A remote control system for power and electrical equipment based on the Internet of Things

Abstract

The present invention belongs to the technical field of remote control of electric equipment, and particularly relates to remote fault diagnosis and control technology of electric equipment. Specifically, a remote control system for electric power equipment based on the Internet of Things is disclosed. The diagnostic effect, network quality and fault deterioration evaluation are performed by retrieving similar fault diagnosis records of electric equipment using comprehensive diagnostic rules. Thus, on the basis of a constructed dynamic network status diagnostic utility evaluation algorithm, historical fault diagnosis evaluation data are used to perform a comprehensive diagnostic rule utility compliance evaluation. When the comprehensive diagnostic rule utility does not meet the standard, the diagnostic rule is optimized and adjusted. The specific adjustment strategy is to select comprehensive diagnostic rules in a strong network state to ensure the integrity of the diagnosis, and to select simplified diagnostic rules in a weak network state to improve the diagnostic efficiency and adaptability. The applicability and pertinence of the fault diagnosis system are significantly enhanced, thereby maximizing the diagnostic quality.

Description

Technical Field

[0001] The present invention belongs to the technical field of remote control of electric equipment, and in particular to remote fault diagnosis and control technology of electric equipment, and specifically discloses a remote control system of electric power equipment based on the Internet of Things. Background Art

[0002] As a key infrastructure in modern society, the power system provides energy support for industrial production, residents' lives and public services, including core components such as overhead high-voltage lines and computer rooms. With the rapid development of social economy, the scale of the power system continues to expand, the types of equipment are becoming increasingly rich, and the operating environment is becoming more complex. In this context, the reliability and stability of power equipment are directly related to the safe operation and power supply quality of the power system.

[0003] However, power equipment is bound to fail or deteriorate in performance during long-term operation. Traditional fault diagnosis relies on manual inspection and on-site maintenance. However, due to the widespread distribution of power equipment, its geographical location is dispersed and some are located in remote or harsh environments, resulting in low efficiency and high cost of manual inspection. At the same time, it is difficult to detect potential hidden dangers in time, which may cause the spread of faults or even major accidents. To overcome the shortcomings of traditional methods, remote diagnosis and control technology has gradually become a mainstream solution.

[0004] There are already some mature solutions in the existing remote fault diagnosis and control technology of power equipment. For example, the Chinese invention patent with publication number CN103558818B proposes a remote monitoring and control system for high-voltage ring network cabinets. The system aims at intelligent operation and management of power grids, integrates multiple online monitoring technologies, deploys monitoring cameras, fans and solar energy equipment in the switch station, and can view the monitoring information and fan operation status in real time in the switch station, and realize the early warning, alarm and post-analysis functions of abnormalities and faults, while supporting dynamic evaluation of equipment health status and remote operation control.

[0005] In addition, the Chinese invention patent with publication number CN113778005B discloses a data room acquisition monitoring method and system, which establishes an equipment fault judgment database, uses the first data acquisition node and the second data acquisition node to communicate with the terminal, matches the collected data with the fault judgment database, and quickly determines whether the power equipment has a fault. Once a fault is detected, the system can automatically control the monitoring device to move to a specified location, remotely monitor the fault point in real time, and display the abnormal information of the power equipment and its specific location on the terminal.

[0006] However, the above two solutions mainly focus on the monitoring of equipment status and the optimization of fault diagnosis logic in the remote fault diagnosis control of power equipment, while ignoring the impact of remote network status on the fault diagnosis process. In fact, the implementation of remote fault diagnosis has a strong dependence on network quality. This is because modern fault diagnosis usually involves complex algorithm processing, which has high requirements for the stability and bandwidth of data transmission. In the actual operating environment of power equipment, since the equipment is often distributed in areas with complex geographical conditions, changeable climate or severe electromagnetic interference, the network quality is easily affected by external factors and deteriorates. In this case, if the fixed original fault diagnosis rules are still used, it may lead to increased diagnosis time, diagnosis interruption or error, resulting in a decrease in diagnosis quality, which in turn affects the timeliness of power equipment fault recovery. Summary of the invention

[0007] In order to solve the above technical problems or at least partially solve the above technical problems, the present application provides a remote control system for power electrical equipment based on the Internet of Things, which dynamically adjusts the fault diagnosis rules according to the changes in network quality, thereby achieving efficient and accurate completion of remote fault diagnosis tasks of power equipment under various network conditions.

[0008] The purpose of the present invention can be achieved through the following technical solutions: A remote control system for power and electrical equipment based on the Internet of Things, including: a historical diagnosis and analysis module: retrieving similar fault diagnosis records of the equipment applying comprehensive diagnostic rules to perform diagnostic effect scoring, network quality scoring and fault deterioration rate evaluation.

[0009] Comprehensive diagnosis evaluation module: Build a dynamic network comprehensive diagnosis effectiveness evaluation algorithm, and evaluate the effectiveness of comprehensive diagnosis rules by importing evaluation data of similar fault diagnosis records.

[0010] Minimum available network determination module: When the effectiveness of comprehensive diagnosis rules does not meet the standards, the same type of fault diagnosis records are divided into valid and invalid diagnosis records, and then the minimum available network for comprehensive diagnosis is determined through the valid diagnosis records.

[0011] Diagnostic rule optimization module: Generates simplified diagnostic rules by comparing the diagnostic steps of valid and invalid diagnostic records, and divides the network status into strong and weak states according to the minimum available network, thereby establishing a selection strategy for fault diagnostic rules.

[0012] Fault diagnosis implementation module: synchronously monitors faults and network status during equipment operation, and uses selected strategies to perform fault diagnosis based on network status when a fault is identified.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) Based on the fault diagnosis effect, fault deterioration rate and dynamic network status evaluation data in the historical fault diagnosis records, the present invention constructs a utility judgment algorithm to realize a data-driven diagnosis rule optimization mechanism. When the utility of the comprehensive diagnosis rules does not meet the standards, the algorithm can promptly identify and screen out the diagnosis scenarios or rule conditions that do not meet the standards, thereby triggering the rule adjustment process, avoiding diagnosis delays or interruptions caused by improper rules, and accelerating fault recovery.

[0014] (2) When the effectiveness of the comprehensive diagnosis rules does not meet the standards, the present invention divides the network status into a strong network status and a weak network status by analyzing the valid and invalid records in the historical fault diagnosis records, and then dynamically selects the diagnosis rules according to the real-time network status, thereby achieving accurate matching of the fault diagnosis rules under different network conditions, significantly enhancing the applicability and pertinence of the fault diagnosis system, and thus maximizing the quality of diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0016] Figure 1 It is a schematic diagram of the connection of each module of the system of the present invention.

[0017] Figure 2 This is an Internet of Things architecture diagram composed of system modules in the present invention.

[0018] Figure 3 It is a schematic diagram of the composition and calling of the database in the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] The present invention proposes a remote control system for electric power equipment based on the Internet of Things, comprising a historical fault analysis module, a comprehensive diagnosis and evaluation module, a minimum available network determination module, a diagnosis rule optimization module and a fault diagnosis implementation module.

[0021] See also Figure 1As shown in Figure 1, the modules mentioned above work closely together through data flow and control flow. The specific connections are as follows: (1) Historical fault analysis module → comprehensive diagnosis and judgment module functional connection: The historical fault analysis module is responsible for deep mining of historical fault diagnosis records, extracting key data (such as fault diagnosis effect, fault deterioration rate, dynamic network status, etc.), and passing these data analysis results to the comprehensive diagnosis and judgment module.

[0022] Data flow: The output key data is used as input to support the comprehensive diagnosis evaluation module to evaluate the effectiveness of the current diagnosis rules.

[0023] (2) Comprehensive diagnosis and evaluation module → minimum available network determination module

[0024] Functional connection: The comprehensive diagnosis evaluation module determines whether the current comprehensive diagnosis rules are effective based on the historical data analysis results. If the effectiveness is not effective, the minimum available network determination module is triggered to clarify the minimum network quality requirements required by the comprehensive diagnosis rules.

[0025] Data flow: The comprehensive diagnosis and evaluation module transmits the utility evaluation results (including network status related indicators) to the minimum available network determination module to define the criteria for dividing weak and strong states.

[0026] (3) Minimum available network determination module → Diagnosis rule optimization module

[0027] Functional connection: After the minimum available network determination module clarifies the network status classification, it sends the classification results to the diagnosis rule optimization module. This module generates or adjusts the diagnosis rules applicable to different network status based on the network status (strong / weak) and historical data analysis results.

[0028] Data flow: The network status classification information output by the minimum available network determination module is used as input to guide the diagnosis rule optimization module to generate simplified diagnosis rules.

[0029] (4) Diagnosis rule optimization module → fault diagnosis implementation module

[0030] Functional connection: The diagnostic rules (including comprehensive diagnostic rules and simplified diagnostic rules) generated by the diagnostic rule optimization module are passed to the fault diagnosis implementation module. This module selects appropriate diagnostic rules according to the real-time network status and performs specific fault diagnosis operations.

[0031] Data flow: The optimized diagnostic rules serve as input to provide execution basis for the fault diagnosis implementation module.

[0032] (5) Fault diagnosis implementation module → Historical fault analysis module

[0033] Functional connection: After completing a diagnostic task, the fault diagnosis implementation module feeds back the data in the diagnostic process (such as diagnostic effect, network status, fault type, etc.) to the historical fault analysis module to form a closed-loop data link.

[0034] Data flow: The real-time diagnostic data output by the fault diagnosis implementation module is used as new historical data input to support subsequent historical fault diagnosis analysis and rule optimization iteration.

[0035] The collaborative work of the above modules fully reflects the core characteristics of the Internet of Things, including perception, transmission, processing and control. The specific Internet of Things architecture is as follows: Figure 2 shown.

[0036] The historical fault analysis module is used to retrieve similar fault diagnosis records of equipment applying comprehensive diagnostic rules to perform diagnostic effect scoring, network quality scoring and fault deterioration rate assessment.

[0037] It should be explained that the comprehensive diagnostic rules mentioned above are a set of systematic, standardized operating procedures and logical specifications designed in advance based on equipment characteristics and for different fault representations in order to ensure the completeness, accuracy and reliability of the diagnostic results in the diagnosis of power and electrical equipment faults. The rules specify in detail all the standardized diagnostic steps required when the equipment fails, and by fully covering possible fault points and their associated factors, it achieves all-round detection and in-depth analysis of equipment faults, thereby accurately identifying the root cause of the fault. Each specific fault representation corresponds to a set of special comprehensive diagnostic rules to ensure the high pertinence and effectiveness of the diagnostic process.

[0038] It should also be understood that the same type of fault diagnosis records mentioned above refer to a set of records with the same or similar fault representations extracted from historical fault diagnosis data. For example, fault diagnosis records caused by abnormal current are classified into one category, while fault diagnosis records related to abnormal indicator lights constitute another category. Such records are classified based on the consistency of fault characteristics, aiming to avoid rule inconsistency and optimization deviation problems caused by mixing historical records of different fault representations during the optimization of fault diagnosis rules, thereby ensuring the accuracy and applicability of diagnosis rules.

[0039] It should be noted that fault diagnosis records refer to data documents or information collections that are automatically generated when a fault occurs in power equipment and comprehensive diagnostic rules are used to locate the cause of the fault. Such records retain various key data in the diagnosis process in detail, including but not limited to the following: Diagnosis step process data: records the execution and results of each standardized step in the diagnosis process, such as parameter collection, data analysis, preliminary fault screening and in-depth analysis, etc. (such as diagnosis time, error log, etc.).

[0040] Evolutionary data of fault characterization: captures the dynamic change process of fault characteristics from initial manifestation to final confirmation, such as the performance data of current anomalies at different times.

[0041] Remote network status data: reflects the status information of the network environment during the diagnosis process, such as key indicators such as latency, packet loss rate, and signal strength.

[0042] Through these detailed data records, fault diagnosis records provide important reference for subsequent fault analysis and rule optimization.

[0043] In the optimized implementation of the above operation, the diagnostic effect scoring process is as follows: define the diagnostic effect parameters as diagnostic efficiency parameters and diagnostic error parameters, and assign weights to each type of parameters.

[0044] Specifically, the diagnostic efficiency parameter reflects the time or resource consumption efficiency of the diagnostic process, and is usually expressed in terms of diagnostic duration.

[0045] The diagnostic error parameter reflects the frequency of errors occurring during the statistical diagnosis process, usually expressed as a false alarm rate.

[0046] In another specific implementation, the weights assigned to each type of parameter can be assigned based on the scenario priority, specifically, judging which is more important, diagnostic efficiency or diagnostic error, according to the specific application scenario of the device. For example, power equipment has high real-time requirements and can tolerate certain errors (such as fault diagnosis in the operation of transmission network equipment). In this scenario, rapid response is a key goal because faults in the power system may spread rapidly, causing power outages or equipment damage on a larger scale. Although false alarms will bring certain waste of resources or risks of misoperation, their impact is relatively small compared to the failure to detect the fault in time. Therefore, the importance of diagnostic efficiency is significantly higher than that of diagnostic error. The diagnostic efficiency parameter can be assigned a weight of 0.7 and the diagnostic error parameter can be assigned a weight of 0.3.

[0047] For example, when power equipment is in a scenario with high reliability requirements (such as maintenance of key equipment, such as nuclear power plants or important substation equipment), false alarms may lead to wrong maintenance decisions and even cause secondary problems. Therefore, the impact of diagnostic errors is far greater than diagnostic efficiency. At this time, the importance of diagnostic errors is significantly higher than diagnostic efficiency. The weight of the diagnostic efficiency parameter can be assigned to be 0.3, and the weight of the diagnostic error parameter can be assigned to be 0.7.

[0048] The above selection of 0.7 and 0.3 as the weight distribution mainly indicates that there is a significant difference in the importance of the two, but it does not completely ignore the secondary factors. This gap not only reflects the dominant role of the main factors, but also retains the focus on the secondary factors, avoiding the system being too one-sided due to extreme weight distribution (such as 0.9 and 0.1).

[0049] Of course, the above weight distribution values ​​are not completely fixed. The system performance can be evaluated by simulating the diagnosis process under different weight combinations. Specifically, a simulation environment can be built to simulate fault diagnosis tasks under different scenarios.

[0050] Use different weight combinations (such as 0.6 / 0.4, 0.7 / 0.3, 0.8 / 0.2, etc.) for diagnosis.

[0051] Record and compare the diagnostic effect parameters under each combination (such as diagnostic time, false alarm rate, etc.)

[0052] Compare the results of different weight combinations to determine the optimal ratio.

[0053] The data of each diagnostic effect parameter is extracted from similar fault diagnosis records, and each data is normalized.

[0054] The purpose of the above normalization process is to ensure that all parameter values ​​are in the same dimensionless range (such as 0 to 1), which helps to make fair comparisons and comprehensive evaluations between different parameters. Generally, for positive evaluation parameters (such as diagnostic efficiency parameters), the larger the value, the better the diagnostic effect. The normalization formula is as follows: ,in is the original parameter value, , Respectively represent the historical maximum and minimum values ​​of the parameter. is the normalized value.

[0055] However, for parameters that require reverse evaluation (such as diagnostic error parameters), the smaller the value, the better the diagnostic effect. The following formula can be used for normalization: .

[0056] In the above normalization example, assume that the historical data of the diagnostic efficiency parameter is , , current diagnosis time , calculated by normalization .

[0057] Assume that the historical data of the diagnostic error parameter is , , the current false alarm rate , calculated by normalization .

[0058] It should be added that the normalization method used above is a linear transformation method, and its core advantage is that it can completely preserve the relative relationship between data and the original distribution characteristics, thereby ensuring the consistency of data before and after transformation. For example, if there are significantly larger or smaller values ​​in the original data, these values ​​will still maintain their original relative differences after normalization, which is particularly important for diagnostic effect evaluation because it can accurately reflect the actual gap between different diagnostic parameters, thereby improving the accuracy and reliability of the evaluation results.

[0059] In addition, the normalization method uniformly maps all data to the interval [0, 1]. This feature not only eliminates the influence of different dimensions on the calculation, but also greatly simplifies the comprehensive calculation process of the diagnostic effect. By standardizing various parameters to the same scale, it is convenient for subsequent weight allocation and multi-index fusion analysis, thus providing a scientific and efficient evaluation framework for the system.

[0060] Incorporate each diagnostic effect parameter data after normalization in the fault diagnosis record into the weighted comprehensive scoring model Calculating the diagnostic performance score , where , They represent the normalized diagnostic efficiency parameter and diagnostic error parameter respectively. , They represent the weight factors assigned to the diagnostic efficiency parameter and the diagnostic error parameter respectively.

[0061] In the above normalized calculation example, assuming that the weight factors assigned to the diagnostic efficiency parameter and the diagnostic error parameter are 0.7 and 0.3, the normalized data of the diagnostic efficiency parameter and the diagnostic error parameter are 0.18 and 0.56 respectively. The diagnostic effect score obtained by substituting into the weighted comprehensive scoring model is .

[0062] Further preferably, the network quality is evaluated in the following process: determining a network status indicator, and establishing a grade scoring set for each network status indicator, each grade scoring set comprising a plurality of scoring value intervals, and each interval corresponding to a scoring value.

[0063] The network status indicators mentioned above include but are not limited to delay time, packet loss rate, jitter, etc.

[0064] The exemplary network status indicator rating sets are shown in Table 1, Table 2, and Table 3.

[0065] Table 1: Delay duration rating set

[0066]

[0067] In Table 1, the shorter the delay time, the better the network quality, so PDU1>PDU2>PDU3>PDU4.

[0068] Table 2: Packet loss rate rating set

[0069]

[0070] In Table 2, the smaller the packet loss rate, the better the network quality, so PPI1>PPI2>PPI3>PPI4.

[0071] Table 3: Jitter rating set

[0072]

[0073] In Table 3, the smaller the jitter, the better the network quality, so PS1>PS2>PS3>PS4.

[0074] The design of the above-mentioned rating set usually refers to international or industry standards, such as the MOS (Mean Opinion Score) rating system: used to evaluate users' subjective feelings about service quality, usually mapping network performance to a rating range of 1-5 points.

[0075] The data of each network status indicator is extracted from the same type of fault diagnosis records and matched with the corresponding grade score set to obtain the score value of each network status indicator.

[0076] It should be explained that the network status indicators extracted from the fault diagnosis records are matched through the graded scoring set to obtain the score value of each indicator, rather than using a normalized processing method similar to the diagnostic effect scoring. This is mainly because the network status indicators are diverse and complex. Specifically, first of all, network status indicators usually cover multiple dimensions (such as delay, packet loss rate, jitter, etc.), and the dimensions and measurement units of these indicators are different. If all indicators are uniformly normalized, some important characteristics may be weakened or lost, thereby affecting the accuracy of the scoring results; secondly, although normalization can simplify the data, it may mask the differences between different indicators to a certain extent. For example, some key indicators (such as high latency) may have a much greater impact on fault diagnosis than other indicators (such as slight jitter). Directly using the graded scoring set matching method can more accurately reflect the actual significance of each indicator according to the predefined scoring rules. In addition, this method makes the scoring results more interpretable and facilitates subsequent analysis and decision-making. In contrast, the normalized values ​​may be difficult to intuitively correspond to specific network status characteristics.

[0077] Therefore, for network status evaluation involving a large number of indicators, the method of matching the grade score set item by item is more efficient than the complex normalization algorithm, and can reduce the computational overhead while ensuring the accuracy of the score.

[0078] The score value of each network status indicator in the fault diagnosis record is calculated comprehensively to obtain the network quality score of the fault diagnosis record.

[0079] Specifically, when the score value of each network status indicator in the fault diagnosis record is used for comprehensive calculation to obtain the network quality score of the fault diagnosis record, the geometric mean method can be used for calculation, wherein the geometric mean method is to multiply the score value of each network status indicator and then take the square root. This method is used to score network quality because the geometric mean is affected by all the score values, and any score value that is too low will significantly lower the overall result. Therefore, it can reflect the balance between various indicators. At the same time, the geometric mean method is more sensitive to extreme values ​​(especially minimum values), but it will not be completely dominated by extreme values ​​like the arithmetic mean method.

[0080] In network quality assessment, various indicators (such as latency, packet loss rate, jitter, etc.) usually need to achieve a certain balance to ensure a good user experience. For example, if the latency is low but the packet loss rate is high, users may experience lag or data loss.

[0081] If jitter is high but other metrics are normal, real-time communications may still be affected.

[0082] The geometric mean method ensures that the deterioration of any indicator will significantly reduce the overall score by multiplying the score values, thereby emphasizing the balance between network status indicators.

[0083] Network quality is usually a system with obvious "short board effect", that is, the performance of the entire system is often determined by the worst indicator. The characteristics of the geometric mean method are exactly in line with this - when a certain indicator performs poorly, the overall score will drop significantly, thus truly reflecting the actual level of network quality.

[0084] As an example of calculating the network quality score, assuming that the scores of delay, packet loss rate, and jitter in the fault diagnosis record are 2, 3, and 4 respectively, the network quality score obtained is .

[0085] Still further preferably, the fault evolution refers to the following evaluation process: extracting a time evolution sequence of fault representations corresponding to the fault diagnosis process from similar fault diagnosis records, the sequence usually including multiple time points and their corresponding fault representations.

[0086] In the above preferred implementation example, assuming that the fault characteristic is current abnormality, the fault characteristic time evolution sequence is a time series set of current.

[0087] The fault characterization data of the first time and the last time as well as the fault diagnosis duration are extracted from the time evolution sequence of the fault characterization.

[0088] Substitute the fault characterization data of the first and last digits and the fault diagnosis time into the fault deterioration rate evaluation formula Get the fault deterioration rate , where , They represent the fault characterization data of the first time and the last time in the fault characterization time evolution sequence respectively. Indicates the fault diagnosis time. Indicates the correction factor.

[0089] It should be noted that during the fault diagnosis process, the fault representation does not remain static, but changes continuously as the fault continues to develop. This phenomenon is called fault deterioration. By evaluating the fault deterioration rate, the development speed of the fault during the diagnosis process can be quantitatively described, thereby providing an important basis for fault trend prediction and maintenance decision-making. In the above fault deterioration rate evaluation formula, It represents the change of fault characterization data in the whole diagnosis process, reflecting the dynamic evolution characteristics of the fault state. At the same time, considering that the actual fault deterioration often presents complex nonlinear characteristics rather than simple linear relationships, a correction coefficient is introduced to adjust the calculation results to make them more consistent with the actual deterioration trend. The specific value of the correction coefficient can be obtained by fitting experimental data or based on historical data analysis, thereby improving the accuracy and applicability of the evaluation model.

[0090] In the above calculation example of fault deterioration rate, it is assumed that the fault is characterized by abnormal current. , , , , then the fault deterioration rate is A / second means that the abnormal increase in current is 0.008 amperes per second.

[0091] The comprehensive diagnosis evaluation module is used to construct a dynamic network comprehensive diagnosis effectiveness evaluation algorithm, and to perform comprehensive diagnosis rule effectiveness evaluation by importing evaluation data of similar fault diagnosis records.

[0092] In the improved implementation of the above scheme, the dynamic network comprehensive diagnostic utility evaluation algorithm is constructed as follows: the network quality score, diagnostic effect score and fault deterioration rate of the same fault diagnosis record are used to generate a network quality change curve, a diagnostic effect change curve and a fault deterioration change curve in a coordinate system with the fault diagnosis record constructed in chronological order as the horizontal axis and the network quality score, diagnostic effect score and fault deterioration rate as the vertical axis.

[0093] Inflection points are captured on the network quality change curve, and among all captured inflection points, inflection points that only indicate a decrease in network quality are further screened out, and the corresponding time points and the fault diagnosis records at the time points are also screened out.

[0094] The fault diagnosis record corresponding to the selected inflection point is mapped onto the diagnostic effect change curve to obtain the tangent slope sign of the mapping point as the diagnostic effect change direction corresponding to the network quality degradation point.

[0095] The fault diagnosis record corresponding to the selected inflection point is mapped onto the fault deterioration change curve to obtain the tangent slope sign of the mapping point as the fault deterioration change direction corresponding to the network quality degradation point.

[0096] The number of network quality degradation points selected on the network quality change curve is counted, and the change direction of the diagnostic effect and the change direction of the fault deterioration corresponding to each network quality degradation point are compared. The proportion of the decrease in fault diagnosis effect and the increase in fault deterioration when the network quality deteriorates are summarized and recorded as p1 and p2 respectively.

[0097] According to the preset thresholds p1 and p2, where p1 and p2 represent the setting thresholds of the fault diagnosis effect decline and fault deterioration increase ratio respectively, if and It is judged that the effectiveness of the comprehensive diagnosis rule meets the standard, otherwise it does not meet the standard.

[0098] It should be added that when setting the p1 threshold and p2 threshold, the evaluation criteria can be adjusted according to actual needs. For example, in scenarios with high reliability requirements, stricter thresholds can be set; in general scenarios, the criteria can be appropriately relaxed.

[0099] Next, we import the evaluation data of similar fault diagnosis records to determine whether the comprehensive diagnosis rules meet the effectiveness standards, see Table 4.

[0100] Table 4: Evaluation of the effectiveness of comprehensive diagnosis rules for partial fault characterization

[0101]

[0102] It should be emphasized that the dynamic network comprehensive diagnosis effectiveness evaluation algorithm constructed above mainly evaluates whether the ratio of the decline in diagnostic effect and the increase in fault deterioration in the fault diagnosis record when the network quality deteriorates exceeds the set threshold. If the ratio is higher than the threshold, the comprehensive diagnosis rule is judged to be ineffective. In other words, when the network quality deteriorates, the comprehensive diagnosis rule can be judged to be ineffective when the diagnostic effect decreases or the fault deterioration increases, and the comprehensive diagnosis rule can be judged to be invalid in this scenario. This evaluation method is relatively strict and aims to minimize the omission of situations where the comprehensive diagnosis rule is not applicable.

[0103] However, in actual operation, a broader evaluation standard can be adopted to improve practicality. Specifically, by quantifying the extent to which the diagnostic effect decreases when the network quality decreases and the extent to which the fault deterioration increases, the comprehensive diagnostic rule is judged to be ineffective only when the network quality decreases accompanied by a significant decrease in the diagnostic effect or a significant increase in the fault deterioration. This approach can ensure the accuracy of the evaluation while being more in line with the needs of the actual operating environment, thereby providing more operational guidance for the optimization and application of comprehensive diagnostic rules.

[0104] The dynamic network comprehensive diagnosis effectiveness evaluation algorithm is based on the actual data of network quality score, diagnosis effect score and fault deterioration rate, and uses a data-driven approach for analysis. By identifying the inflection points in the data and statistically analyzing the change trend, the algorithm can objectively evaluate whether there is a synchronous correlation between the diagnosis effect and the fault deterioration in the context of dynamic network quality degradation.

[0105] It should be understood that in a dynamic network environment, when the network quality decreases, if the diagnostic effect does not change significantly but the degree of fault deterioration increases, this may be attributed to the high robustness of the diagnostic system. Specifically, the performance of the diagnostic system shows a strong tolerance for fluctuations in network quality, such as by relying on local cache data for analysis, thereby reducing dependence on real-time network transmission and ensuring that its diagnostic capabilities remain stable under adverse conditions. However, fault characterization is often affected by the external environment and has the characteristics of self-aggravation. In this case, although the diagnostic effect is not affected, due to the accelerated deterioration of the fault itself, the existing comprehensive diagnostic rules may not be able to effectively respond, and further optimization is needed to improve diagnostic efficiency and shorten the duration of fault deterioration. Therefore, in the context of network quality degradation, even if the diagnostic effect has not degraded, if the fault characterization shows a trend of deteriorating too quickly, it means that the actual effectiveness of the current comprehensive diagnostic rules has not met the standards and still needs to be improved to better adapt to the complex and changing operating environment.

[0106] The minimum available network determination module is used to classify similar fault diagnosis records into valid and invalid diagnosis records when the effectiveness of the comprehensive diagnosis rules does not meet the standards, and then determine the minimum available network for comprehensive diagnosis through the valid diagnosis records.

[0107] In a possible implementation of the above scheme, the same type of fault diagnosis records are divided into valid and invalid diagnostic records according to the following process: the diagnostic effect score values ​​and fault deterioration rates of the same type of fault diagnosis records are compared with the set effective critical values, thereby screening out the fault diagnosis records whose diagnostic effect score values ​​and fault deterioration rates both meet the effective critical values ​​as valid diagnostic records, and the other fault diagnosis records are regarded as invalid diagnostic records.

[0108] The above-mentioned effective critical values ​​of the diagnostic effect score and the fault deterioration rate can be determined by statistical methods, as follows: Critical value of the diagnostic effect score: By calculating the distribution characteristics of the diagnostic effect score (such as mean, standard deviation, quantile, etc.), a reasonable threshold is set in combination with business needs. For example, a value higher than a certain percentile (such as the 80th percentile) in the score distribution is selected as the effective critical value.

[0109] Critical value of fault deterioration rate: Perform a similar statistical analysis on the fault deterioration rate and select a value below a certain percentile (such as the 20th percentile) as the effective critical value to reflect a lower deterioration rate.

[0110] The following process is used to determine the minimum available network for comprehensive diagnosis through effective diagnostic records: the standard deviation of the network quality score of each effective diagnostic record is calculated, where the standard deviation reflects the overall fluctuation of the network quality score, and then compared with the tolerance threshold of the preset standard deviation. If the standard deviation of the network quality score is less than or equal to the tolerance threshold, it means that the network quality score has a small range of variation and is relatively stable, which represents the minimum operating requirements of the system when the network quality score fluctuates slightly. The network state corresponding to the minimum network quality score value is selected as the minimum available network for comprehensive diagnosis, which represents the minimum operating requirements of the system when the network quality score fluctuates slightly. Otherwise, the median score of the network quality score of each effective diagnostic record is calculated, and the network state corresponding to the median network quality score value is selected as the minimum available network for comprehensive diagnosis. This is because the median reflects the intermediate level of the network quality score and can provide a more robust reference value within a larger fluctuation range.

[0111] It should be noted that the above tolerance thresholds need to be set in combination with actual needs. If the tolerance threshold is too large, it may lead to a standard that is too loose; if it is too small, it may be too strict. Through multiple tests and verifications, a reasonable value that can reflect system performance and adapt to actual needs can be found.

[0112] The diagnostic rule optimization module is used to generate a simplified diagnostic rule by comparing the diagnostic steps of valid and invalid diagnostic records, and divide the network state into a strong state and a weak state according to the minimum available network, so as to establish a selection strategy for fault diagnostic rules.

[0113] Preferably, the above scheme generates a simplified version of the diagnostic rules as follows: extracting steps that fail to locate the cause of the fault from valid diagnostic records and invalid diagnostic records, marking them as non-locked steps, and extracting steps that can locate the cause of the fault and marking them as locked steps.

[0114] It is important to understand that the classification of diagnostic steps into "non-locking steps" (failed to locate the cause of the fault) and "locking steps" (able to locate the cause of the fault) is a clear classification method based on functional contribution. By distinguishing these two types of steps, it is clear which steps directly contribute to fault location and which steps may only be auxiliary or redundant operations.

[0115] The extracted non-locked steps and locked steps are combined to identify the execution order before and after, and those non-locked steps that have a front-to-back execution dependency relationship with the locked steps are screened out. These non-locked steps with dependencies are excluded, and the remaining non-locked steps that do not depend on the locked steps are defined as independent non-locked steps.

[0116] It is further understood that by analyzing the pre- and post-execution dependencies between non-locked steps and locked steps, non-locked steps with practical significance (i.e., those operations that provide necessary prerequisites for locked steps) are screened out. This method ensures that the existence of each step in the diagnostic process is reasonable and avoids the risk of mistakenly deleting key pre- or post-operations that may be caused by simply removing all non-locked steps.

[0117] The proportion of each independent non-locking step that appears simultaneously in the valid diagnosis records and the invalid diagnosis records is counted, and then the independent non-locking step with the largest proportion is taken as the non-critical step.

[0118] It is important to understand that counting the percentage of independent non-locked steps that appear simultaneously in valid and invalid diagnostic records can quantify the commonality and relevance of these steps in different diagnostic results. Selecting the independent non-locked steps with the largest percentage of simultaneous appearance as non-critical steps is a data-driven method that can objectively evaluate the importance of steps.

[0119] The non-critical steps identified above are removed from the comprehensive diagnostic rules, and the remaining diagnostic steps are sequentially generated into a simplified diagnostic rule.

[0120] Finally, it is important to understand that non-critical steps are removed from the comprehensive diagnostic rules, and the remaining steps are usually those operations that directly contribute to fault location, that is, the core diagnostic steps. This method can effectively reduce redundant operations and improve diagnostic efficiency. The generated simplified diagnostic rules retain the core diagnostic logic and streamline unnecessary steps. Although the diagnostic rules are simplified, the combined identification and synchronous proportion statistics of the front and back executions ensure that the simplified rules can still cover the key diagnostic logic, thereby maintaining the accuracy of the diagnosis and being suitable for actual application scenarios.

[0121] The above scheme is further preferably divided into a strong state and a weak state according to the minimum available network. See the following process: taking the comprehensive diagnosis of the minimum available network state as a benchmark, a network state with a network quality lower than the benchmark is defined as a weak state, and a network state with a network quality reaching or higher than the benchmark is defined as a strong state.

[0122] In a further preferred embodiment of the above scheme, a selection strategy for establishing fault diagnosis rules is performed as follows: comprehensive diagnosis rules are selected when the network state is strong, and simplified diagnosis rules are selected when the network state is weak.

[0123] It should be noted that although the simplified diagnostic rules retain the core diagnostic logic by streamlining unnecessary steps, this does not mean that the comprehensive diagnostic rules are invalid, nor does it mean that the simplified rules should be used preferentially in good network conditions. As a standardized process developed for equipment fault characterization research, the comprehensive diagnostic rules are highly comprehensive and systematic, and can maximize the accurate location of the cause of the fault. In contrast, the simplified diagnostic rules are an adaptive solution designed specifically for poor network conditions. Their main goal is to quickly locate the cause of the fault when resources are limited or the environment is unfavorable. However, due to the relatively limited diagnostic processes contained in the rules, some complex faults may not be effectively identified, resulting in a certain risk of missed diagnosis. Therefore, in practical applications, appropriate diagnostic rules should be selected according to the network status and specific needs to achieve a balance between efficiency and accuracy.

[0124] The fault diagnosis implementation module is used to synchronously monitor the fault and network status during the operation of the equipment, and when a fault is identified, a fault diagnosis is performed based on the network status using a selection strategy.

[0125] As a specific description of the above operations, during the operation of the equipment, it is necessary to monitor the fault situation and network status simultaneously. Once a fault is detected, the specific manifestation of the fault should be identified first to clarify what type of fault it belongs to. On this basis, combined with the current network status, fault diagnosis is implemented according to the preset selection strategy.

[0126] After selecting a diagnostic strategy based on the current network status and completing fault diagnosis, the diagnostic results can be fed back to the front end, including the evaluation of the diagnostic effect and fault deterioration rate. If a simplified diagnostic rule is currently used, it is necessary to further judge the effectiveness of the diagnosis based on the diagnostic effect and fault deterioration rate. If the diagnosis is deemed effective, it means that the simplified diagnostic rule can be applied under poor network conditions; conversely, if the diagnosis is invalid, it means that the rule is still insufficient in dealing with scenarios with poor network conditions and needs to be further optimized to improve its adaptability and reliability. This evaluation mechanism helps to continuously improve diagnostic rules to ensure efficient and accurate fault diagnosis under different network conditions.

[0127] Finally, it should be made clear that if the network status is poor when remotely implementing equipment fault diagnosis and control, the network status should be optimized and adjusted first, and comprehensive diagnostic rules should be enabled to ensure the completeness and accuracy of the diagnosis. Only when the network status cannot be adjusted should simplified diagnostic rules be enabled as an alternative. This will cover as many possible fault possibilities as possible and improve the accuracy and comprehensiveness of fault location.

[0128] In the implementation of the present invention, since a variety of data are used, a relational database and a memory database are used to facilitate the call of data, wherein the relational database is used to store historical fault diagnosis records and diagnosis rules, and the memory database is used to store the operation data generated during the operation of the equipment. The specific data call relationship is as follows: Figure 3 shown.

[0129] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they shall all fall within the protection scope of the present invention.

Claims

1. A remote control system for electric power equipment based on the Internet of Things, characterized in that: include: Historical fault analysis module: retrieves similar fault diagnosis records of equipment using comprehensive diagnostic rules to perform diagnostic effect scoring, network quality scoring and fault deterioration rate assessment; Comprehensive diagnosis evaluation module: Build a dynamic network comprehensive diagnosis effectiveness evaluation algorithm, and evaluate the effectiveness of comprehensive diagnosis rules by importing evaluation data of similar fault diagnosis records; Minimum available network determination module: when the effectiveness of comprehensive diagnosis rules does not meet the standard, the same type of fault diagnosis records are divided into valid and invalid diagnosis records, and then the minimum available network for comprehensive diagnosis is determined through valid diagnosis records; Diagnostic rule optimization module: Generates simplified diagnostic rules by comparing the diagnostic steps of valid and invalid diagnostic records, and divides the network status into strong and weak states according to the minimum available network, thereby establishing a selection strategy for fault diagnostic rules. When the network status is strong, comprehensive diagnostic rules are selected, and when the network status is weak, simplified diagnostic rules are selected. Fault diagnosis implementation module: synchronously monitors faults and network status during equipment operation, and uses selected strategies to perform fault diagnosis based on network status when a fault is identified; Determining the minimum available network for comprehensive diagnosis through valid diagnostic records includes: calculating the standard deviation of the network quality score of each valid diagnostic record, and comparing it with the tolerance threshold of the preset standard deviation. If the standard deviation of the network quality score is less than or equal to the tolerance threshold, the minimum network quality score is selected as the minimum available network for comprehensive diagnosis; otherwise, the median score of the network quality score of each valid diagnostic record is calculated, and the median network quality score value is selected as the minimum available network for comprehensive diagnosis; Generating a simplified version of the diagnostic rule includes: extracting steps that fail to locate the cause of the fault from valid diagnostic records and invalid diagnostic records, marking them as non-locked steps, extracting steps that can locate the cause of the fault and marking them as locked steps, performing combined identification of the extracted non-locked steps and locked steps in the order of execution, screening out non-locked steps that have a before-and-after execution dependency relationship with the locked steps, excluding non-locked steps with dependencies, and defining the remaining non-locked steps that do not depend on the locked steps as independent non-locked steps; counting the proportion of each independent non-locked step that appears simultaneously in the valid diagnostic record and the invalid diagnostic record, taking the independent non-locked step with the largest proportion as a non-critical step, eliminating the above-determined non-critical steps from the comprehensive diagnostic rule, and sequentially generating a simplified version of the diagnostic rule for the retained diagnostic steps.

2. The remote control system for electric power equipment based on the Internet of Things as claimed in claim 1, characterized in that: The diagnostic effect score is based on the following evaluation process: Define the diagnostic effect parameters as diagnostic efficiency parameters and diagnostic error parameters, and assign weights to each type of parameter; Extract the data of each diagnostic effect parameter from similar fault diagnosis records and normalize each data; Incorporate each diagnostic effect parameter data after normalization in the fault diagnosis record into the weighted comprehensive scoring model Calculating the diagnostic performance score , where , They represent the normalized diagnostic efficiency parameter and diagnostic error parameter respectively. , They represent the weight factors assigned to the diagnostic efficiency parameter and the diagnostic error parameter respectively.

3. The remote control system for electric power equipment based on the Internet of Things as claimed in claim 1, characterized in that: The network quality score adopts the following evaluation process: Determine a network status indicator, and establish a grade scoring set for each network status indicator, each grade scoring set includes multiple scoring value intervals, and each interval corresponds to a scoring value; Extract the data of each network status indicator from similar fault diagnosis records, and match them with the corresponding grade score set to obtain the score value of each network status indicator; The score value of each network status indicator in the fault diagnosis record is calculated comprehensively to obtain the network quality score of the fault diagnosis record.

4. The remote control system for electric power equipment based on the Internet of Things as claimed in claim 1, characterized in that: The failure deterioration rate refers to the following evaluation process: Extracting a fault representation time evolution sequence corresponding to the fault diagnosis process from similar fault diagnosis records, the sequence including multiple time points and their corresponding fault representations; Extract the fault characterization data of the first time and the last time and the fault diagnosis duration from the time evolution sequence of the fault characterization; Substitute the fault characterization data of the first and last digits and the fault diagnosis time into the fault deterioration rate evaluation formula Get the fault deterioration rate , where , They represent the fault characterization data of the first time and the last time in the fault characterization time evolution sequence respectively. Indicates the fault diagnosis time. Indicates the correction factor.

5. The remote control system for electric power equipment based on the Internet of Things as claimed in claim 1, characterized in that: The dynamic network comprehensive diagnostic utility evaluation algorithm is constructed as follows: The network quality scores, diagnostic effect scores and fault deterioration rates of the same fault diagnosis records are used to generate a network quality change curve, a diagnostic effect change curve and a fault deterioration change curve in a coordinate system with the fault diagnosis records constructed in chronological order as the horizontal axis and the network quality score, diagnostic effect score and fault deterioration rate as the vertical axis; Capture inflection points on the network quality change curve, and further select inflection points that only indicate network quality degradation from all captured inflection points, and simultaneously select the corresponding time points and the fault diagnosis records at the time points; The fault diagnosis record corresponding to the selected inflection point is mapped onto the diagnostic effect change curve to obtain the tangent slope sign of the mapping point as the diagnostic effect change direction corresponding to the network quality degradation point; The fault diagnosis record corresponding to the selected inflection point is mapped onto the fault deterioration change curve to obtain the tangent slope sign of the mapping point as the fault deterioration change direction corresponding to the network quality degradation point; Count the number of network quality degradation points selected on the network quality change curve, and compare the change direction of the diagnostic effect and the change direction of the fault deterioration corresponding to each network quality degradation point. The proportion of the decrease in fault diagnosis effect and the increase in fault deterioration when the network quality deteriorates are recorded as p1 and p2 respectively; According to the preset thresholds p1 and p2, where p1 and p2 represent the setting thresholds of the fault diagnosis effect decline and fault deterioration increase ratio respectively, if and It is judged that the effectiveness of the comprehensive diagnosis rule meets the standard, otherwise it does not meet the standard.

6. The remote control system for electric power equipment based on the Internet of Things as claimed in claim 1, characterized in that: The classification of the same type of fault diagnosis records into valid and invalid diagnosis records refers to the following process: The diagnostic effect score values ​​and fault deterioration rates of similar fault diagnostic records are compared with the set effective critical values, thereby screening out the fault diagnostic records whose diagnostic effect score values ​​and fault deterioration rates meet the effective critical values ​​as valid diagnostic records, and the other fault diagnostic records as invalid diagnostic records.

7. The remote control system for electric power equipment based on the Internet of Things as claimed in claim 1, characterized in that: The network state is divided into a strong state and a weak state according to the minimum available network, see the following process: Taking the minimum available network for comprehensive diagnosis as a benchmark, a network state whose network quality is lower than the benchmark is defined as a weak state, and a network state whose network quality reaches or exceeds the benchmark is defined as a strong state.

Images