A high-voltage switch mechanical state diagnosis evaluation control method and system

Abstract

The application provides a high-voltage switch mechanical state diagnosis evaluation control method and system, relates to the technical field of power system maintenance and management, and comprises the following steps: collecting real-time data of a high-voltage switch, collecting real-time data, historical residual life, historical data and corresponding fault labels of each element in the high-voltage switch; constructing a residual life prediction model to obtain the residual life of the element; constructing a failure rate prediction model to obtain the failure rate of the element; judging whether the failure rate of the element exceeds a preset failure threshold; if yes, performing high-voltage switch mechanical state failure alarm prompting; if no, calculating the failure rates of all commands in the command tree based on the failure rate of the element, judging whether the failure rates of the commands are all lower than a preset alarm threshold; if yes, displaying the result that the mechanical state of the high-voltage switch is qualified; and if no, performing high-voltage switch mechanical state abnormality alarm prompting. The application is helpful for capturing instability superposition in element execution commands and improves the accuracy of high-voltage switch mechanical state diagnosis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system maintenance and management, in particular to a high-voltage switch mechanical state diagnosis and evaluation control method and system. BACKGROUND

[0002] High-voltage switches are important protection components in power systems and are widely used in substations, power transmission and distribution systems, and large industrial facilities. However, due to the complex working environment of high-voltage switches, which involves the coordinated work of a large number of mechanical components, especially during frequent closing and opening operations, these mechanical components are prone to wear, aging, and other mechanical failures. Therefore, it is of great significance to effectively diagnose and evaluate the mechanical state of high-voltage switches.

[0003] In related technologies, historical state data of each mechanical component in the high-voltage switch and corresponding high-voltage switch failure labels are used to train a neural network model, and then the current failure condition of the high-voltage switch is predicted based on the current state data of each mechanical component and the trained model.

[0004] However, the current high-voltage switch functions are becoming more and more complex, and such high-voltage switches control multiple related circuits at the same time. When using the model to predict, the model can only learn the influence of component state data, single component or different component combinations on high-voltage switch failure under different working conditions, but in actual application, when the high-voltage switch executes a complete command, each component may be used more than once. Therefore, when the component is used multiple times, the instability of the component itself will have a cumulative effect on the failure of the high-voltage switch.

[0005] However, the existing model cannot accurately capture the instability of the component during the execution of the command, ignores the state drift caused by the accumulation of operation times, and results in the failure prediction result of the high-voltage switch being unable to reflect the change in failure risk caused by the accumulation of execution times, causing the failure prediction result to deviate from the actual situation. SUMMARY

[0006] In order to improve the accuracy of high-voltage switch mechanical state diagnosis, the present application provides a high-voltage switch mechanical state diagnosis and evaluation control method and system.

[0007] In a first aspect, the present application provides a high-voltage switch mechanical state diagnosis and evaluation control method, which adopts the following technical solution:

[0008] A high-voltage switch mechanical state diagnosis and evaluation control method, comprising the following steps:

[0009] Data acquisition: Collect real-time and historical data of the high-voltage switch, and collect real-time data, historical remaining life, historical data and corresponding fault tags of each component in the high-voltage switch;

[0010] Predicting the remaining lifetime of components: Construct a remaining lifetime prediction model by inputting real-time and historical data of the components into the remaining lifetime prediction model to obtain the remaining lifetime of the components;

[0011] Obtaining component failure rate: Construct a failure rate prediction model. Train the model based on historical data of high-voltage switches, historical remaining life of components, and corresponding fault labels. Input the remaining life of components, real-time data of high-voltage switches and components into the trained model to obtain the component failure rate.

[0012] Component assessment: Determine whether the failure rate of each component exceeds the preset failure threshold.

[0013] If so, a high-voltage switch mechanical status fault alarm will be issued;

[0014] If not, proceed with the status assessment step;

[0015] Status assessment: Based on the failure rate of each component, calculate the failure rate of all commands in the preset command tree, and determine whether the failure rate of each command is lower than the preset alarm threshold.

[0016] If so, it displays the result that the current mechanical condition of the high-voltage switch is qualified;

[0017] If not, then issue an alarm for abnormal mechanical status of the high-voltage switch.

[0018] The failure rate of the command is calculated using the following model:

[0019] ;

[0020] in, This indicates the command failure rate. Indicates the first The number of times each component is executed in the current command. , Indicates the first Failure rate of individual components.

[0021] By adopting the technical scheme, the residual life prediction model is constructed to obtain the residual life of the element, the instability of the current state of the element is comprehensively predicted based on the aging degree of the element and the change of the operating state under different working conditions, which helps to more accurately predict the risk of element execution error, and further improves the prediction accuracy of the overall operating state of the high-voltage switch. At the same time, unlike the traditional high-voltage switch prediction method, the complex coupling effect between the execution commands of the elements is included in the evaluation range, based on the characteristics of the instability superposition in the execution command process of the elements, through the dynamic calculation of the command combination and the element failure rate, considering the cumulative effect of the command execution times on the instability of the elements, the failure probability of the entire high-voltage switch system is calculated, which helps to provide more actual quantitative basis for the stability evaluation of the high-voltage switch system, and provides effective quantitative indexes for accurately evaluating the mechanical state of the entire high-voltage switch system, greatly improving the prediction ability of the potential fault hidden danger of the high-voltage switch existing in the command execution, and helping the long-term stable and reliable operation of the high-voltage switch system.

[0022] Optionally, after the state evaluation step, it further includes:

[0023] Obtaining the risk degree of the element: extracting the command with a failure rate not lower than the preset alarm threshold as a risk command, obtaining the overall failure rate caused by each element in the current risk command, denoted as total failure rate;

[0024] Element maintenance verification: the elements are gradually simulated and replaced in the order from large to small according to the total failure rate, the failure rate of the risk command after simulation and replacement is calculated, and the failure rate of the risk command after simulation and replacement is lower than the preset alarm threshold:

[0025] Formulating a preset maintenance plan: performing the element maintenance verification step on all risk commands until the failure rates of all risk commands are lower than the preset alarm threshold, taking all simulation replacement results as the preset maintenance plan, and outputting the preset maintenance plan.

[0026] By adopting the technical scheme, the command with a failure rate not lower than the preset alarm threshold is extracted, and the total failure rate caused by each element in the command is identified, and the simulation replacement is performed in the order of the total failure rate, so as to verify the influence of the replaced element on the failure rate of the command, which helps to simulate the implementation effect of different maintenance schemes without interfering with the actual production. At the same time, this dynamic optimization method helps to evaluate the elements that are most prone to failure when frequently used, so that the preset maintenance plan is more personalized and accurate, which realizes the technical effect of relieving the instability superposition of the elements after multiple uses, and based on the simulation replacement of the elements and the failure rate evaluation after the element replacement, not only improves the efficiency of the maintenance process, but also improves the scientificity of the maintenance plan, saves the operation and maintenance cost, and further improves the failure prevention ability of the high-voltage switch.

[0027] Optionally, in the risk degree obtaining step of the obtaining element, the maximum of the overall failure rate of the same element in all risk commands is taken as the total failure rate.

[0028] By taking the maximum of the overall failure rate of the same element in all risk commands as the total failure rate of the element, it is helpful to identify and prioritize the elements that exhibit high risk under multiple conditions when the elements are repaired, so as to more effectively allocate resources for failure prevention and maintenance.

[0029] Optionally, before the preset maintenance plan making step, it further comprises:

[0030] Risk command verification: judging whether the risk command exists a simulated replacement element:

[0031] If yes, the element maintenance verification step is performed on the risk command whose failure rate after simulated replacement is still not lower than the preset alarm threshold;

[0032] If no, the element maintenance verification step is performed on all risk commands.

[0033] By adopting the above technical solution, it is helpful to reduce unnecessary maintenance intervention, avoid repeated calculation on the risk command that has been alleviated or solved, thereby optimizing the simulation process, improving the accuracy of the maintenance strategy, saving time and cost.

[0034] Optionally, it further comprises a backup element database, and the backup element database comprises the element type and the remaining life of the backup element.

[0035] In the element maintenance verification step, the backup element in the backup element database is extracted for simulated replacement.

[0036] By adopting the above technical solution, the backup element database is set, which is helpful to quickly obtain the type and remaining life information of the backup element, so that in the element failure or maintenance process, a suitable backup element can be quickly found for replacement, which can reduce the maintenance time and avoid downtime or system failure due to lack of backup elements.

[0037] Optionally, in the state evaluation step, it further comprises: if there is a multiple set-subset relationship between the commands in the preset command tree, only the failure rate of the non-subset command is calculated.

[0038] By adopting the above technical solution, the multiple set-subset relationship judgment mechanism is introduced, which is helpful to exclude redundant commands, optimize the calculation process, improve the failure evaluation efficiency of the high-voltage switch system, and further improve the rationality of the maintenance resource allocation.

[0039] Optionally, before the element judging step, further comprising: if the trigger condition, execution object and execution effect of the plurality of elements are completely same, merging the plurality of elements into one virtual element.

[0040] The calculation model of the failure rate of the merged virtual element is:

[0041]

[0042] wherein, represents the failure rate of the virtual element, represents the number of elements with the same trigger condition, execution object and execution effect as the current trigger condition, execution object and execution effect, represents the failure rate of the i-th element in all elements with the same trigger condition, execution object and execution effect as the current trigger condition, execution object and execution effect.

[0043] By adopting the above technical solution, the elements with the same trigger condition, execution effect and execution object are merged into one virtual element, which helps to summarize the reliability of the redundantly designed elements, comprehensively consider the overall failure rate of the redundantly designed elements, reduce the calculation burden, reduce the repeated evaluation work, and help to improve the reliability of the overall failure evaluation of the high-voltage switch system with complex structure and redundant elements. Moreover, it helps to simplify the element maintenance strategy and reduce resource waste. In addition, considering the redundancy design of the high-voltage switch system by merging the elements with the same execution effect and execution object into one virtual element to calculate the failure rate of the virtual element helps to further evaluate the effectiveness of the redundancy design of the high-voltage switch system, so as to achieve the technical effect of improving the stability and safety of the high-voltage switch system.

[0044] Optionally, in the state evaluation step, the preset alarm threshold of each command is determined according to the usage frequency of all commands in the command tree within a preset time.

[0045] By adopting the above technical solution, the alarm threshold is set according to the usage frequency of the command, the accelerating effect of the frequently used command on the instability of the element is considered, and the risk level of the frequently used command is more accurately reflected, so as to provide more personalized and accurate failure evaluation results for the high-voltage switch system. Moreover, this dynamic threshold setting can focus on the usage of high-frequency commands, which helps to more accurately evaluate the potential failure level of the high-voltage switch system and improve the failure evaluation response efficiency.

[0046] Optionally, in the state evaluation step, the preset alarm threshold of each command is determined according to the importance of the execution result of all commands in the command tree.

[0047] ​​By adopting the above technical solution and setting alarm thresholds based on the importance of command execution results, it helps reduce overreactions to unimportant tasks, thereby reducing false alarms caused by minor anomalies in unimportant commands. Simultaneously, focusing on the risk status of important commands allows for timely detection of potential problems, reducing missed alarms. Furthermore, this dynamic threshold mechanism helps to more accurately reflect the risk level of important commands, thus providing more personalized and accurate fault assessment results for high-voltage switchgear systems.

[0048] Secondly, this application provides a high-voltage switch mechanical condition diagnosis and evaluation control system, which adopts the following technical solution:

[0049] A high-voltage switch mechanical condition diagnosis and assessment control system, including

[0050] Data acquisition module: used to collect real-time and historical data of high-voltage switch, collect real-time data, historical remaining life, historical data and corresponding fault tags of each component in high-voltage switch, and send the collection results to the failure rate prediction module;

[0051] Failure rate prediction module: It is used to predict the failure rate of components by combining the remaining life prediction model and the failure rate prediction model, combined with the real-time data and historical data of components, the historical data of high voltage switches, the historical remaining life and historical data of components, and the corresponding fault tags, and then send the prediction results to the evaluation module.

[0052] Evaluation module: Used to evaluate the fault condition of components, determine the current mechanical state of the high-voltage switch, and transmit the results to the alarm module;

[0053] Alarm module: Used to issue alarm prompts for abnormal mechanical status of high-voltage switches or alarm prompts for mechanical faults in high-voltage switches.

[0054] In summary, this application includes at least one of the following beneficial technical effects:

[0055] 1. Constructing a residual life prediction model to obtain the residual life of the element, based on the aging degree of the element and the change of the operating state under different working conditions, comprehensively predicting the instability of the current state of the element, which helps to more accurately predict the risk of element execution error, and further improves the prediction accuracy of the overall operating state of the high-voltage switch. At the same time, unlike traditional high-voltage switch prediction methods, this scheme takes into account the complex coupling effect between elements in the execution command, based on the characteristics of the instability superposition in the element execution command process, through the dynamic calculation of the command combination and the element failure rate, considering the cumulative effect of the command execution times on the instability of the element to calculate the failure probability of the entire high-voltage switch system, which helps to provide more realistic quantitative basis for the stability evaluation of the high-voltage switch system, and provides an effective quantitative index for accurately evaluating the mechanical state of the entire high-voltage switch system, greatly improving the prediction ability of the potential fault hidden danger of the high-voltage switch existing in the command execution, which helps the long-term stable and reliable operation of the high-voltage switch system.

[0056] 2. Extracting the command with a failure rate not lower than the preset alarm threshold, and identifying the total failure rate of each element caused by the command, and simulating replacement according to the order of the total failure rate to verify the influence of the replacement element on the command failure rate, which helps to simulate the implementation effect of different maintenance schemes without interfering with the actual production. At the same time, this dynamic optimization method helps to evaluate the elements that are most prone to failure when frequently used, making the preset maintenance plan more personalized and accurate, and achieving the technical effect of relieving the instability superposition of the element after multiple uses. Moreover, based on the simulation replacement of the element and the failure rate evaluation after the replacement of the element, not only the efficiency of the maintenance process is improved, but also the scientificity of the maintenance plan is improved, the operation and maintenance cost is saved, and the fault prevention ability of the high-voltage switch is further improved.

[0057] 3. Merging elements with the same trigger condition, execution effect and execution object into a virtual element, which helps to summarize the reliability of the redundantly designed elements, and comprehensively considers the overall failure rate of the redundantly designed elements, reducing the computational burden and reducing repeated evaluation work, which helps to improve the reliability of the overall failure evaluation of the high-voltage switch system with redundant elements and complex structure. Moreover, it is also helpful to simplify the element maintenance strategy and reduce resource waste. In addition, considering the redundancy design of the high-voltage switch system by merging elements with the same execution effect and execution object into a virtual element to calculate the failure rate of the virtual element, which helps to further evaluate the effectiveness of the redundancy design of the high-voltage switch system, to achieve the technical effect of improving the stability and safety of the high-voltage switch system.

[0058] 4. Set the alarm threshold according to the frequency of use of the command, considering the accelerated effect of frequently used commands on the instability of the element, which helps to more accurately reflect the risk level of frequently used commands, thereby providing more personalized and accurate failure assessment results for the high-voltage switch system. Moreover, this dynamic threshold setting can focus on the use of high-frequency commands, helping to more accurately assess the potential failure level of the high-voltage switch system and improve the response efficiency of failure assessment. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 is a flowchart of embodiment 1 of the present application;

[0060] Figure 2 is a flowchart of embodiment 2 of the present application;

[0061] Figure 3 is a flowchart of embodiment 4 of the present application;

[0062] Figure 4 is a schematic diagram of module connection of embodiment 8 of the present application. DETAILED DESCRIPTION

[0063] The following Figures 1 to 4 The present application is further described in detail.

[0064] Embodiment 1: Referring to Figure 1 , the present embodiment discloses a high-voltage switch mechanical state diagnosis and evaluation control method, as shown in Figure 1 , the system comprises: collecting real-time data and historical data of the high-voltage switch, collecting real-time data, historical residual life, historical data and corresponding failure labels of each element in the high-voltage switch; constructing a residual life prediction model, inputting the real-time data and historical data of the element into the residual life prediction model to obtain the residual life of the element; constructing a failure rate prediction model and training the model, inputting the residual life of the element, the real-time data of the high-voltage switch and the element into the trained model to obtain the failure rate of the element; judging whether the failure rate of each element exceeds the preset failure threshold: if yes, performing high-voltage switch mechanical state failure alarm prompt; if no, based on the failure rate of each element, calculating the failure rate of all commands in the preset command tree, judging whether the failure rate of the command is lower than the preset alarm threshold: if yes, displaying the result that the current mechanical state of the high-voltage switch is qualified; if no, performing high-voltage switch mechanical state abnormal alarm prompt. The present embodiment comprises the following steps:

[0065] S1 data collection: collecting real-time data and historical data of the high-voltage switch, collecting real-time data, historical residual life, historical data and corresponding failure labels of each element in the high-voltage switch.

[0066] The real-time data of the high-voltage switch includes ambient temperature, ambient humidity, input current and input voltage. The historical data of the high-voltage switch includes historical ambient temperature, historical ambient humidity, historical input current and historical input voltage.

[0067] The elements in the high-voltage switch include switch elements, execution elements, protection elements, detection elements and operation elements, etc. The switch elements include circuit breakers, disconnectors and grounding switches, etc. The execution elements include load switches and fuses, etc. The protection elements include voltage transformers and current transformers, etc. The detection elements include relays and discharge detectors, etc. The operation elements include electric actuators and spring energy storage devices, etc.

[0068] The real-time data of the elements in the high-voltage switch includes real-time current, real-time voltage and real-time temperature. The historical data of the elements in the high-voltage switch includes historical current, historical voltage and historical temperature. The fault label of the historical data is the fault that occurs before the element is replaced with a new element.

[0069] S2 predicts the remaining life of the element: a remaining life prediction model is constructed, the real-time data and the historical data of the element are input into the remaining life prediction model, and the remaining life of the element is obtained. A long short-term memory network model, a convolutional neural network model or a time convolution network model can be used as a basic model of the remaining life prediction model.

[0070] In this embodiment, the remaining life prediction model is constructed based on a long short-term memory network model. The training process of the remaining life prediction model is as follows:

[0071] The temperature data, current data and voltage data of the entire life cycle of various elements are obtained to construct a sample training set, the constructed sample training set is input into the remaining life prediction model for model training, and the trained remaining life prediction model is obtained.

[0072] In this embodiment, the real-time data and the historical data of each element are respectively input into the trained remaining life prediction model, and the remaining life of each element is obtained.

[0073] S3 obtains the element failure rate: a failure rate prediction model is constructed, the model is trained based on the historical data of the high-voltage switch, the historical remaining life and the historical data of the element and the corresponding fault label, the remaining life of the element, the real-time data of the high-voltage switch and the element are input into the trained model, and the element failure rate is obtained.

[0074] The Transformer model was chosen as the base model for the failure rate prediction model. Therefore, a portion of the Transformer encoder layers were used as shared layers, while the remaining Transformer encoder layers and the fully connected layers performing specific tasks within the Transformer model were designated as task-specific layers. Task-specific branches were created based on the components within the high-voltage switchgear, each branch containing a failure rate prediction task for that component. The number of task-specific branches created corresponded to the number of different types of components within the high-voltage switchgear. A failure rate prediction sample training set was constructed based on historical data from the high-voltage switchgear, the historical remaining lifespan and historical data of the components, and the corresponding fault labels. The failure rate prediction model was then trained using this training set, resulting in the trained failure rate prediction model.

[0075] S4 Component Judgment: Determines whether the failure rate of each component exceeds the corresponding preset failure threshold.

[0076] If so, output the component and provide a high-voltage switch mechanical status fault alarm.

[0077] If not, proceed with the S5 status assessment step.

[0078] S5 Status Assessment: Determines whether the failure rate of all calculated commands is lower than the preset alarm threshold.

[0079] If so, it will display the result that the current mechanical condition of the high-voltage switch is qualified.

[0080] If not, an alarm will be triggered indicating an abnormal mechanical condition of the high-voltage switch.

[0081] The failure rate of the command is calculated using the following model:

[0082] .

[0083] in, This indicates the command failure rate. Indicates the first The number of times each component is executed in the current command. , Indicates the first Failure rate of individual components.

[0084] In this embodiment, a preset fault threshold range for each component is set according to industry standards and the recommendations of experts in the field, and the maximum value in the preset fault threshold range for each component is selected as the preset fault threshold for each component.

[0085] The construction process of the preset command tree is: constructing multiple command trees based on different scene requirements, taking an externally triggered operation or a received command started under a certain condition as a root node, taking an element in a high-voltage switch that needs to be triggered under a certain condition as an intermediate node, and taking an operation triggered under the condition of all intermediate nodes, that is, a target result, as a leaf node, to construct the command tree. The condition met and the elements involved between the root node and each leaf node are taken as a command, and one command includes an externally triggered operation or a received command started under a certain condition, a triggering condition of an element, a triggering operation of the element, and a target result.

[0086] For example, a detection element 1, a detection element 2, an execution element 1, and an execution element 2 are used in one command tree.

[0087] The detection element 1 obtains a detection result, and if the result is greater than a certain threshold, the execution element 1 executes a command, otherwise the execution element 2 executes a command, and then the detection element 2 detects another detection result. After the detection element 2 obtains another detection result, if the result is greater than a certain threshold, the execution element 2 executes another command, otherwise the execution element 1 executes a command.

[0088] Therefore, the command tree has the following three commands:

[0089] Command 1: detection element 1 — execution element 1;

[0090] Command 2: detection element 1 — execution element 2 — detection element 2 — execution element 2;

[0091] Command 3: detection element 1 — execution element 2 — detection element 2 — execution element 1.

[0092] If the failure rate of the detection element 1 is 0.1, the failure rate of the detection element 2 is 0.15, the failure rate of the execution element 1 is 0.2, and the failure rate of the execution element 2 is 0.25, according to the state evaluation method of the embodiment of the application,

[0093] the failure rate of the command 1 is:

[0094] P1=1-(1-0.1)×(1-0.2)=0.28;

[0095] the failure rate of the command 2 is:

[0096] P2=1-(1-0.1)×(1-0.25)×(1-0.15)×(1-0.25)=0.5696875;

[0097] the failure rate of the command 3 is:

[0098] P3 = 1 - (1 - 0.1) x (1 - 0.25) x (1 - 0.15) x (1 - 0.2) = 0.541.

[0099] In this embodiment, the execution command time complex coupling effect between elements is taken into account in the evaluation range, based on the characteristics of the instability superposition in the element execution command process, through the dynamic calculation of the command combination and the element failure rate, considering the cumulative effect of the command execution times on the instability of the elements, the failure probability of the entire high-voltage switch system is calculated, which helps to provide more realistic quantitative basis for the stability evaluation of the high-voltage switch system, provides effective quantitative index for accurately evaluating the mechanical state of the entire high-voltage switch system, greatly improves the prediction ability of the potential fault hidden danger of the high-voltage switch existing in the command execution, and helps the long-term stability and reliable operation of the high-voltage switch system.

[0100] Embodiment 2: Refer to Figure 2 The embodiment of the present application is basically the same as embodiment 1, the difference lies in that after obtaining the element failure rate in S3, S31 element integration operation can be further included.

[0101] S31 element integration operation: if the trigger conditions, execution objects and execution effects of multiple elements are exactly the same, these elements are merged into a virtual element;

[0102] The calculation model of the failure rate of the virtual element after merging is:

[0103] ;

[0104] Wherein, represents the failure rate of the virtual element, represents the number of elements with the same trigger condition, execution object and execution effect as the current, represents the failure rate of the first element in all elements with the same trigger condition, execution object and execution effect as the current.

[0105] Example: three relays are provided in the high-voltage switch, which are relay 1, relay 2 and relay 3, the failure rate of relay 1 is 0.2, the failure rate of relay 2 is 0.4, and the failure rate of relay 3 is 0.3, if the execution objects of the three relays are all the circuit breaking of control circuit A, and the trigger conditions of the three relays are exactly the same, then relay 1, relay 2 and relay 3 are merged as a virtual relay, then the failure rate of the virtual relay is The virtual relay after merging is taken as a virtual element to execute S4 element judgment.

[0106] When the state is evaluated, the multiple trigger conditions, execution objects and execution effects of the elements with the same elements are integrated into a virtual element, so that the calculation result of the command failure rate is more accurate, the probability of a large number of element replacements caused by false positives is reduced, the evaluation difficulty is reduced, and the evaluation efficiency is improved.

[0107] Embodiment 3: The embodiment of the application is substantially the same as embodiment 1, and the difference lies in that S5 state evaluation: judging whether there is a multiple set subset relationship between the commands in the preset command tree, if yes, only based on the element failure rate, the failure rate of the non-subset command is calculated; if not, based on the element failure rate, the failure rate of all commands in the preset command tree is calculated. Then judge whether the failure rates of all the calculated commands are lower than the preset alarm threshold.

[0108] If yes, the result that the current mechanical state of the high-voltage switch is qualified is displayed.

[0109] If not, the high-voltage switch mechanical state abnormality alarm prompt is performed.

[0110] The calculation model of the failure rate of the command is as follows:

[0111] .

[0112] Wherein, represents the failure rate of the command, represents the execution times of the element in the current command, , , represents the failure rate of the element.

[0113] In this embodiment, the multiple set subset relationship refers to that the execution times of the elements in a command and the elements corresponding to the execution times are contained in another command.

[0114] ​For example, there are three commands, including command 1, command 2 and command 3. Command 1 includes element 1, element 2, element 3, element 4, element 1 and element 5, command 2 includes element 1, element 2, element 2, element 1, element 4, element 1 and element 5, and command 3 includes element 1, element 2, element 3, element 2, element 1, element 4, element 4 and element 5. The execution times of element 1 in command 1 are 2, and the execution times of element 2, element 3, element 4 and element 5 are all 1; the execution times of element 1 in command 2 are 3, the execution times of element 2 are 2, and the execution times of element 3, element 4 and element 5 are all 1; the execution times of element 1 in command 3 are 2, the execution times of element 2 are 2, the execution times of element 4 are 2, and the execution times of element 3 and element 5 are both 1. Since all the elements in command 1 exist in command 2, and the execution times of each element in command 1 are within the execution times of the corresponding elements in command 2, there is a multiple set-subset relationship between command 1 and command 2, and command 1 is a subset of command 2. All the elements in command 3 exist in command 2, but the execution times of element 4 in command 3 are greater than the execution times of element 4 in command 2, so there is no multiple set-subset relationship between command 2 and command 3.

[0115] When the failure rates required for all commands are consistent, the non-subset commands can be screened out by judging whether there is a multiple set-subset relationship between the commands in the preset command tree, which can greatly improve the efficiency of calculating the failure rate.

[0116] Embodiment 4: Reference Figure 3 The difference from embodiment 1 is that:

[0117] After the S5 state evaluation step, it further includes S6 element maintenance, which includes S61 obtaining the risk degree of the element, S62 element maintenance verification and S63 making a preset maintenance plan.

[0118] S61 Obtain the risk degree of the element: extract the command with a failure rate not lower than the preset alarm threshold as a risk command, and obtain the total failure rate caused by each element in the current risk command to the risk command.

[0119] The calculation model of the total failure rate caused by the element to the risk command is as follows:

[0120] .

[0121] Wherein, represents the total failure rate caused by the element in the risk command to the current risk command, represents the execution times of the i th element in the current risk command, represents the execution times of the i th element in the current risk command, represents the execution times of the i th element in the current risk command, Failure rate of an element.

[0122] S62 Risk command verification: select a risk command, and determine whether the risk command has a simulated replacement element: if yes, execute S63 element maintenance verification step on the risk command whose failure rate after the simulated replacement is still not lower than the preset alarm threshold;

[0123] if no, directly execute S63 element maintenance verification step on the risk command.

[0124] S63 Element maintenance verification: sort the elements according to the total failure rates of the elements in the selected risk command from large to small, and perform step-by-step simulated replacement on the elements according to the order, and calculate the failure rate of the corresponding risk command after each simulated replacement, until the failure rate of the risk command after the simulated replacement is lower than the preset alarm threshold.

[0125] That is, in the current risk command, first simulate the failure rate of the current risk command after replacing the element ranked first, and determine whether the failure rate of the current risk command after replacing the element ranked first is lower than the preset alarm threshold: if no, replace the element ranked second, and determine whether the failure rate of the current risk command after replacing the element ranked second is lower than the preset alarm threshold, until the failure rate of the risk command after the simulated replacement is lower than the preset alarm threshold, and stop the simulation of the risk command.

[0126] S64 Formulate a preset maintenance plan: execute the element maintenance verification step on all risk commands until the failure rates of all risk commands are lower than the preset alarm threshold, aggregate all simulated replacement results as a preset maintenance plan (i.e., all simulated replacement elements are the elements to be replaced in the maintenance plan), and output the preset maintenance plan.

[0127] Example 1: Assuming that the failure rates of the following elements are obtained through S3: the failure rate of element A1 is 0.5, the failure rate of element A2 is 0.6, the failure rate of element A3 is 0.3, the failure rate of element A4 is 0.2, and the failure rate of element A5 is 0.4. There are risk command 1 and risk command 2, wherein risk command 1 includes element A1, element A2, element A3, element A1, element A1, and element A2, and risk command 2 includes element A4, element A3, element A2, element A2, element A3, element A5, and element A3.

[0128] Then the failure rate P1 of risk command 1 is:

[0129] P1 = 1 - (1 - 0.5) × (1 - 0.6) × (1 - 0.3) × (1 - 0.5) × (1 - 0.5) × (1 - 0.6) = 0.986;

[0130] The overall failure rate of element A1 for risk command 1 is PA1:

[0131] PA1=1-(1-0.5)×(1-0.5)×(1-0.5)=0.875;

[0132] The overall failure rate of element A2 for risk command 1 is PA2:

[0133] PA2=1-(1-0.6)×(1-0.6)=0.84;

[0134] The overall failure rate of element A3 for risk command 1 is PA3: PA3=1-(1-0.3)=0.3.

[0135] The overall failure rates of element A1, element A2 and element A3 in risk command 1 are 0.875, 0.84 and 0.3 respectively. The above operations are performed on risk command 2, and the overall failure rates of element A2, element A3, element A4 and element A5 in risk command 2 are 0.84, 0.675, 0.2 and 0.4 respectively.

[0136] According to the above calculation process, the ranking of elements in risk command 1 is element A1, element A2 and element A3. Element A1 is replaced, and the failure rate of the current risk command 1 after the replacement of element A1 is calculated. It is determined that the failure rate of the current risk command 1 after the replacement of element A1 is not lower than the preset alarm threshold. Element A2 is replaced, and the failure rate of the current risk command 1 after the replacement of element A2 is calculated on the basis of the replacement of element A1. At this time, the calculated failure rate of risk command 1 is lower than the preset alarm threshold, and element A1 and element A2 are taken as the simulation replacement result. Then, according to the overall failure rates of the elements in risk command 2, the ranking of elements in risk command 2 is element A2, element A3, element A5 and element A4. Since element A2 in risk command 1 has been replaced in the above steps, the failure rate of the current risk command 2 after the replacement of element A2 is calculated. It is determined that the failure rate of risk command 2 after the replacement of element A2 is not lower than the preset alarm threshold. Element A3 is replaced, and the failure rate of the current risk command 2 after the replacement of element A3 is calculated on the basis of the replacement of element A2. At this time, the calculated failure rate of risk command 2 is lower than the preset alarm threshold, and element A2 and element A3 are taken as the simulation replacement result. The simulation replacement results of risk command 1 and risk command 2 are summarized, and element A1, element A2 and element A3 are taken as the preset maintenance plan output.

[0137] Example 5: The difference from example 4 is that:

[0138] S61: Obtain the risk degree of the element: extract the command with the failure rate not lower than the preset alarm threshold as a risk command, obtain the overall failure rate of each element in the current risk command to the risk command, take the maximum value of the overall failure rate of the same element in all risk commands as the total failure rate, and record the risk command corresponding to the maximum value of the overall failure rate of the same element in all risk commands as the total failure rate.

[0139] For example, the failure rate of element A in command 1 is 0.5, the failure rate of element A in command 2 is 0.6, and the failure rate of element A in command 3 is 0.7. The total failure rate of element 1 is 0.7, and the corresponding risk command is command 3.

[0140] S62: Risk command verification: sort the total failure rates from large to small, select a risk command according to the order of the total failure rate, and judge whether there is a simulated replacement element in the risk command: if yes, execute the element maintenance verification step S63 for the risk command whose failure rate is still not lower than the preset alarm threshold after simulated replacement;

[0141] If not, directly execute the element maintenance verification step S63 for the risk command.

[0142] S63: Element maintenance verification: simulate replacement of elements step by step according to the order of the total failure rate of each element, and calculate the failure rate of the corresponding risk command after each simulated replacement.

[0143] S64: Formulate a preset maintenance plan: execute the element maintenance verification step for all risk commands until the failure rates of all risk commands are lower than the preset alarm threshold, aggregate all simulated replacement results as a preset maintenance plan (i.e., all elements that have been simulated and replaced are regarded as elements to be replaced in the maintenance plan), and output the preset maintenance plan.

[0144] That is, first simulate and calculate the failure rate of the current risk command after replacing the first ranked element, judge whether the failure rate of the current risk command after replacing the first ranked element is lower than the preset alarm threshold, then replace the second ranked element, and judge whether the failure rate of the current risk command after replacing the second ranked element is lower than the preset alarm threshold, until the failure rates of all risk commands after simulated replacement are all lower than the preset alarm threshold, and stop the simulation of the risk command.

[0145] The difference between the embodiment and embodiment 4 in implementation is that embodiment 4 first simulates and replaces the risk command with the highest failure rate, and replaces the elements in the risk command in order from large to small according to the failure rate of the elements in the risk command during simulated replacement, until the failure rate of the risk command is lower than the preset alarm threshold, and then simulates and replaces the next risk command, until all risk commands are lower than the preset alarm threshold.

[0146] Embodiment 5 is to integrate the failure rates of elements first, take the maximum value of the overall failure rate of the same element in all risk commands as the total failure rate, and then replace the elements according to the total failure rate from large to small until all risk commands are below the preset alarm threshold. Compared with embodiment 4, embodiment 5 can more accurately capture the faulty elements and reduce unnecessary replacement of elements.

[0147] Embodiment 6: The difference between the embodiment of the application and embodiment 4 is that it further includes a spare element database, which includes the element type and remaining life of the spare element. In the element maintenance verification step, the spare element in the spare element database is extracted for simulated replacement.

[0148] In this embodiment, the spare elements in the spare element database are all unused by default. Since the spare elements in the spare element database are in the unused state, the failure rate of the default spare element is the lowest failure rate in the preset production standard failure rate range. In other embodiments, the spare elements in the spare element database can be used, such as in embodiment 2, since multiple elements constitute a virtual element, when the virtual element fails, all elements in the virtual element will be replaced, but not all elements cannot continue to be used, at this time, these still usable elements can be placed in the spare element library for backup, reducing waste.

[0149] Embodiment 7: The difference between embodiment 7 and embodiment 4 is that:

[0150] In the S5 state evaluation step, the preset alarm threshold of each command is determined according to the usage frequency of all commands in the command tree within the preset time.

[0151] In this embodiment, the preset time is 6 months, and in this embodiment, the preset alarm threshold of the command gradually decreases as the usage frequency of the command increases. The calculation model of the preset alarm threshold is as follows:

[0152] .

[0153] wherein, represents the preset alarm threshold of the command, is the initial alarm threshold, that is, the alarm threshold when the usage frequency of the command is 0, is the usage frequency of the command, is the decay factor. The value of can be adjusted by system simulation, and the false alarm rate and the missed alarm rate under each setting are analyzed, the alarm effect of the system under different decay factors is compared, and finally a most suitable value is selected as the decay factor .

[0154] In other embodiments, the preset alarm threshold of each command can also be determined by weighted summation according to the importance of the execution result of all commands in the command tree and the usage frequency of all commands within the preset time. Combining the importance of the execution result of the command with the usage frequency of the command within the preset time, the relative importance of each command is comprehensively evaluated, which enhances the fault evaluation capability of the system and helps to find potential faults of the high-voltage switch system. The calculation model of the preset alarm threshold is as follows:

[0155] .

[0156] wherein, represents the weight of the usage frequency of the command, represents the weight of the importance of the execution result of the command. and The settings of and can be set according to the suggestions of experts in the field, and herein = 0.6, = 0.4. represents the importance of the execution result of the command. The commands are classified based on the influence range after the execution of each command, and the corresponding command influence value is set according to the classification result, and the command influence value corresponding to the execution result of each command is taken as the importance of the execution result of each command. The command influence value corresponding to each classification result can be set according to the suggestions of experts.

[0157] In this embodiment, the alarm threshold is set according to the usage frequency of the command, the accelerating effect of the frequently used command on the instability of the element is considered, which helps to more accurately reflect the risk level of the frequently used command, thereby providing more personalized and accurate fault evaluation results for the high-voltage switch system. Moreover, this dynamic threshold setting can focus on the usage of high-frequency commands, which helps to more accurately evaluate the potential fault level of the high-voltage switch system and improves the fault evaluation response efficiency.

[0158] Embodiment 8: The application provides a high-voltage switch mechanical state diagnosis and evaluation control system; comprising

[0159] The data acquisition module is configured to acquire real-time data and historical data of the high-voltage switch, acquire real-time data, historical residual life, historical data and corresponding fault labels of each element in the high-voltage switch, and deliver the acquisition results to the failure rate prediction module.

[0160] The failure rate prediction module is configured to predict the failure rate of the element by combining the real-time data and historical data of the element and the historical data of the high-voltage switch, the historical residual life and historical data of the element, and the corresponding fault labels, and deliver the prediction results to the evaluation module.

[0161] evaluation module: used for evaluating the failure condition of the element, judging the current mechanical state of the high-voltage switch. If the failure rate of the element exceeds the preset failure threshold, the result is that the mechanical state of the high-voltage switch is faulty; if not, further judge the failure rate of all commands in the command tree. When judging the failure rate of the command, if the failure rate of the command is lower than the preset alarm threshold, the result is that the current mechanical state of the high-voltage switch is qualified, otherwise the result is that the mechanical state of the high-voltage switch is abnormal. Then the result is delivered to the alarm module;

[0162] alarm module: used for issuing alarm prompts. When the result is that the current mechanical state of the high-voltage switch is qualified, no alarm is issued; when the result is that the mechanical state of the high-voltage switch is abnormal, an alarm prompt of the abnormal mechanical state of the high-voltage switch is issued; when the result is that the mechanical state of the high-voltage switch is faulty, a prompt of the faulty mechanical state of the high-voltage switch is issued.

[0163] The above are preferred embodiments of the present application, and are not intended to limit the protection scope of the present application, therefore: any equivalent changes made according to the structure, shape, principle of the present application should be covered within the protection scope of the present application.

Claims

1. A high-voltage switch mechanical state diagnosis evaluation control method, characterized by, The method comprises the following steps: Data collection: collecting real-time data and historical data of the high-voltage switch, collecting real-time data, historical residual life, historical data and corresponding fault labels of each element in the high-voltage switch; Predicting the residual life of the element: constructing a residual life prediction model, inputting the real-time data and historical data of the element into the residual life prediction model to obtain the residual life of the element; Obtaining the failure rate of the element: constructing a failure rate prediction model, training the model based on the historical data of the high-voltage switch, the historical residual life and historical data of the element and the corresponding fault labels, inputting the residual life of the element, the real-time data of the high-voltage switch and the element into the trained model to obtain the failure rate of the element; Element judgment: judging whether the failure rate of each element exceeds a preset failure threshold: If yes, a high-voltage switch mechanical state failure alarm is prompted; If no, a state evaluation step is performed; State evaluation: based on the failure rate of each element, the failure rate of all commands in a preset command tree is calculated, and it is judged whether the failure rate of the command is lower than a preset alarm threshold: If yes, a result that the current mechanical state of the high-voltage switch is qualified is displayed; If no, a high-voltage switch mechanical state abnormal alarm is prompted; The calculation model of the failure rate of the command is as follows: ； wherein, represents the failure rate of the command, represents the number of executions of the element in the current command, , represents the failure rate of the element; The construction process of the preset command tree is as follows: based on different scene requirements, a plurality of command trees are constructed, an externally triggered operation or a received command started under a certain condition is taken as a root node, an element in the high-voltage switch that needs to be triggered to meet a certain condition is taken as an intermediate node, and an operation triggered to meet the conditions of all intermediate nodes, that is, a target result, is taken as a leaf node to construct the command tree; The conditions met and the elements involved between the root node and each leaf node are taken as a command, and a command includes an externally triggered operation or a received command started under a certain condition, a triggering condition of the element, a triggering operation of the element and a target result.

2. The high-voltage switch mechanical state diagnosis evaluation control method according to claim 1, characterized by, After the state evaluation step, the following steps are further included: Obtaining the risk degree of the element: extracting the command whose failure rate is not lower than the preset alarm threshold as a risk command, obtaining the overall failure rate of each element in the current risk command to the risk command, and recording the overall failure rate as a total failure rate; Element maintenance verification: the elements are gradually simulated and replaced in the order from large to small according to the total failure rate, and the failure rate of the corresponding risk command after the simulation and replacement is calculated; Formulating a preset maintenance plan: the element maintenance verification step is performed on all risk commands until the failure rates of all risk commands are lower than the preset alarm threshold, the simulation and replacement results of all elements are taken as the preset maintenance plan, and the preset maintenance plan is output.

3. The high-voltage switch mechanical state diagnosis evaluation control method according to claim 2, characterized by, In the step of obtaining the risk degree of the element, the overall failure rate of the same element in all risk commands is taken as the maximum value, and the maximum value is taken as the total failure rate.

4. The high-voltage switch mechanical state diagnosis evaluation control method according to claim 2, characterized by, Before the step of formulating a preset maintenance plan, the following steps are further included: Risk command verification: judging whether the risk command has a simulated and replaced element: If yes, the element maintenance verification step is performed on the risk command whose failure rate is still not lower than the preset alarm threshold after the simulation and replacement; If no, the element maintenance verification step is performed on the risk command.

5. The high-voltage switch mechanical condition diagnosis evaluation control method according to claim 4, characterized by, A spare element database is further included, and the spare element database includes the element type and residual life of the spare element; In the component maintenance verification step, a spare component in the spare component database is extracted for simulation replacement.

6. The diagnostic evaluation control method of a high-voltage switchgear mechanical state according to any one of claims 1 to 5, characterized by, In the state evaluation step, if there is a multiple set subset relationship between commands in the preset command tree, only the failure rate of the command that is not a subset is calculated.

7. The diagnostic evaluation control method of a high-voltage switchgear mechanical state according to any one of claims 1 to 5, characterized by, Before the component judgment step, if the trigger conditions, execution objects and execution effects of multiple components are completely the same, the components are merged into a virtual component. The calculation model of the failure rate of the virtual component after merging is: ； wherein, represents a failure rate of a virtual element, represents the number of elements identical to the current trigger condition, execution object, and execution effect, represents a failure rate of the first element among all elements identical to the current trigger condition, execution object, and execution effect. represents a failure rate of the first element among all elements identical to the current trigger condition, execution object, and execution effect.

8. The high-voltage switch mechanical condition diagnosis evaluation control method according to claim 1, characterized by, In the state evaluation step, the preset alarm threshold of each command is determined according to the usage frequency of all commands in the command tree within a preset time.

9. The high-voltage switch mechanical condition diagnostic evaluation control method of claim 1, wherein, In the state evaluation step, the preset alarm threshold of each command is determined according to the importance of the execution result of all commands in the command tree.

10. A high voltage switch mechanical condition diagnostic evaluation control system characterized by, The high-voltage switch mechanical state diagnosis evaluation control method according to any one of claims 1-9, comprising a data acquisition module for acquiring real-time data and historical data of the high-voltage switch, acquiring real-time data, historical residual life, historical data and corresponding failure labels of each component in the high-voltage switch, and delivering the acquisition results to the failure rate prediction module; a failure rate prediction module for predicting the failure rate of the component by combining the residual life prediction model and the failure rate prediction model, the real-time data and the historical data of the component, the historical data of the high-voltage switch, the historical residual life and the historical data of the component, and the corresponding failure labels, and delivering the prediction results to the evaluation module; an evaluation module for evaluating the failure condition of the component, judging the current mechanical state of the high-voltage switch, and delivering the results to the alarm module; an alarm module for issuing an abnormal mechanical state alarm prompt or a mechanical state failure alarm prompt of the high-voltage switch.

Images