable program code.

[0107] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0108] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0109] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0110] This invention also provides a computer program product, which is used to execute any of the above-described automatic point-to-point methods between a substation and a dispatching master station. Since the computer program product provided by this invention and the above-described automatic point-to-point method between a substation and a dispatching master station belong to the same inventive concept, the computer program product provided by this invention has all the advantages of the above-described automatic point-to-point method between a substation and a dispatching master station. Therefore, the beneficial effects of the computer program product provided by this invention will not be elaborated upon here.

[0111] In this invention, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0112] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the scope of the technology disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention.

Claims

1. An automatic point-to-point method between a substation and a dispatch master station, characterized in that, include: The verification platform stores standard point table information in advance. The verification platform is set up on the dispatch master station side and is connected to the substation and the dispatch master station respectively. In verification mode, the verification platform sends a point-to-point trigger command to the substation; The verification platform captures the point-to-point response signal uploaded by the substation in response to the trigger command, and simultaneously captures the alarm signal generated by the dispatch master station. The point response signal and the alarm signal are compared with the standard point table information to generate preliminary comparison results; Extract the multidimensional feature information of this pairing process, and input the multidimensional feature information into a pre-trained machine learning model based on gradient boosting decision tree to calculate the confidence score of this pairing result; Based on the comparison result between the confidence score and the preset threshold, the corresponding point-to-point decision operation is performed.

2. The automatic point-to-point method between a substation and a dispatch master station according to claim 1, characterized in that, The multidimensional feature information includes at least two of the following: time feature factors, logical feature factors, consistency feature factors, and quality feature factors; The time characteristic factor characterizes the degree of matching between the signal transmission delay and the expected delay; The logical characteristic factor characterizes whether the signal state change conforms to the physical logic rules of the power system; The consistency feature factor characterizes the stability of the results of the same signal in multiple verifications; The quality characteristic factors characterize the bit error rate and / or packet loss rate of communication messages.

3. The automatic point-to-point method between a substation and a dispatch master station according to claim 1, characterized in that, The machine learning model based on gradient boosting decision trees is obtained through the following steps: Collect historical point-to-point records, extract multi-dimensional feature vectors for each historical point-to-point record, and construct corresponding true confidence labels; The true confidence label is constructed as follows: if the test is successful on the first attempt and no retry is needed, the label value is set to 1; if the test is successful after a certain number of attempts, the label value is set to 1. k If the retries are successful, the label value is set to 1 / ( k +1); if it ultimately fails, the tag value is set to 0; Using the multidimensional feature vector as input and the true confidence label as the prediction target, the mean squared error is used as the loss function. Multiple regression trees are iteratively trained through the gradient boosting algorithm, and the weighted sum of the multiple regression trees constitutes the machine learning model.

4. The automatic point-to-point method between a substation and a dispatch master station according to claim 1, characterized in that, The step of performing the corresponding point-to-point decision operation based on the comparison result of the confidence score and the preset threshold specifically includes: If the confidence score is higher than the first threshold, the pairing is considered successful, the pairing result is recorded, and the pairing task for the next signal is triggered. If the confidence score is between the first threshold and the second threshold, it is determined that a retry is required, and the current signal point-to-point instruction is automatically resent to the substation, and the number of retries is recorded. If the confidence score is lower than the second threshold, or the number of retries exceeds the preset upper limit, the point is determined to have failed and an alarm message is generated.

5. The automatic point-to-point method between a substation and a dispatch master station according to claim 1, characterized in that, It also includes intelligent inspection steps in the operating mode: Under normal system operation, select some signals from the point table to generate an inspection task list; According to the inspection task list, point-to-point verification is automatically performed within a preset time period; The step of selecting some signals includes: constructing a feature vector for each signal to characterize its importance and / or failure risk, calculating a comprehensive risk score, and sorting and filtering the signals from high to low according to the comprehensive risk score.

6. The automatic point-to-point method between a substation and a dispatch master station according to claim 5, characterized in that, The step of selecting a portion of the signal further includes: Cluster analysis is performed on the high-priority signal set after sorting and filtering. The cluster features include at least one or more of the following: physical location features, logical function features, and communication path features of the signals. At least one signal is selected from each clustering result and combined to form the inspection task list.

7. The automatic point-to-point method between a substation and a dispatch master station according to claim 1, characterized in that, In the steps of capturing point response signals and alarm signals: The point response signal includes point table information uploaded from the substation remote control device and alarm window text information uploaded from the substation back-end computer. The alarm signals generated by the dispatch master station include the alarm window information of the dispatch master station itself. The comparison steps specifically involve: performing a multi-source comparison of the standard point table information, the alarm window text information uploaded by the back-end machine, and the alarm window information on the dispatch master station side.

8. The automatic point-to-point method between a substation and a dispatch master station according to claim 1, characterized in that, Also includes: Once all point-to-point tasks are completed, a structured point-to-point report is automatically generated based on the point-to-point results and confidence scores for each signal.

9. The automatic point-to-point method between a substation and a dispatch master station according to claim 1, characterized in that, The dispatch master station is directly connected to the substation through the first communication channel for real-time data interaction during normal operation. The substation is connected to the verification platform via a second communication channel, which is used to upload point-to-point response signals to the verification platform in verification mode. The first communication channel and the second communication channel are independent physical channels or logical channels.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements an automatic point-to-point method between a substation and a dispatch master station as described in any one of claims 1 to 9.

Images