Automatic generation method based on large hydropower station equipment operation monthly report function

Abstract

An automatic generation method based on a large hydropower station equipment operation monthly report function comprises the following steps: firstly, pre-designing a format of an equipment operation monthly report needing to be generated, and automatically extracting a corresponding feature data sequence from a database stored in an equipment server; noise reduction processing is further carried out on the extracted feature data sequence, feature analysis is carried out, and a conclusion is obtained; then generating a monthly text report of equipment operation according to requirements, and marking places where the equipment operation parameters are abnormal; and finally, inserting the typical operation data trend chart into the equipment operation monthly text report to form a complete equipment operation monthly report, performing manual auditing, correction and finalizing, and outputting corresponding contents. The method provided by the invention has the advantages of liberating human resources, being high in analysis report accuracy, highlighting abnormal data key points and the like, and greatly promoting the development of the intelligent power station.

Description

Technical Field

[0001] The invention belongs to the technical field of automatic report generation, and in particular relates to an automatic generation method based on a monthly report function of equipment operation in a large hydropower station. Background Art

[0002] In the context of energy interconnection in today's society, the safe and stable operation of hydropower plant power generation equipment plays an increasingly important role in the interconnection of power systems and energy supply. Completing the monthly operation report of hydropower station equipment on time as one of the daily work tasks of operators plays an important reference role in accurately judging the health of equipment, timely discovering potential fault information of equipment, and further formulating equipment maintenance plans, providing strong support for the stable operation of the power system.

[0003] At present, the monthly operation report of hydropower station equipment mostly takes a lot of human resources to extract effective data information from scattered raw data, organize and analyze the data, and mostly rely on manual experience to finally draw conclusions. Obviously, with the rapid development of energy Internet, more and more clean energy power generation equipment is put into operation, and it is urgently needed to use digital, information and intelligent technology as a carrier to realize the function of automatically generating monthly operation reports of large hydropower station equipment.

[0004] In summary, the problems to be solved by the present invention are: 1. The production process of the monthly report on the operation of hydropower station equipment is inefficient, inaccurate, easily affected by human factors, and has great uncertainty. 2. If the accuracy and timeliness of the acquisition of a large amount of data are not fully guaranteed, the equipment analysis results will be inaccurate, and important latent equipment failures will be easily missed. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide an automatic generation method based on the monthly report function of the equipment operation of a large hydropower station, which has the advantages of liberating human resources, high accuracy of analysis reports, and highlighting abnormal data, thereby achieving the purpose of visual analysis of the operating parameters of the hydropower station equipment.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: The automatic generation method based on the monthly report function of large hydropower station equipment operation includes the following steps: Step 1: pre-design the format of the monthly equipment operation report to be generated; the format includes one or more of the equipment operation report function template, the picture layout function template, the natural language processing function template and the report document overall layout function template; Step 2: According to the characteristic item data required by the equipment operation report function template in step 1, the corresponding data sequence is automatically extracted from the database stored in the equipment server; Step 3: Perform data noise reduction processing on the feature data sequence extracted in step 2, and eliminate invalid data based on the equipment operation principle or experience; Step 4: Perform feature analysis on the data after noise reduction in step 3 to obtain conclusions, and extract corresponding feature values ​​according to the requirements of the equipment operation report function template in step 1; Step 5: For the feature values ​​extracted in step 4, use the natural language processing function template in step 1 to generate a monthly text report on equipment operation according to the requirements of the equipment operation report function template in step 1, and mark the abnormalities of equipment operation parameters in the report; Step 6: Based on the monthly text report on equipment operation in step 4, use the picture layout function template in step 1 and the equipment operation report function template in step 1 to form a complete monthly report on equipment operation; Step 7: Manually review, finalize, and output the complete equipment operation monthly report formed in step 6. If it does not meet the requirements, re-execute steps 1 to 7.

[0007] Preferably, the equipment operation report function template in step 1 refers to: comprehensively reflecting the operating status and health status of the equipment, including the change range and change trend conclusions of each specific parameter of the equipment, the number of times the equipment analog quantity exceeds the limit, the number of switch actions and unit start-up and shutdown times, and the equipment operation time, combined with the actual operation status of the hydropower station equipment.

[0008] Preferably, the image layout function template in step 1 refers to: inserting a typical operation data trend chart into the generated equipment operation monthly report, adding a layout item of the typical operation data trend chart, and setting the size and position of the layout item.

[0009] Preferably, the natural language processing function template in step 1 refers to: setting the detailed description content of the equipment operation report function template in step 1, using the natural language processing function to complete text error correction, classification, understanding, grammatical analysis, etc., and expressing it in a standardized formal language in a computer.

[0010] Preferably, the report document overall layout function template in step 1 refers to: for the text report formed after the natural language processing function in step 1, the report document overall layout function template is further used to design various levels of titles, font size, font type, paragraph spacing and line spacing for the text report content as the layout template of the final equipment operation monthly report.

[0011] Preferably, the characteristic item data required by the equipment operation report function template in step 2 is automatically extracted from the database stored in the equipment server in the following way: the data to be extracted is the operation parameters in brackets in the detailed description of the equipment operation report function template, and all data values ​​of this characteristic value stored in the current month are extracted from the database to form a new characteristic sequence, and corresponding characteristic sequences are formed for different analysis aspects of the equipment operation report.

[0012] Preferably, the data noise reduction processing for the extracted feature data sequence in step 3 means: in the process of collecting equipment status operating parameters in the database, due to the short-term failure of the data monitoring system, transmission system, equipment external environment, and sensors, blank values ​​or abnormal value bad data exist in the database. The operating status of the equipment cannot be accurately judged based on the analysis of the bad data, so the blank values ​​or abnormal values ​​in the feature sequence need to be repaired and restored.

[0013] Preferably, in step 3, the support vector machine model is used to implement the noise reduction process of the initial data, and the initial feature sequence data , using the support vector machine model for regression fitting, and selecting the model independent variables based on the correlation between the data , n is the total number of samples, set the mapping function , the initial feature sequence data is mapped to a high-dimensional feature space, and a linear function is used for regression in this high-dimensional space, as shown in formula (1): (1); in, w , b are the threshold and regression parameters, respectively; The process of solving the threshold and regression parameters can be expressed as formula (2): (2); Where: C , , y i are the penalty parameter, slack variable and fitting marker variable of the support vector machine model respectively. The support vector machine model needs to be solved with the assistance of the duality of the Lagrangian algorithm, according to the KKT condition, as shown in formula (3): (3); (4); The calculation can be obtained: (5); Where: , , , , is the key parameter of the Lagrange multiplier method, is the kernel function; The expression is as follows: (6); It is the key width parameter of the kernel function, and its value is usually greater than 0.

[0014] Preferably, in step 3, data noise reduction is performed on the extracted feature data sequence, and the steps are as follows: Step 3.1: Input the feature data sequence to be cleaned , using the support vector machine model to calculate ,in , is the original feature sequence, is the model prediction value; Step 3.2: Determine the abnormality of the data and calculate Is it established? is the standard deviation of the model residual, which is applicable to the operating data of hydropower plant equipment monitoring. If it is established, it is judged as a normal value, otherwise it is judged as an abnormal value; When the characteristic sequence data is judged as an outlier, ,Compare i=n Is it true? If not, return to step 1 and re-execute the process; If true, the cleaned data is output.

[0015] Preferably, in step 4, feature analysis is performed on the data after noise reduction processing in step 3, and the method is: counting the extreme values ​​and average values ​​of the feature data, and the fluctuation range of the feature data; according to the threshold value of the parameter setting during normal operation of the large hydropower station equipment, automatically counting the frequency of feature value exceeding the limit, and the change law of the feature data, and extracting the corresponding feature value according to the requirements of the equipment operation report function template in step 1.

[0016] Preferably, in step 5, the feature values ​​extracted in step 4 are used to output a monthly text report on equipment operation that meets the requirements using a natural language processing function template, and the text content is modified, grammatically corrected, correctly classified, etc., and expressed in a computer using a standardized formal language; and in the text report, special marks are made where the equipment operating parameters are abnormal, and this special marking method is red bold.

[0017] Preferably, in step 6, based on the monthly text report on equipment operation in step 4, a typical operation data trend chart is inserted into the generated monthly report on equipment operation using the picture layout function template in step 1.

[0018] The present invention can achieve the following beneficial effects: The present invention uses digital, information and intelligent technologies as means to regularly analyze the data of hydropower station equipment operation and automatically generate a monthly report on the unit operation. Compared with the traditional analysis report acquisition method, the method proposed in the present invention has the advantages of liberating human resources, high accuracy of analysis reports, and highlighting abnormal data. It achieves the purpose of visual analysis of the operating parameters of hydropower station equipment and greatly promotes the development of smart power stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1 It is a flow chart of the present invention; Figure 2 This is a diagram of the noise reduction process of feature data sequence based on the SVM model of the present invention; Figure 3 It is a startup process curve diagram of the present invention; Figure 4 A typical operating data trend diagram of the present invention (pattern 1); Figure 5 This is a typical operation data trend chart of the present invention (pattern 2); Figure 6 This is a typical operation data trend chart of the present invention (pattern three); Figure 7 This is a typical operating data trend diagram of the present invention (pattern four). DETAILED DESCRIPTION

[0020] The preferred solution is Figures 1 to 7 As shown in the figure, the automatic generation method based on the monthly report function of the equipment operation of large hydropower stations uses digital, information and intelligent technologies as means to regularly analyze the data of the equipment operation of hydropower stations and automatically generate monthly unit operation reports. Compared with the traditional analysis report acquisition method, the method proposed in the present invention has the advantages of liberating human resources, high accuracy of analysis reports, and highlighting abnormal data, which greatly promotes the development of smart power stations. Figure 1 As shown, the specific steps are as follows: Step 1: Pre-design the format of the monthly equipment operation report to be generated, including the equipment operation report function template, picture layout function template, natural language processing function template, and report document overall layout function template; The equipment operation report function template means: it can comprehensively reflect the operation status and health status of the equipment, including the change range and change trend conclusion of each specific parameter of the equipment, the number of times the equipment analog quantity exceeds the limit, the number of switch actions and unit start-up and shutdown times, the equipment operation time, etc. Combined with the actual operation of the hydropower station equipment, the equipment operation report function template in step 1 is elaborated from the following aspects: 1. Unit vibration, swing, and pressure pulsation; 2. Bearing temperature, oil level and oil level exceeding limit of three parts of the unit; 3. Unit generator temperature; 4. Pure water system of the unit; 5. Oil and gas content of the main transformer of the unit; 6. Dry-type transformer for self-use of the unit; 7. Unit speed control hydraulic system; 8. Rapid door of water inlet of the unit; 9. Unit top cover drainage system; 10. Technical water supply system of the unit; 11. The unit start-up and shutdown process.

[0021] The picture layout function template refers to: an automatic generation method based on the monthly report function of large hydropower station equipment operation proposed by the present invention, in which a typical operation data trend chart is inserted into the generated monthly equipment operation report, and layout items of the typical operation data trend chart are further added, and the size and position of the layout items are set.

[0022] The natural language processing function template refers to: setting the detailed description content of the equipment operation report function template in step 1, using the natural language processing function to complete text error correction, classification, understanding, grammatical analysis, etc., and expressing it in a standardized formal language in a computer.

[0023] Taking the unit vibration, swing and pressure pulsation as an example, the following are the expressions: upper guide X-direction swing: ()~()μm, upper guide Y-direction swing: ()~()μm, water guide X-direction swing: ()~()μm, water guide Y-direction swing: ()~()μm, lower guide X-direction swing: ()~()μm, lower guide Y-direction swing: ()~()μm, stator core horizontal X-direction vibration: ()~()μm, stator core horizontal Y-direction vibration: ()~()μm, top cover X-direction horizontal vibration: ()~()μm, top cover Y Horizontal vibration in X direction: ()~()μm, horizontal vibration in X direction of upper frame: ()~()μm, horizontal vibration in Y direction of upper frame: ()~()μm, horizontal vibration in X direction of stator frame: ()~()μm, horizontal vibration in Y direction of stator frame: ()~()μm, pressure pulsation under top cover: ()~()%, pressure pulsation between runner and guide vane: ()~()%, pressure pulsation at volute inlet: ()~()%, pressure pulsation at upstream of tailwater pipe: ()~()%, pressure pulsation at downstream of tailwater pipe: ()~()%.

[0024] The content in () is the characteristic value inserted after device data analysis.

[0025] The report document overall layout function template refers to: for the text report formed after the natural language processing function in step 1, the report document overall layout function template is further used to design the titles, font size, font type, paragraph spacing, line spacing, etc. of the text report content as the layout template of the final equipment operation monthly report.

[0026] Step 2: According to the characteristic item data required by the equipment operation report function template in step 1, the corresponding data sequence is automatically extracted from the database stored in the equipment server; The characteristic item data required by the equipment operation report function template automatically extracts the corresponding data sequence from the database saved by the equipment server; the specific extraction method is: the data to be extracted is the operating parameters in brackets in the detailed description of the equipment operation report function template, that is, the operating parameters in "()", and all data values ​​of this characteristic value stored in the current month are extracted from the database to form a new characteristic sequence, and corresponding characteristic sequences are formed for different analysis aspects of the equipment operation report (the 11 aspects pointed out in step 1).

[0027] Step 3: Perform data noise reduction on the feature data sequence extracted in step 2, and eliminate invalid data based on the equipment operation principle or experience, such as only retaining the equipment data in a stable operating state and discarding the data in the transition process such as startup and shutdown, so as to maintain the valid information in the data sequence while minimizing the impact of interference information on the conclusion.

[0028] Data noise reduction is performed on the extracted feature data sequence. Specifically, in the process of collecting equipment status operating parameters in the database, due to adverse factors such as the data monitoring system, transmission system, equipment external environment, and short-term sensor failure, blank values ​​or abnormal values ​​may exist in the database. The operating status of the equipment cannot be accurately judged based on the analysis of the bad data. Therefore, the blank values ​​or abnormal values ​​in the feature sequence need to be repaired and restored. This not only maintains the valid information in the data sequence, but also minimizes the impact of interference information on the conclusion.

[0029] In step 3 above, the support vector machine model (SVM) is used to implement the noise reduction process of the initial data. , SVM is used for regression fitting, and the model independent variables are selected according to the correlation between the data , n is the total number of samples, set the mapping function , the initial feature sequence data is mapped to a high-dimensional feature space, and a linear function is used to perform regression in this high-dimensional space, as shown in formula (1). w , b are the threshold and regression parameters respectively.

[0030] (1); The process of solving the threshold and regression parameters can be expressed as formula (2): (2); Where: C , , y i are the penalty parameter, slack variable and SVM fitting marker variable respectively. SVM needs to use the duality of Lagrangian algorithm to assist in solving it, according to KKT (KarushKuhn-Tucker) condition, as shown in formula (3): (3); (4); The calculation can be obtained: (5); Where: , , , , is the key parameter of the Lagrange multiplier method, is the kernel function.

[0031] The classification of kernel functions generally includes Gaussian kernel function, linear kernel function, polynomial kernel function, radial kernel function, Sigmoid, etc. The kernel function of this paper selects Gaussian kernel function, which has few internal key parameters, can map multi-dimensional spatial relationships, and has stable performance. The expression is as follows: It is the key width parameter of the kernel function, and its value is usually greater than 0.

[0032] (6); In step 3, the extracted feature data sequence is subjected to data noise reduction processing. The specific flow chart is shown in the attached figure. Figure 2 As shown, the steps are as follows: 1. Input the feature data sequence to be cleaned , using the SVM model to calculate ,in , is the original feature sequence, is the model prediction value; 2. Analyze the abnormality of the data and calculate Is it established? is the standard deviation of the model residual, which is applicable to the operating data of hydropower plant equipment monitoring. If it is established, it is judged as a normal value, otherwise it is judged as an abnormal value.

[0033] When the characteristic sequence data is judged as an outlier, ,Compare i=n Is it true? If not, return to step 1 and re-execute the process. If it is true, output the cleaned data.

[0034] Step 4: Perform feature analysis on the data after noise reduction in step 3 to obtain conclusions, and extract corresponding feature values ​​according to the requirements of the equipment operation report function template in step 1; Perform feature analysis on the data after noise reduction processing in step 3. The specific method is: count the extreme values ​​and average values ​​of the feature data, and the fluctuation range of the feature data; according to the threshold value set by the parameters during normal operation of the large hydropower station equipment, automatically count the frequency of feature value exceeding the limit, and the change law of the feature data, and extract the corresponding feature value according to the requirements of the equipment operation report function template in step 1.

[0035] Step 5: For the feature values ​​extracted in step 4, use the natural language processing function template in step 1 to generate a monthly text report on equipment operation in accordance with the specific requirements of the equipment operation report function template in step 1, and mark the abnormalities in the equipment operation parameters in the report.

[0036] For the feature values ​​extracted in step 4, use the natural language processing function template to output a monthly text report on equipment operation that meets the requirements, and perform text modification, grammatical correction, correct classification, etc. on the text content, and express it in a standardized formal language in the computer. In the text report, special marks are made for abnormal equipment operation parameters, and this special marking method is red bold.

[0037] Step 6: Based on the monthly text report on equipment operation in step 4, use the picture layout function template in step 1 and the equipment operation report function template in step 1 to form a complete monthly report on equipment operation.

[0038] Based on the monthly text report on equipment operation in step 4, using the image layout function template in step 1, typical operation data trend charts are inserted into the generated monthly report on equipment operation, such as the unit start-up and shutdown curves, the temperature change curves of the three bearing bushes of the unit, etc., and the size and position of the trend chart are set to finally form a complete monthly report on equipment operation.

[0039] Step 7: Manually review, finalize, and output the complete equipment operation monthly report formed in step 6. If it does not meet the requirements, re-execute steps 1 to 7.

[0040] This example analysis part takes the 4F hydro-turbine generator unit of the Three Gorges Power Station as an example to demonstrate the process of generating its equipment operation monthly report. First, through step 1, pre-design the format of the equipment operation monthly report to be generated, select the equipment area, unit number and corresponding time; then click the query button to generate the corresponding equipment operation monthly report. This process executes steps 2 to 6; finally, execute step 7 to manually review, finalize, and output the equipment operation monthly report content that meets the requirements. The typical operation data trend chart is shown in the attached figure (3-7).

[0041] The specific content of the monthly equipment operation report is as follows: Operation report of Unit 4 of Three Gorges Power Station Period: August 1, 2023 to August 31, 2023 Unit 4 of the Three Gorges Power Station was put into operation on October 28, 2003, with a cumulative power generation of 48.29 billion kWh. The most recent maintenance: D-level maintenance, construction period 2023-11-27 to 2023-12-09, maintenance project: D-level maintenance of the #4 generator-transformer unit of the Three Gorges Left Bank Power Plant.

[0042] During the statistical period: the machine was started 6 times and stopped 7 times, with a total monthly operation time of 616.62 hours, normal operation data, and 5 items were manually analyzed and confirmed. The basic operation status of the subsystem is as follows: 1. Unit vibration: The second-level alarms of unit vibration, swing and pressure pulsation alarm occurred 0 times in total.

[0043] Upper guide X-direction swing: 26.3~143.4μm, upper guide Y-direction swing: 36.6~158.7μm, water guide X-direction swing: 80~145.9μm, water guide Y-direction swing: 75.1~149.5μm, lower guide X-direction swing: 92.8~143.4μm, lower guide Y-direction swing: 107.4~143.4μm, stator core horizontal X-direction vibration: 3~9.2μm, stator core horizontal Y-direction vibration: 16.9~27.1μm, top cover X-direction horizontal vibration: 26.6~43.3μm, top cover Y-direction water Flat vibration: 12~22.1μm, horizontal vibration of upper frame in X direction: 6.7~32.2μm, horizontal vibration of upper frame in Y direction: 6.7~48.2μm, horizontal vibration of stator frame in X direction: 76~84.7μm, horizontal vibration of stator frame in Y direction: 51.1~60.3μm, pressure pulsation under top cover: 0~2.47%, pressure pulsation between runner and guide vane: 0~1.23%, pressure pulsation at volute inlet: 0~5.66%, pressure pulsation upstream of tailwater pipe: 0~3.4%, pressure pulsation downstream of tailwater pipe: 0~1.49%.

[0044] 2. Unit bearings: The bearing temperature, oil temperature, and oil level of the three bearings of the unit exceeded the limit 0 times, and the abnormal trend of each data occurred 0 times. After manual analysis and inspection, no manual focus is needed. Upper guide bearing temperature: 40.5~46.3℃, stable trend; lower guide bearing temperature: 41.2~46.8℃, stable trend; thrust bearing temperature: 74~78.5℃, stable trend; water guide bearing temperature: 32.16~58.94, stable trend.

[0045] 3. Generator temperature: stator slot, core, air cooler, pure water (water-cooled unit) temperature exceeded the limit 0 times, and each data trend abnormality occurred 0 times. After manual analysis and inspection, no manual focus is required. Stator slot temperature: yes~yes℃, stable trend; stator core temperature: 43.9~71.7℃, trend query data is empty, and the trend cannot be determined; stator pure water temperature before water inlet: 37.2~61.2℃, stable trend; generator cold air temperature: 30.1~44.1℃, hot air temperature: 39.9~67℃.

[0046] 4. Pure water system: conductivity: 1.4~1.42us / cm. Expansion water tank: 63~76%.

[0047] 5. Oil and gas content of main transformer: main transformer oil temperature: 51.2~55.5℃, oil and gas content exceeded the limit 0 times, and oil and gas content data trend abnormality occurred 0 times. After manual analysis and inspection, no manual focus is needed; acetylene content: 0~0ppm, hydrogen content: 1.6~2.2ppm, total hydrocarbon content: 3.3~3.6ppm.

[0048] 6. Self-use dry-type transformers: The temperature of each dry-type transformer exceeded the limit 0 times. After manual analysis and inspection, no manual focus is needed; 400V self-use transformer winding temperature: 20~149.9℃, stable trend, excitation transformer winding temperature: 45.7~87.7℃, stable trend, factory high-voltage transformer operating temperature: 28.6~33.8℃, trend query data is empty, and the trend cannot be determined.

[0049] 7. Speed ​​regulating hydraulic system: The main operating data of the speed regulating hydraulic system exceeded the limit 0 times. After manual analysis and inspection, no manual focus is needed. Oil pump loading interval: 0.0~0.0 minutes, oil pump single loading duration: 0.0~0.0 seconds; air tank refill times 0, the total oil volume of the system is 6.89.

[0050] 8. Water inlet rapid door: The rapid door slides down and automatically rises 0 times.

[0051] 9. Top cover drainage system: No. 1 top cover pump started and ran 0 times, cumulative running time: 0 minutes, No. 2 top cover pump started and ran 0 times, cumulative running time: 0 minutes. Number of times the top cover water level rose too quickly: 0 times.

[0052] 10. Technical water supply: upstream water level: 153.261~159.711 meters, main pipe pressure at water intake: 0.969~1.101MPa, pressure after unit pressure reducing valve: 0.04~0.39MPa, pressure after main transformer pressure reducing valve: 0.04~0.41MPa, technical water supply flow of generator: 0~1790.78m³ / h, technical water supply flow of main transformer: 0~513.3m³ / h.

[0053] 11. Start-up and shutdown process: The last startup was normal, the last shutdown was normal, and the startup and shutdown process curves are attached.

[0054] The present invention uses digital, information and intelligent technologies as means. First, the format of the monthly equipment operation report to be generated is pre-designed, and the corresponding data sequence is automatically extracted from the database stored in the equipment server; further, the extracted characteristic data sequence is subjected to data noise reduction processing, and characteristic analysis is performed to obtain conclusions; then, a monthly equipment operation text report is generated as required, and the places where the equipment operation parameters are abnormal are marked; finally, a typical operation data trend chart is inserted into the monthly equipment operation text report to form a complete monthly equipment operation report, which is manually reviewed, revised, and finalized, and the corresponding content is output.

[0055] The above embodiments are only preferred technical solutions of the present invention and should not be regarded as limiting the present invention. The protection scope of the present invention shall be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. Automatic generation method based on the monthly report function of large hydropower station equipment operation, characterized by The following steps are involved: Step 1: pre-design the format of the monthly equipment operation report to be generated; the format includes one or more of the equipment operation report function template, the picture layout function template, the natural language processing function template and the report document overall layout function template; Step 2: According to the characteristic item data required by the equipment operation report function template in step 1, the corresponding data sequence is automatically extracted from the database stored in the equipment server; Step 3: Perform data noise reduction processing on the feature data sequence extracted in step 2, and eliminate invalid data based on the equipment operation principle or experience; Step 4: Perform feature analysis on the data after noise reduction in step 3 to obtain conclusions, and extract corresponding feature values ​​according to the requirements of the equipment operation report function template in step 1; Step 5: For the feature values ​​extracted in step 4, use the natural language processing function template in step 1 to generate a monthly text report on equipment operation according to the requirements of the equipment operation report function template in step 1, and mark the abnormalities of equipment operation parameters in the report; Step 6: Based on the monthly text report on equipment operation in step 4, use the picture layout function template in step 1 and the equipment operation report function template in step 1 to form a complete monthly report on equipment operation; Step 7: Manually review, finalize, and output the complete equipment operation monthly report formed in step 6. If it does not meet the requirements, re-execute steps 1 to 7.

2. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: The equipment operation report function template in step 1 refers to: comprehensively reflecting the operating status and health status of the equipment, including the change range and change trend conclusion of each specific parameter of the equipment, the number of times the equipment analog quantity exceeds the limit, the number of switch actions and unit start-up and shutdown times, and the equipment operation time, combined with the actual operation status of the hydropower station equipment.

3. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: The image layout function template in step 1 refers to: inserting a typical operation data trend chart into the generated equipment operation monthly report, adding a layout item of the typical operation data trend chart, and setting the size and position of the layout item.

4. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: The natural language processing function template in step 1 refers to: setting the detailed description content of the equipment operation report function template in step 1, using the natural language processing function to complete text error correction, classification, understanding, grammatical analysis, etc., and expressing it in a standardized formal language in the computer.

5. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: The report document overall layout function template in step 1 refers to: for the text report formed after the natural language processing function in step 1, the report document overall layout function template is further used to design various levels of titles, font size, font type, paragraph spacing and line spacing for the text report content, as the layout template of the final equipment operation monthly report.

6. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: The characteristic item data required by the equipment operation report function template in step 2 is automatically extracted from the database saved by the equipment server. The extraction method of the corresponding data sequence is as follows: the data to be extracted is the operation parameters in brackets in the detailed description of the equipment operation report function template, and all data values ​​of this characteristic value stored in the current month are extracted from the database to form a new characteristic sequence. Corresponding characteristic sequences are formed for different analysis aspects of the equipment operation report.

7. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: The data noise reduction processing for the extracted feature data sequence in step 3 means: in the process of collecting equipment status operating parameters in the database, due to the short-term failure of the data monitoring system, transmission system, equipment external environment, and sensors, blank values ​​or abnormal value bad data exist in the database. The operating status of the equipment cannot be accurately judged based on the analysis of the bad data, so the blank values ​​or abnormal values ​​in the feature sequence need to be repaired and restored.

8. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: In step 3, the support vector machine model is used to implement the noise reduction process of the initial data. , using the support vector machine model for regression fitting, and selecting the model independent variables based on the correlation between the data , n is the total number of samples, set the mapping function , the initial feature sequence data is mapped to a high-dimensional feature space, and a linear function is used for regression in this high-dimensional space, as shown in formula (1): (1)； in, w , b are the threshold and regression parameters, respectively; The process of solving the threshold and regression parameters can be expressed as formula (2): (2)； Where: C , , y i are the penalty parameter, slack variable and fitting marker variable of the support vector machine model respectively. The support vector machine model needs to be solved with the assistance of the duality of the Lagrangian algorithm, according to the KKT condition, as shown in formula (3): (3)； (4)； The calculation can be obtained: (5)； Where: , , , , is the key parameter of the Lagrange multiplier method, is the kernel function; The expression is as follows: (6)； It is the key width parameter of the kernel function, and its value is usually greater than 0.

9. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: In step 3, the extracted feature data sequence is subjected to data denoising, and the steps are as follows: Step 3.1: Input the feature data sequence to be cleaned , using the support vector machine model to calculate ,in , is the original feature sequence, is the model prediction value; Step 3.2: Determine the abnormality of the data and calculate Is it established? is the standard deviation of the model residual, which is applicable to the operating data of hydropower plant equipment monitoring. If it is established, it is judged as a normal value, otherwise it is judged as an abnormal value; When the characteristic sequence data is judged as an outlier, ,Compare i=n Is it true? If not, return to step 1 and re-execute the process; If true, the cleaned data is output.

10. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: In the step 4, feature analysis is performed on the data after noise reduction processing in step 3, and the method is: counting the extreme values ​​and average values ​​of the feature data, and the fluctuation range of the feature data; according to the threshold value of the parameter setting during the normal operation of the large hydropower station equipment, the frequency of feature value exceeding the limit and the change law of the feature data are automatically counted, and the corresponding feature value is extracted according to the requirements of the equipment operation report function template in step 1.

11. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: In the step 5, the feature values ​​extracted in the step 4 are used to output a monthly text report on equipment operation that meets the requirements using a natural language processing function template, and the text content is modified, grammatically corrected, correctly classified, etc., and expressed in a computer using a standardized formal language; and in the text report, special marks are made where the equipment operation parameters are abnormal, and this special marking method is red bold.

12. The automatic generation method based on the monthly report function of large hydropower station equipment operation according to claim 1 is characterized by: In the step 6, based on the monthly text report on equipment operation in step 4, a typical operation data trend chart is inserted into the generated monthly report on equipment operation using the picture layout function template in step 1.

Images