A monitoring method for the entire life cycle of power equipment based on multi-parameter comprehensive analysis

Abstract

The present application discloses a method for monitoring the entire life cycle of electric power equipment based on multi-parameter comprehensive analysis, and relates to the field of electric power technology. The method includes determining the four processing situations of emergency, serious, general and suspected according to the maintenance processing steps, formulating the first and second maintenance plans, and according to the second maintenance plan, respectively performing fault detection on the four types of maintenance data of overall disassembly inspection, local maintenance, routine inspection and appearance inspection of the electric power equipment after the maintenance processing steps, and respectively performing comparison processing on the first and second types of maintenance data and the first and third types of maintenance data to achieve comprehensive fault risk determination, and according to the first maintenance plan, detecting the first or second type of project data of all projects within the first or second preset time period after the maintenance processing steps, and performing abnormal analysis on the first or second type of project data to achieve comprehensive fault risk determination. The method for monitoring the entire life cycle of electric power equipment based on multi-parameter comprehensive analysis provided by the present invention improves the efficiency of maintenance work.

Description

Technical Field

[0001] The present application relates to the field of electric power technology, and in particular to a method for monitoring the entire life cycle of electric power equipment based on comprehensive analysis of multiple parameters. Background Art

[0002] Power equipment life cycle fault monitoring plays an important role in the stable operation of the power system, equipment maintenance management, and the improvement of power supply reliability. Through an effective fault monitoring system, early detection and prevention of faults can be achieved, equipment life can be extended, maintenance costs can be optimized, and data support can be provided for equipment management, thereby improving the operating efficiency and safety of the entire power system.

[0003] The existing maintenance data are not interconnected, such as state quantity detection results, power outage maintenance results, manual analysis results, including background data. All in-station maintenance data are separated. This results in low reliability of fault detection and judgment results and waste of resources. Summary of the invention

[0004] In order to overcome the above-mentioned deficiencies of the prior art, the present application provides a method and system for monitoring the entire life cycle of power equipment based on multi-parameter comprehensive analysis.

[0005] In a first aspect, the present application provides a method for monitoring the entire life cycle of electric power equipment based on multi-parameter comprehensive analysis, the method comprising:

[0006] Obtain maintenance processing steps formulated for the power equipment, determine four processing situations of emergency, serious, general and suspected for the maintenance processing steps, output processing situation one or processing situation two or processing situation three or processing situation four, formulate a first maintenance plan according to processing situation one or processing situation two, and formulate a second maintenance plan according to processing situation three or processing situation four;

[0007] According to the second maintenance scheme, fault detection is performed on the first type of maintenance data of overall disassembly inspection, the second type of maintenance data of local maintenance, the third type of maintenance data of routine inspection, and the fourth type of maintenance data of appearance inspection of the power equipment after the maintenance processing step, and a first fault prediction result, a second fault prediction result, a third fault prediction result, and a fourth fault prediction result are output;

[0008] Compare the first type of maintenance data with the second type of maintenance data to obtain a first fault variation ratio, and compare the first type of maintenance data with the third type of maintenance data to obtain a second fault variation ratio;

[0009] When at least one of the first fault variation ratio and the second fault variation ratio is equal to a preset ratio threshold, output the information to be determined; when both the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold, perform a comprehensive fault risk determination on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and output a comprehensive fault risk warning result;

[0010] According to the first maintenance plan or the information to be judged, the first category project data of all items of the power equipment in the first preset time period after the maintenance step is detected. When it is judged that the first category project data is normal, the second category project data of all items of the power equipment in the second preset time period after the maintenance step is continued to be detected. When it is judged that the second category project data is abnormal, a comprehensive fault risk judgment is performed on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and the comprehensive fault risk warning result is output. The maintenance processing steps are optimized accordingly, and a maintenance step optimization plan is output.

[0011] Preferably, real-time detection data and historical record data of the power equipment are obtained, and according to the real-time detection data, the recorded data to be analyzed of the same category as the real-time detection data are extracted from the historical record data, and the real-time detection data are compared with the recorded data to be analyzed to obtain abnormal data comparison results, and the corresponding fault risk is determined according to the abnormal data comparison results;

[0012] According to the failure risk, timely adopt the corresponding maintenance treatment steps;

[0013] Counting the data change status of the power equipment after the maintenance processing step to obtain expected reference change data;

[0014] An abnormal comparison is performed on the data change status and the expected reference change data to obtain an abnormal comparison result of the change data, and the fault severity of the abnormal comparison result of the change data is determined. When the fault severity is an emergency fault, it is immediately processed and the processing situation 1 is output. When the fault severity is a serious fault, a recent planned processing is performed and the processing situation 2 is output. When the fault severity is a general fault, a preset plan is performed to carry out a power outage processing and the processing situation 3 is output. When the fault severity is a suspected fault, further verification is required and the processing situation 4 is output.

[0015] According to the processing situation 1 or the processing situation 2, a first maintenance plan is formulated;

[0016] According to the processing situation three or processing situation four, formulate a second maintenance plan.

[0017] Preferably, according to the second maintenance plan, the power equipment after the maintenance processing step is subjected to an overall disassembly inspection, and the first type of maintenance data is output;

[0018] Performing local maintenance on the power equipment after the maintenance processing step, and outputting second type maintenance data;

[0019] Performing routine inspection on the power equipment after the maintenance processing steps, and outputting third-category maintenance data;

[0020] Performing a visual inspection on the electric power equipment after the maintenance processing step, and outputting fourth type maintenance data;

[0021] Acquire historical fault record information of power equipment in historical periods, and train a first fault prediction model based on the historical fault record information and historical record data;

[0022] The first type of maintenance data, the second type of maintenance data, the third type of maintenance data and the fourth type of maintenance data are respectively input into the first fault prediction model for testing to obtain a first fault prediction result, a second fault prediction result, a third fault prediction result and a fourth fault prediction result.

[0023] Preferably, the first type of maintenance data and the second type of maintenance data are subjected to difference processing to obtain a difference value of one;

[0024] According to the difference value one, extracting a historical difference value one from the historical record data, and training a second fault prediction model according to the historical difference value one and the historical fault record information;

[0025] Inputting the difference value 1 into the second fault prediction model for testing to obtain a new fault prediction result 1;

[0026] According to the first type of maintenance data and the second type of maintenance data, reference maintenance data 1 is obtained, the difference value 1 is averaged to obtain an average difference value 1, and the average difference value 1 and the reference maintenance data 1 are integrated to obtain first comparative analysis data;

[0027] Performing statistics on the variation amplitudes of the corresponding data of the first type of maintenance data and the first comparative analysis data to obtain a first fault variation amplitude;

[0028] Performing statistics on the variation amplitudes of the respective data of the second type of maintenance data and the first comparative analysis data to obtain a second fault variation amplitude;

[0029] Calculating the ratio between the first fault variation range and the second fault variation range to obtain a first fault variation range ratio;

[0030] Performing difference processing on the first type of maintenance data and the third type of maintenance data to obtain a difference value 2;

[0031] According to the difference value 2, extracting the historical difference value 2 from the historical record data, and training a third fault prediction model according to the historical difference value 2 and the historical fault record information;

[0032] Inputting the difference value 2 into the third fault prediction model for testing to obtain a new fault prediction result 2;

[0033] According to the second type of maintenance data and the third type of maintenance data, reference maintenance data 2 is obtained, the difference value 2 is averaged to obtain average difference value 2, and the average difference value 2 and reference maintenance data 2 are integrated to obtain second comparative analysis data;

[0034] Performing statistics on the variation amplitudes of the corresponding data of the first type of maintenance data and the second comparative analysis data to obtain a third fault variation amplitude;

[0035] Performing statistics on the variation amplitudes of the corresponding data of the third type of maintenance data and the second comparative analysis data to obtain a fourth fault variation amplitude;

[0036] A second fault variation ratio is obtained by calculating a ratio between the third fault variation range and the fourth fault variation range.

[0037] Preferably, when at least one of the first fault variation ratio and the second fault variation ratio is equal to a preset ratio threshold, it is impossible to perform a comprehensive fault risk determination on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and output the information to be determined;

[0038] When the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold, the new fault prediction result one and the new fault prediction result two are combined to perform a comprehensive fault risk judgment on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and output a comprehensive fault risk warning result.

[0039] Preferably, according to the first maintenance plan or the information to be determined, the first type of project data of all projects of the power equipment in the first preset time period after the maintenance step is detected;

[0040] Obtaining first-category standard reference data, performing abnormal data comparison on the first-category project data and the first-category standard reference data to obtain comparison result one, and when comparison result one is abnormal, performing comprehensive fault risk determination on the first fault prediction result, the second fault prediction result, the third fault prediction result, and the fourth fault prediction result according to comparison result one, outputting a comprehensive fault risk warning result, optimizing the maintenance processing steps accordingly, and outputting a maintenance step optimization plan;

[0041] When the comparison result 1 is normal, the second type of project data of all projects of the power equipment in the second preset time period after the detection maintenance step, the second preset time period is earlier than the first preset time period, and the second type of project data is other project data except the first type of project data;

[0042] Obtain the second category standard reference data, perform abnormal data comparison on the second category project data and the second category standard reference data to obtain comparison result two; when comparison result two is abnormal, perform a comprehensive fault risk judgment on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result according to comparison result two, output a comprehensive fault risk warning result, optimize the maintenance processing steps accordingly, and output a maintenance step optimization plan.

[0043] Second, the full life cycle monitoring system for power equipment based on multi-parameter comprehensive analysis includes:

[0044] A processing plan formulation unit is used to obtain maintenance processing steps formulated for the power equipment, determine four processing situations of emergency, serious, general and suspected for the maintenance processing steps, output processing situation one or processing situation two or processing situation three or processing situation four, formulate a first maintenance plan according to processing situation one or processing situation two, and formulate a second maintenance plan according to processing situation three or processing situation four;

[0045] The first fault analysis unit is used to perform fault detection on the first type of maintenance data of overall disassembly inspection, the second type of maintenance data of local maintenance, the third type of maintenance data of routine inspection and the fourth type of maintenance data of appearance inspection of the power equipment after the maintenance processing step according to the second maintenance plan, and output a first fault prediction result, a second fault prediction result, a third fault prediction result and a fourth fault prediction result;

[0046] The second fault analysis unit is used to compare the first type of maintenance data with the second type of maintenance data to obtain a first fault variation ratio, and compare the first type of maintenance data with the third type of maintenance data to obtain a second fault variation ratio;

[0047] The first comprehensive determination unit is used to determine that when at least one of the first fault variation ratio and the second fault variation ratio is equal to a preset ratio threshold, output information to be determined, and when both the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold, output a comprehensive fault risk warning result;

[0048] The second comprehensive judgment unit is used to detect the first category of project data of all projects of the power equipment in the first preset time period after the maintenance step according to the first maintenance plan or the information to be judged. When it is judged that the first category of project data is normal, the second category of project data of all projects of the power equipment in the second preset time period after the maintenance step is continued to be detected. When it is judged that the second category of project data is abnormal, the comprehensive fault risk warning result is output, and the maintenance processing steps are optimized accordingly, and the maintenance step optimization plan is output.

[0049] Compared with the prior art, the present invention has the following characteristics and beneficial effects:

[0050] By performing corresponding maintenance processing steps according to the faults appearing in the real-time detection data of the power equipment, it is judged whether there is a fault risk and the fault severity is divided, so as to formulate corresponding treatment opinions for different fault severities, so as to avoid the high errors caused by the implementation of unified treatment opinions and the low efficiency of the entire information processing work. According to the four judged treatment situations, two maintenance plans are formulated, namely the first maintenance plan and the second maintenance plan. The first maintenance plan is to perform data maintenance on the power equipment under four different conditions after the maintenance processing steps, and predict the corresponding fault risks according to the established model to obtain four corresponding fault prediction results. According to the degree of correlation between the four fault prediction results, it is judged whether a comprehensive maintenance plan can be performed. The judgment of the risk of abnormal failure is carried out. If it is possible, the four fault prediction results are directly combined to make the final fault judgment. If not, the second maintenance plan is carried out, that is, the first type of project data of the power equipment in the first preset time period after the maintenance step is judged as abnormal. If there is an abnormal condition, the four fault prediction results are directly combined to make a comprehensive fault risk judgment. If there is no abnormality, the time period is shortened, that is, the second type of project data of all projects of the power equipment in the second preset time period after the maintenance step is continued to be detected, and the previous judgment step is repeated. Finally, the maintenance processing step is optimized accordingly to obtain the maintenance step optimization plan, thereby improving the efficiency of the entire information processing work and the accuracy of fault judgment, and conducting scientific research on developability information to enhance the reliability of information analysis results. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a flowchart of the steps of the method for monitoring the entire life cycle of power equipment based on multi-parameter comprehensive analysis mainly embodied in this embodiment.

[0052] Figure 2 It is a structural block diagram of the electric power equipment full life cycle monitoring system based on multi-parameter comprehensive analysis mainly embodied in this embodiment. DETAILED DESCRIPTION

[0053] The present invention is further described in detail below with reference to the following examples.

[0054] Reference Figure 1 A method and system for monitoring the entire life cycle of power equipment based on multi-parameter comprehensive analysis, the method comprising the following steps:

[0055] S1. Obtain the maintenance processing steps formulated for the power equipment, determine the four processing situations of the maintenance processing steps, namely, emergency, serious, general, and suspected, and output processing situation one, processing situation two, processing situation three, or processing situation four. According to processing situation one or processing situation two, formulate a first maintenance plan, and according to processing situation three or processing situation four, formulate a second maintenance plan.

[0056] S2. According to the second maintenance plan, fault detection is performed on the first type of maintenance data of overall disassembly inspection, the second type of maintenance data of local maintenance, the third type of maintenance data of routine inspection and the fourth type of maintenance data of appearance inspection of the power equipment after the maintenance processing steps, and a first fault prediction result, a second fault prediction result, a third fault prediction result and a fourth fault prediction result are output.

[0057] S3. Compare the first type of maintenance data with the second type of maintenance data to obtain a first fault variation ratio, and compare the first type of maintenance data with the third type of maintenance data to obtain a second fault variation ratio.

[0058] S4. When at least one of the first fault variation ratio and the second fault variation ratio is equal to the preset ratio threshold, output the information to be judged; when both the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold, perform a comprehensive fault risk judgment on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and output a comprehensive fault risk warning result.

[0059] S5. According to the first maintenance plan or the information to be judged, the first category of project data of all projects of the power equipment in the first preset time period after the maintenance step is detected. When it is determined that the first category of project data is normal, the second category of project data of all projects of the power equipment in the second preset time period after the maintenance step is continued to be detected. When it is determined that the second category of project data is abnormal, a comprehensive fault risk judgment is performed on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and a comprehensive fault risk warning result is output. The maintenance processing steps are optimized accordingly, and a maintenance step optimization plan is output.

[0060] Specifically, the data collected here involve historical data of power equipment (including power outage test data, state quantity detection data, equipment failure records), real-time detection data (including temperature, current, voltage, power, micro-water, chromatographic vibration, etc.), and equipment operation data (such as equipment inspection conditions, state quantity detection conditions, oil product testing conditions, etc. These three types of data can constitute a database, so that when the following same analysis steps are performed in the future, the data objects can be referenced to improve the utilization rate of resources, and a model can be established based on the database to achieve scientific research on the developable data and analysis of data regularity changes. First, the real-time detection data of the power equipment is obtained (the real-time detection data here also includes temperature, current, voltage, power, micro-water, Equipment operation data other than chromatographic vibration, etc.), by comparing the real-time detection data with the recorded data to be analyzed of the same category, timely determine whether there is a fault risk, and adopt corresponding maintenance processing steps according to the fault risk, and then make a differentiated analysis of the multiple data changes of the power equipment after the maintenance processing steps are implemented, and make judgments on the severity of different faults, that is, make a distinction between the four levels of emergency faults, serious faults, general faults and suspected faults, that is, divide them into processing situation one, processing situation two, processing situation three and processing situation four, so as to facilitate subsequent targeted processing, improve the orderliness and accuracy of the entire information processing process, and formulate and allocate two maintenance plans for the above four processing situations, Processing situations 1 and 2 are both assigned to the first maintenance plan for subsequent information detection and analysis, and processing situations 3 and 4 are both assigned to the second maintenance plan for subsequent information detection and analysis, so as to facilitate diversified and differentiated analysis and processing for conditions of different fault severity. The second maintenance plan is to analyze and judge the interrelated changes in the data detected by the above four processing situations. First, fault prediction is performed on the four processing situations to obtain four fault prediction results, and then the occurrence of new faults is predicted based on the data differences between them. The fault change amplitude between the new fault and the two fault prediction results is used to determine whether a comprehensive fault risk can be determined. When a more accurate judgment cannot be made, then Processing the first maintenance plan, that is, judging whether the first type of project data of the power equipment in the first preset time period after the maintenance processing step is abnormal, when it is judged to be normal, continuing the same step analysis of the second type of project data in the second preset time period, when it is judged to be abnormal, the above-mentioned four fault prediction results can be combined to perform a comprehensive fault risk judgment, and based on the analysis of abnormal data, the cause of the fault risk can be extracted, so as to facilitate targeted optimization of the maintenance steps based on the maintenance processing steps, and finally obtain an optimized maintenance step plan. Through the above-mentioned entire information processing steps, the more accurate verification of the effect of the maintenance processing steps is improved.

[0061] The specific step S1 includes the following sub-steps:

[0062] Acquire real-time detection data and historical record data of the power equipment, extract the record data to be analyzed that is of the same category as the real-time detection data from the historical record data based on the real-time detection data, compare the real-time detection data with the record data to be analyzed to obtain abnormal data comparison results, and determine the corresponding fault risk based on the abnormal data comparison results.

[0063] According to the risk of failure, take appropriate maintenance and treatment steps in a timely manner.

[0064] Statistics are collected on the data changes of the power equipment after the maintenance process to obtain the expected reference change data.

[0065] An abnormal comparison is performed on the data change status and the expected reference change data to obtain the abnormal comparison result of the change data, and the fault severity of the abnormal comparison result of the change data is determined. When the fault severity is an emergency fault, it is immediately processed and the processing situation one is output. When the fault severity is a serious fault, the short-term plan processing is carried out and the processing situation two is output. When the fault severity is a general fault, the preset plan is carried out to carry out power outage processing and the processing situation three is output. When the fault severity is a suspected fault, further verification is required and the processing situation four is output.

[0066] According to treatment situation one or treatment situation two, formulate the first maintenance plan.

[0067] According to treatment situation three or treatment situation four, formulate the second maintenance plan.

[0068] Specifically, such as real-time detection data (if the category of the target data to be detected is transformer data, then the real-time detection data refers to the data of detecting the transformer. The transformer mainly has the function of voltage transformation, and the internal structure is coils, cores, etc. The data includes high and low voltage side voltage, current, power, core clamp current, bushing temperature, bushing oil test results, on-load voltage regulation excessive resistance, partial discharge, fan heat plate temperature sampling, micro-water sampling, etc., and can also refer to other categories of data, such as: current transformer, voltage transformer and other categories of data. Here, the data of monitoring transformers is taken as an example), historical record data (refers to all reference data of normal and abnormal monitored and recorded in the historical period). According to the data, the data category contains transformers. According to the transformer category to which the real-time detection data belongs, the recorded data to be analyzed is extracted from the historical record data, that is, the normal reference data and abnormal reference data recorded in the historical period), fault risk (that is, the real-time detection data and the recorded data to be analyzed are compared to see if the data is abnormal, and the fault risk is preliminarily judged according to the data abnormality comparison result. The fault risk judged here is determined according to the historical fault characteristics), maintenance processing steps (according to the fault and maintenance plan matching table, to match the corresponding most reasonable maintenance processing steps), data change status (after the implementation of the maintenance processing steps The data changes that occurred), expected reference change data (referring to the characteristics of previous historical data, there will be a fixed range of changes for each maintenance processing step after implementation, that is, the interval value of normal data changes, which is the expected reference change data), abnormal comparison results of change data (such as data change conditions: the data of a type of casing temperature has changed, and the data change range is evaluated according to the numerical degree of data change. If it is expressed as a percentage, the data change degree is 60%, and the range interval value of the expected reference change data is 40%-50%, then this type of data is abnormal), judge the fault severity of the abnormal comparison results of the change data (according to the fault level Comparison table: If 0%-10% and including 10% is a suspected fault, 10%-20% and including 20% ​​is a general fault, 20%-50% and including 50% is a serious fault, and 50%-100% but not including 100% is an emergency fault, then according to the difference between the above data change degree and the range interval value of the expected reference change data is 10%, it is determined to be a suspected fault, which is processing situation four. The explanations for processing situation one, processing situation two, and processing situation three are the same and will not be explained here). According to the severity of the processing situation, the subsequent analysis of different processing methods is carried out accordingly to reduce the large errors caused by a single processing method.

[0069] The specific step S2 includes the following sub-steps:

[0070] According to the second maintenance plan, the power equipment is disassembled and inspected as a whole after the maintenance process, and the first type of maintenance data is output.

[0071] Perform local maintenance on the power equipment after the maintenance processing steps and output the second type of maintenance data.

[0072] Carry out routine inspection on the power equipment after the maintenance process and output the third type of maintenance data.

[0073] Perform visual inspection on the power equipment after the maintenance process and output the fourth type of maintenance data.

[0074] The historical fault record information of the power equipment in the historical period is obtained, and the first fault prediction model is trained according to the historical fault record information and the historical record data.

[0075] The first type of maintenance data, the second type of maintenance data, the third type of maintenance data and the fourth type of maintenance data are respectively input into the first fault prediction model for testing to obtain the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result.

[0076] Specifically, first of all, in order to discover hidden dangers of equipment in operation and prevent accidents or equipment damage, the equipment is inspected, tested or monitored. Preventive tests include power outage tests, live detection and online monitoring. Power outage test: various inspections and tests on the equipment status performed by operating personnel on site under the condition of exiting operation. Live detection: on-site detection of equipment status quantities under operating conditions, also including tests conducted by taking oil or gas samples. Generally, portable detection equipment is used for short-term detection, which is different from continuous or periodic online monitoring. Online monitoring: continuous or periodic automatic monitoring and detection of the status of power equipment without power outage. The detection of subsequent data is carried out separately under the above three conditions, such as the first type of maintenance data detected under power-off test conditions or the first type of maintenance data detected under energized detection conditions or the first type of maintenance data detected under online monitoring conditions. The interpretation of the other three types of maintenance data is the same, such as the overall dismantling inspection of the power equipment after the maintenance processing steps (refers to the overall dismantling inspection, repair, replacement and related tests of the power equipment, and it is Class A maintenance, and the data obtained by the detection is the first type of maintenance data), local maintenance of the power equipment after the maintenance processing steps (local maintenance of the power equipment, dismantling inspection, repair, replacement and related tests of major components and parts, and it is Class B maintenance, and the data obtained by the detection is the second type of maintenance data), routine inspection of the power equipment after the maintenance processing steps (routine inspection, testing and maintenance of the power equipment, including a few The data obtained by the test are the third type of maintenance data, the appearance inspection of the power equipment after the maintenance processing steps (appearance inspection of the power equipment, simple fault elimination and live detection, and it is a D-level maintenance, the data obtained by the test are the fourth type of maintenance data, and live detection is performed during D-level maintenance), the first fault prediction model (that is, the abnormal data in the historical record data is input into the model, and the historical faults corresponding to the abnormal data are also input into the model for training to obtain a matching table between the abnormal data and the faults, and the first type of maintenance data is subsequently input into the trained first fault prediction model for testing. According to the abnormal condition of the first type of maintenance data, the corresponding fault type can be predicted and matched, that is, the first fault prediction result is obtained, and so on. The explanations for the remaining three fault prediction results are the same and will not be explained here).

[0077] The specific step S3 includes the following sub-steps:

[0078] The first type of maintenance data and the second type of maintenance data are processed to obtain a difference value of one.

[0079] According to the difference value one, a historical difference value one is extracted from the historical record data, and according to the historical difference value one and the historical fault record information, a second fault prediction model is trained to obtain.

[0080] The difference value one is input into the second fault prediction model for testing to obtain a new fault prediction result one.

[0081] According to the first type of maintenance data and the second type of maintenance data, reference maintenance data one is obtained, the difference value one is averaged to obtain an average difference value one, and the average difference value one and the reference maintenance data one are integrated to obtain first comparative analysis data.

[0082] The first fault variation amplitude is obtained by performing variation amplitude statistics of each corresponding data of the first type of maintenance data and the first comparative analysis data.

[0083] The second fault variation range is obtained by performing variation range statistics of each corresponding data of the second type maintenance data and the first comparative analysis data.

[0084] The ratio between the first fault variation range and the second fault variation range is calculated to obtain the first fault variation range ratio.

[0085] The first type of maintenance data and the third type of maintenance data are processed for difference to obtain a difference value of two.

[0086] According to the difference value two, the historical difference value two is extracted from the historical record data, and according to the historical difference value two and the historical fault record information, a third fault prediction model is trained to obtain.

[0087] The difference value 2 is input into the third fault prediction model for testing to obtain a new fault prediction result 2.

[0088] According to the second type of maintenance data and the third type of maintenance data, reference maintenance data 2 is obtained, the difference value 2 is averaged to obtain average difference value 2, and the average difference value 2 and reference maintenance data 2 are integrated to obtain second comparative analysis data.

[0089] The third fault variation range is obtained by performing variation range statistics of each corresponding data of the first type of maintenance data and the second comparative analysis data.

[0090] The variation amplitude of each corresponding data of the third type maintenance data and the second comparative analysis data is statistically performed to obtain the fourth fault variation amplitude.

[0091] A second fault variation ratio is obtained by calculating the ratio between the third fault variation range and the fourth fault variation range.

[0092] Specifically, such as the difference value one (multiple categories of data in the first category of maintenance data include multiple categories of data in the second category of maintenance data, so the change difference values ​​between the multiple data of the same category in the two are statistically analyzed), the second fault prediction model (such as extracting from the historical record data the same processing steps as the difference value one, and both are the historical difference value one obtained by processing the first category maintenance data and the second category maintenance data of the same category in the history. The historical difference value one here does not refer to the data of one historical period, but to the data of different historical periods. The historical difference values ​​one of different historical periods correspond to a kind of fault. Then, statistical training can be used to obtain a matching table between data and faults, that is, the second fault prediction model, and the difference value one is subsequently input into the training The second fault prediction model is tested in the next step. According to the abnormal condition of the difference value one, the corresponding fault type can be predicted and matched, that is, a new fault prediction result one is obtained), reference maintenance data one (refers to the data that changes normally after the effective effect of the maintenance processing step, if temperature is taken as an example), average difference one (that is, the average temperature value between the abnormal temperature data change range under the A-level maintenance condition and the abnormal temperature data change range under the B-level maintenance condition, in order to obtain a relatively stable abnormal temperature data change range), the first comparative analysis data (that is, the relatively stable abnormal temperature data change range and the reference maintenance data one: temperature, are summed to obtain a relatively stable abnormal temperature data, and used as a data reference object for subsequent comparison) , the first fault variation range and the second fault variation range (such as calculating the difference between the first type of maintenance data belonging to the category of temperature and the first control analysis data belonging to the category of temperature to obtain the difference a, and calculating the difference between the second type of maintenance data belonging to the category of temperature and the first control analysis data belonging to the category of temperature to obtain the difference b, and then calculating the ratio of the difference a and the difference b of the same temperature to obtain a ratio, and so on, the other remaining abnormal data of the same category in the first type of maintenance data and the second type of maintenance data are all subjected to the same difference calculation processing with the first control analysis data, and then the difference obtained between the two is calculated to obtain the remaining multiple ratios, and these calculated and statistically obtained multi-category ratios are calculated. Draw the change curve route to obtain the first fault change amplitude: that is, the first fault change correlation degree between the first fault prediction result and the new fault prediction result one and the second fault change amplitude: that is, the second fault change correlation degree between the second fault prediction result and the new fault prediction result one), the first fault change amplitude ratio (that is, the ratio of the first fault change amplitude and the second fault change amplitude obtained above is calculated one by one: the ratio of each change node in the change curve route is calculated to obtain the first fault change amplitude ratio: that is, the ratio between the first fault change correlation degree and the second fault change correlation degree), and process the first type of maintenance data and the third type of maintenance data in the same steps as the first type of maintenance data and the second type of maintenance data mentioned above,Among them, the third fault prediction model (such as extracting the same processing steps as the difference value 2 from the historical record data, and both are the historical difference value 2 obtained by processing the first-category maintenance data and the third-category maintenance data of the same category in the history, the historical difference value 2 here does not refer to the data of one historical period, but refers to the data of different historical periods. The historical difference values ​​1 of different historical periods all correspond to a kind of fault, then the matching table between data and faults can be obtained by statistical training, that is, the third fault prediction model, and then the difference value 2 is input into the trained third fault prediction model for testing. According to the abnormal condition of the difference value 2, the corresponding fault type can be predicted and matched, that is, the new fault prediction result 2 is obtained), and finally the second fault variation ratio is obtained, and the step explanations and instructions are the same, so no further explanation is given here. ,

[0093] The specific step S4 includes the following sub-steps:

[0094] When at least one of the first fault variation ratio and the second fault variation ratio is equal to the preset ratio threshold, it is impossible to perform a comprehensive fault risk judgment on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and output the information to be judged.

[0095] When the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold, the new fault prediction result one and the new fault prediction result two are combined to perform a comprehensive fault risk judgment on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and output a comprehensive fault risk warning result.

[0096] Specifically, when at least one of the first fault variation ratio and the second fault variation ratio is equal to the preset ratio threshold (such as the preset ratio threshold is 1, the first fault variation ratio is 1 / 2, and the second fault variation ratio is 1), it means that the second fault prediction result is very likely to be misjudged, so it is impossible to perform a comprehensive fault risk judgment on the first fault prediction result, the second fault prediction result, the third fault prediction result, and the fourth fault prediction result. When both the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold (such as the preset ratio threshold is 1, the first fault variation ratio is 1 / 2, and the second fault variation ratio is 2 / 3), it means that it can be ruled out. In the event of misjudgment, and based on the size obtained by the ratio of the two fault change amplitudes, the second fault prediction result is fine-tuned to tend to the first fault prediction result, and finally, combined with the new fault prediction result one and the new fault prediction result two, the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result are comprehensively judged for fault risk (for example, circuit parameter abnormality and equipment overheating are two common faults in the operation of electronic equipment, which combined may lead to comprehensive faults, such as performance degradation, intermittent operation, component damage and even complete system collapse), and finally, timely early warning notification of the comprehensive fault risk is given, and the detection and analysis data information are transmitted.

[0097] The specific step S5 includes the following sub-steps:

[0098] According to the first maintenance plan or the information to be determined, the first category of project data of all projects of the power equipment within the first preset time period after the maintenance step is detected.

[0099] Obtain the first type of standard reference data, perform abnormal data comparison between the first type of project data and the first type of standard reference data to obtain comparison result one; when comparison result one is abnormal, perform a comprehensive fault risk judgment on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result based on comparison result one, output a comprehensive fault risk warning result, optimize the maintenance processing steps accordingly, and output a maintenance step optimization plan.

[0100] When the comparison result is normal, the second type of project data of all projects of the power equipment in the second preset time period after the detection maintenance step, the second preset time period is earlier than the first preset time period, and the second type of project data is other project data except the first type of project data.

[0101] Obtain the second type of standard reference data, perform abnormal data comparison on the second type of project data and the second type of standard reference data to obtain comparison result two; when comparison result two is abnormal, perform a comprehensive fault risk judgment on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result according to comparison result two, output a comprehensive fault risk warning result, optimize the maintenance processing steps accordingly, and output a maintenance step optimization plan.

[0102] Specifically, such as the first preset time period (such as the test cycle of the test equipment is within 6 years), the first type of project data (such as the insulation resistance test, chromatography, insulation resistance, oil chromatography, water-soluble acid, acid value and the dielectric loss factor and capacitance of the winding together with the bushing, if there are 65 project data in total), the first type of standard reference data (referring to the standard reference data of the above 65 project data under normal circumstances), the comparison result one (referring to the comparison result of the first type of project data and the first type of standard reference data one by one, when the comparison result one is abnormal (when there is abnormal project data in the comparison result one, and the difference processing of the abnormal project data is performed to obtain the abnormal difference), according to the comparison result one, the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result are comprehensively judged for fault risk (that is, according to the abnormal degree of the specific subordinate project, the above four fault prediction results are comprehensively judged for the comprehensive fault risk of aggregation, and the aggregation here refers to which of the above four fault prediction results is specifically concentrated on for the evaluation and judgment of the main comprehensive fault. , and the comprehensive judgment method is the same as the above explanation, which will not be further explained here). When the comparison result is normal, the data detection and analysis of the second preset period (such as reducing the time on the basis of the first preset period, adjusting the detection of the trial cycle from 6 years to the detection of the trial cycle once every 3 years) will continue. The second type of project data (referring to other project data except the first type of project data) and the subsequent processing of the second type of project data are the same as the steps of the processing of the first type of project data, and the explanation is also the same, which will not be further explained here. Finally, according to the abnormal data obtained from all the above detections and analyses, the specific cause of the fault is evaluated, and the defect judgment of one or more specific sub-steps in the above maintenance processing steps is made, so as to facilitate the subsequent targeted adjustment of the sub-step, and according to the specific cause of the fault, the failure cause of the defective sub-step is judged accordingly. Finally, the overall optimization of the maintenance processing steps is realized to obtain a new maintenance step optimization plan, and the maintenance work is carried out. All the data information analyzed and processed above is transmitted to the database for timely data update and information warning transmission notification.

[0103] The electric power equipment full life cycle monitoring system based on multi-parameter comprehensive analysis, by applying the electric power equipment full life cycle monitoring method based on multi-parameter comprehensive analysis as mentioned above, includes a processing plan formulation unit, a first fault analysis unit, a second fault analysis unit, a first comprehensive judgment unit and a second comprehensive judgment unit, referring to Figure 2 , obtain the maintenance processing steps formulated by the power equipment through the processing plan formulation unit, judge the maintenance processing steps according to four processing situations: emergency, serious, general and suspected, output processing situation one or processing situation two or processing situation three or processing situation four, formulate a first maintenance plan according to processing situation one or processing situation two, and formulate a second maintenance plan according to processing situation three or processing situation four; through the first fault analysis unit, according to the second maintenance plan, perform fault detection on the first type of maintenance data of overall disassembly inspection, the second type of maintenance data of local maintenance, the third type of maintenance data of routine inspection and the fourth type of maintenance data of appearance inspection of the power equipment after the maintenance processing steps, and output a first fault prediction result, a second fault prediction result, a third fault prediction result and a fourth fault prediction result; through the second fault analysis unit, compare and process the first type of maintenance data and the second type of maintenance data to obtain a first fault variation ratio, and compare and process the first type of maintenance data and the third type of maintenance data to obtain a second fault variation ratio; through the comprehensive judgment unit, judge when the first When at least one of the first fault variation ratio and the second fault variation ratio is equal to the preset ratio threshold, the information to be judged is output; when both the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold, a comprehensive fault risk judgment is performed on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and a comprehensive fault risk warning result is output; through the comprehensive judgment second unit, according to the first maintenance plan or the information to be judged, the first category of project data of all projects of the power equipment in the first preset time period after the maintenance step is detected, when it is determined that the first category of project data is not abnormal, the second category of project data of all projects of the power equipment in the second preset time period after the maintenance step is continued to be detected; when it is determined that the second category of project data is abnormal, a comprehensive fault risk judgment is performed on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and a comprehensive fault risk warning result is output, and the maintenance processing step is optimized accordingly, and a maintenance step optimization plan is output.

[0104] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for monitoring the entire life cycle of power equipment based on multi-parameter comprehensive analysis, characterized in that: The following steps are involved: Obtain maintenance processing steps formulated for the power equipment, determine four processing situations of emergency, serious, general and suspected for the maintenance processing steps, output processing situation one or processing situation two or processing situation three or processing situation four, formulate a first maintenance plan according to processing situation one or processing situation two, and formulate a second maintenance plan according to processing situation three or processing situation four; According to the second maintenance scheme, fault detection is performed on the first type of maintenance data of overall disassembly inspection, the second type of maintenance data of local maintenance, the third type of maintenance data of routine inspection, and the fourth type of maintenance data of appearance inspection of the power equipment after the maintenance processing step, and a first fault prediction result, a second fault prediction result, a third fault prediction result, and a fourth fault prediction result are output; Compare the first type of maintenance data with the second type of maintenance data to obtain a first fault variation ratio, and compare the first type of maintenance data with the third type of maintenance data to obtain a second fault variation ratio; When at least one of the first fault variation ratio and the second fault variation ratio is equal to a preset ratio threshold, output the information to be determined; when both the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold, perform a comprehensive fault risk determination on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and output a comprehensive fault risk warning result; According to the first maintenance plan or the information to be judged, the first category project data of all items of the power equipment in the first preset time period after the maintenance step is detected. When it is judged that the first category project data is normal, the second category project data of all items of the power equipment in the second preset time period after the maintenance step is continued to be detected. When it is judged that the second category project data is abnormal, a comprehensive fault risk judgment is performed on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and the comprehensive fault risk warning result is output. The maintenance processing steps are optimized accordingly, and a maintenance step optimization plan is output.

2. The method for monitoring the entire life cycle of electric power equipment based on multi-parameter comprehensive analysis according to claim 1 is characterized in that: The steps of obtaining maintenance processing steps formulated for the power equipment, determining four processing situations of emergency, serious, general and suspected for the maintenance processing steps, outputting processing situation one or processing situation two or processing situation three or processing situation four, formulating a first maintenance plan according to processing situation one or processing situation two, and formulating a second maintenance plan according to processing situation three or processing situation four are specifically as follows: Acquire real-time detection data and historical record data of the power equipment, extract the to-be-analyzed recorded data of the same category as the real-time detection data from the historical record data according to the real-time detection data, compare the real-time detection data with the to-be-analyzed recorded data to obtain an abnormal data comparison result, and determine the corresponding fault risk according to the abnormal data comparison result; According to the failure risk, timely adopt the corresponding maintenance treatment steps; Counting the data change status of the power equipment after the maintenance processing step to obtain expected reference change data; An abnormal comparison is performed on the data change status and the expected reference change data to obtain an abnormal comparison result of the change data, and the fault severity of the abnormal comparison result of the change data is determined. When the fault severity is an emergency fault, it is immediately processed and the processing situation 1 is output. When the fault severity is a serious fault, a recent planned processing is performed and the processing situation 2 is output. When the fault severity is a general fault, a preset plan is performed to carry out a power outage processing and the processing situation 3 is output. When the fault severity is a suspected fault, further verification is required and the processing situation 4 is output. According to the processing situation 1 or the processing situation 2, a first maintenance plan is formulated; According to the processing situation three or processing situation four, formulate a second maintenance plan.

3. The method for monitoring the entire life cycle of electric power equipment based on multi-parameter comprehensive analysis according to claim 2 is characterized in that: According to the second maintenance scheme, the steps of performing fault detection on the first type of maintenance data of overall disassembly inspection, the second type of maintenance data of local maintenance, the third type of maintenance data of routine inspection, and the fourth type of maintenance data of appearance inspection of the power equipment after the maintenance processing step, and outputting the first fault prediction result, the second fault prediction result, the third fault prediction result, and the fourth fault prediction result are specifically: According to the second maintenance plan, the power equipment is disassembled and inspected as a whole after the maintenance process step, and the first type of maintenance data is output; Performing local maintenance on the power equipment after the maintenance processing step, and outputting second type maintenance data; Performing routine inspection on the power equipment after the maintenance processing steps, and outputting third-category maintenance data; Performing a visual inspection on the electric power equipment after the maintenance processing step, and outputting fourth type maintenance data; Acquire historical fault record information of power equipment in historical periods, and train a first fault prediction model based on the historical fault record information and historical record data; The first type of maintenance data, the second type of maintenance data, the third type of maintenance data and the fourth type of maintenance data are respectively input into the first fault prediction model for testing to obtain a first fault prediction result, a second fault prediction result, a third fault prediction result and a fourth fault prediction result.

4. The method for monitoring the entire life cycle of electric power equipment based on multi-parameter comprehensive analysis according to claim 2 is characterized in that: The steps of comparing the first type of maintenance data with the second type of maintenance data to obtain a first fault variation ratio, and comparing the first type of maintenance data with the third type of maintenance data to obtain a second fault variation ratio are specifically: Performing difference processing on the first type of maintenance data and the second type of maintenance data to obtain a difference value of one; According to the difference value one, extracting a historical difference value one from the historical record data, and training a second fault prediction model according to the historical difference value one and the historical fault record information; Inputting the difference value 1 into the second fault prediction model for testing to obtain a new fault prediction result 1; According to the first type of maintenance data and the second type of maintenance data, reference maintenance data 1 is obtained, the difference value 1 is averaged to obtain an average difference value 1, and the average difference value 1 and the reference maintenance data 1 are integrated to obtain first comparative analysis data; Performing statistics on the variation amplitudes of the corresponding data of the first type of maintenance data and the first comparative analysis data to obtain a first fault variation amplitude; Performing statistics on the variation amplitudes of the respective data of the second type of maintenance data and the first comparative analysis data to obtain a second fault variation amplitude; Calculating the ratio between the first fault variation range and the second fault variation range to obtain a first fault variation range ratio; Performing difference processing on the first type of maintenance data and the third type of maintenance data to obtain a difference value 2; According to the difference value 2, extracting the historical difference value 2 from the historical record data, and training a third fault prediction model according to the historical difference value 2 and the historical fault record information; Inputting the difference value 2 into the third fault prediction model for testing to obtain a new fault prediction result 2; According to the second type of maintenance data and the third type of maintenance data, reference maintenance data 2 is obtained, the difference value 2 is averaged to obtain average difference value 2, and the average difference value 2 and reference maintenance data 2 are integrated to obtain second comparative analysis data; Performing statistics on the variation amplitudes of the corresponding data of the first type of maintenance data and the second comparative analysis data to obtain a third fault variation amplitude; Performing statistics on the variation amplitudes of the corresponding data of the third type of maintenance data and the second comparative analysis data to obtain a fourth fault variation amplitude; A second fault variation ratio is obtained by calculating a ratio between the third fault variation range and the fourth fault variation range.

5. The method for monitoring the entire life cycle of electric power equipment based on multi-parameter comprehensive analysis according to claim 4 is characterized in that: When at least one of the first fault variation ratio and the second fault variation ratio is equal to a preset ratio threshold, outputting information to be determined; when both the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold, performing a comprehensive fault risk determination on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and outputting a comprehensive fault risk warning result, specifically, the step of: When at least one of the first fault variation ratio and the second fault variation ratio is equal to a preset ratio threshold, it is impossible to perform a comprehensive fault risk determination on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and output the information to be determined; When the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold, the new fault prediction result one and the new fault prediction result two are combined to perform a comprehensive fault risk judgment on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and output a comprehensive fault risk warning result.

6. The method for monitoring the entire life cycle of electric power equipment based on multi-parameter comprehensive analysis according to claim 5 is characterized in that: According to the first maintenance plan or the information to be determined, the first type of project data of all projects of the power equipment in the first preset time period after the maintenance step is detected, and when it is determined that the first type of project data is normal, the second type of project data of all projects of the power equipment in the second preset time period after the maintenance step is continuously detected, and when it is determined that the second type of project data is abnormal, a comprehensive fault risk determination is performed on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and a comprehensive fault risk warning result is output, and the maintenance processing step is optimized accordingly, and the steps of outputting the maintenance step optimization plan are specifically as follows: According to the first maintenance plan or the information to be determined, detecting the first type of project data of all projects of the power equipment within a first preset time period after the maintenance step; Obtaining first-category standard reference data, performing abnormal data comparison on the first-category project data and the first-category standard reference data to obtain comparison result one, and when comparison result one is abnormal, performing comprehensive fault risk determination on the first fault prediction result, the second fault prediction result, the third fault prediction result, and the fourth fault prediction result according to comparison result one, outputting a comprehensive fault risk warning result, optimizing the maintenance processing steps accordingly, and outputting a maintenance step optimization plan; When the comparison result 1 is normal, the second type of project data of all projects of the power equipment in the second preset time period after the detection maintenance step, the second preset time period is earlier than the first preset time period, and the second type of project data is other project data except the first type of project data; Obtain the second category standard reference data, perform abnormal data comparison on the second category project data and the second category standard reference data to obtain comparison result two; when comparison result two is abnormal, perform a comprehensive fault risk judgment on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result according to comparison result two, output a comprehensive fault risk warning result, optimize the maintenance processing steps accordingly, and output a maintenance step optimization plan.

7. The power equipment full life cycle monitoring system based on multi-parameter comprehensive analysis is characterized by: The system is used to implement the method for monitoring the entire life cycle of electric power equipment based on multi-parameter comprehensive analysis as described in any one of claims 1 to 6, comprising: A processing plan formulation unit is used to obtain maintenance processing steps formulated for the power equipment, determine four processing situations of emergency, serious, general and suspected for the maintenance processing steps, output processing situation one or processing situation two or processing situation three or processing situation four, formulate a first maintenance plan according to processing situation one or processing situation two, and formulate a second maintenance plan according to processing situation three or processing situation four; The first fault analysis unit is used to perform fault detection on the first type of maintenance data of overall disassembly inspection, the second type of maintenance data of local maintenance, the third type of maintenance data of routine inspection and the fourth type of maintenance data of appearance inspection of the power equipment after the maintenance processing step according to the second maintenance plan, and output a first fault prediction result, a second fault prediction result, a third fault prediction result and a fourth fault prediction result; The second fault analysis unit is used to compare the first type of maintenance data with the second type of maintenance data to obtain a first fault variation ratio, and compare the first type of maintenance data with the third type of maintenance data to obtain a second fault variation ratio; A first comprehensive determination unit is used to determine that when at least one of the first fault variation ratio and the second fault variation ratio is equal to a preset ratio threshold, output information to be determined; when both the first fault variation ratio and the second fault variation ratio are not equal to the preset ratio threshold, perform a comprehensive fault risk determination on the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result, and output a comprehensive fault risk warning result; The second comprehensive judgment unit is used to detect the first category of project data of all projects of the power equipment in the first preset time period after the maintenance step according to the first maintenance plan or the information to be judged. When it is judged that the first category of project data is normal, the second category of project data of all projects of the power equipment in the second preset time period after the maintenance step continues to be detected. When it is judged that the second category of project data is abnormal, the first fault prediction result, the second fault prediction result, the third fault prediction result and the fourth fault prediction result are comprehensively judged for fault risk, the comprehensive fault risk warning result is output, the maintenance processing steps are optimized accordingly, and the maintenance step optimization plan is output.

Images