A method for evaluating and early warning of equipment state of a hydropower station based on transfer learning

Abstract

This invention proposes a method for assessing and issuing early warnings for hydropower station equipment status based on transfer learning. The method includes: acquiring operating parameters of core hydropower station equipment; verifying the completeness of the collected operating parameters; and integrating them to generate a set of basic parameters for the core equipment. Based on the set of basic parameters, the method calculates the correlation between equipment characteristics and operating conditions according to preset correlation calculation rules, generating a set of characteristic operating condition correlation coefficients. The method compares the set of characteristic operating condition correlation coefficients with the baseline coefficients of rated parameters to establish a hierarchical classification result of evaluation indicators. Based on the hierarchical classification result of evaluation indicators, the method matches similar source domain data to generate a data source adapted to the transfer learning model. Combining the data source adapted to the transfer learning model and real-time operating parameters, the method obtains early warning thresholds for each level, compares the real-time correlation coefficients with the early warning thresholds for each level, and generates an equipment status early warning judgment result.

Description

Technical Field

[0001] This invention belongs to the field of state early warning technology, specifically relating to a method for state assessment and early warning of hydropower station equipment based on transfer learning. Background Technology

[0002] The field of status early warning technology encompasses technologies related to the monitoring, assessment, and anomaly early warning of industrial equipment operating status. Its core lies in identifying potential failure risks through the collection, processing, and analysis of equipment operating data, providing support for equipment maintenance and safe operation. It is widely used in large-scale industrial sectors such as hydropower stations. Hydropower station equipment has complex structures and high operating loads; its stable operation directly impacts overall production efficiency and safety. Therefore, status assessment methods and anomaly early warning technologies based on equipment operating data are crucial components of the hydropower station operation and maintenance system.

[0003] Existing technologies for hydropower station equipment status assessment and early warning based on transfer learning have several shortcomings: data collection lacks specificity, and the key parameters of core equipment are not systematically screened and classified for verification, which can easily lead to missing or redundant basic data, affecting the reliability of subsequent analysis results; the data processing stage does not achieve deep correlation calculation between equipment parameters and operating conditions, resulting in insufficient data representation and difficulty in accurately reflecting the equipment's operating status; the assessment indicators are not hierarchically divided based on the equipment's rated parameters, lacking adaptability to the actual operating characteristics of the equipment, leading to weak assessment specificity; when applying the transfer learning model, the source domain data is not accurately matched and screened, resulting in insufficient model adaptability and affecting the accuracy of the assessment; and the early warning judgment does not combine real-time parameters with hierarchical early warning thresholds for accurate comparison, making it difficult to achieve accurate and efficient early warning. Summary of the Invention

[0004] The present invention aims to at least partially solve one of the technical problems in the related art.

[0005] Therefore, the first objective of this invention is to propose a method for assessing and providing early warning of the condition of hydropower station equipment based on transfer learning.

[0006] The second objective of this invention is to propose a hydropower station equipment status assessment and early warning device based on transfer learning.

[0007] The third objective of this invention is to provide a computer device.

[0008] A fourth objective of this invention is to provide a non-transitory computer-readable storage medium.

[0009] To achieve the above objectives, a first aspect of the present invention proposes a method for assessing and issuing early warnings of the condition of hydropower station equipment based on transfer learning, comprising:

[0010] Obtain the operating parameters of the core equipment of the hydropower station, verify the completeness of the collected operating parameters, and integrate them to generate a set of basic parameters for the core equipment; Based on the core equipment basic parameter set, the correlation between equipment characteristics and operating conditions is calculated according to the preset correlation calculation rules, and a characteristic operating condition correlation coefficient set is generated. The set of correlation coefficients for the characteristic working conditions is compared with the benchmark coefficients of the rated parameters to establish the hierarchical division results of the evaluation indicators; Based on the hierarchical division results of the evaluation indicators, similar source domain data are matched to generate a transfer learning model that adapts to the data source. By combining the transfer learning model with the data source and real-time operating parameters, the warning thresholds for each level are obtained. The real-time correlation coefficients are compared with the warning thresholds for each level to generate the device status warning judgment result.

[0011] In one embodiment of the present invention, the step of acquiring the operating parameters of the core equipment of the hydropower station, verifying the completeness of the collected operating parameters, and integrating them to generate a basic parameter set for the core equipment includes: The system obtains the turbine runner diameter, blade reference angle, runner inlet and outlet pressure, and speed fluctuation value; the generator stator winding resistance, stator winding temperature, rotor vibration amplitude, and output current stability; and the main transformer rated capacity, rated voltage, and rated current parameters. It then integrates the original parameters of various equipment to generate a summary of the original parameters of the equipment. Based on the summary of the original parameters of the equipment, the parameters are broken down according to the type of hydro-generator set and main transformer, the order of similar parameters is sorted out, the matching of the parameters that should be collected for each type of equipment with the actual collected parameters is checked, and the status of missing parameters is determined. The parameters that have been verified and confirmed to be correct are integrated to generate a set of basic parameters for the core equipment of the hydropower station. The set of basic parameters for the core equipment of the hydropower station includes a subset of core parameters for the turbine, a subset of core parameters for the generator, and a subset of core parameters for the main transformer.

[0012] In one embodiment of the present invention, the step of calculating the correlation between equipment characteristics and operating conditions based on the core equipment basic parameter set and according to preset correlation calculation rules to generate a characteristic operating condition correlation coefficient set includes: Based on the core equipment basic parameter set, the turbine runner inlet and outlet pressures, blade reference angles, and speed fluctuation values ​​are extracted. The ratio of the runner inlet and outlet pressure difference to the blade reference angle is calculated. This ratio is then multiplied by the speed fluctuation value to generate turbine characteristic operating condition related parameters. Based on the core equipment basic parameter set, the generator stator winding temperature, stator winding resistance, and rotor vibration amplitude are extracted. The ratio of stator winding temperature to stator winding resistance is calculated. The difference between this ratio and the rotor vibration amplitude is calculated to generate generator characteristic operating condition related parameters. Based on the associated parameters of the turbine characteristic operating conditions and the associated parameters of the generator characteristic operating conditions, the calculation results of the two types of equipment are summarized to generate a set of associated coefficients of equipment characteristic operating conditions. The set of associated coefficients of equipment characteristic operating conditions includes a subset of associated coefficients of turbine generator set characteristic operating conditions and a subset of associated coefficients of main transformer characteristic operating conditions.

[0013] In one embodiment of the present invention, the step of comparing the set of characteristic working condition correlation coefficients with the baseline coefficients of rated parameters to establish the evaluation index hierarchy results includes: Call the set of correlation coefficients for equipment characteristic operating conditions, obtain the reference coefficients for rated parameters of the hydro-generator unit and the reference coefficients for rated parameters of the main transformer, and integrate the two types of reference coefficients to generate a set of rated reference coefficients for the equipment. Based on the set of correlation coefficients for equipment characteristic operating conditions and the set of equipment rated reference coefficients, the correlation coefficients of the hydro-generator unit and the corresponding reference coefficients, and the correlation coefficients of the main transformer and the corresponding reference coefficients are compared group by group to determine the parameter combination attribution and obtain the parameter combination attribution determination result.

[0014] In one embodiment of the present invention, the step of matching similar source domain data according to the hierarchical division results of the evaluation index to generate a transfer learning model-adapted data source includes: Based on the hierarchical classification of hydropower station equipment evaluation indicators, the equipment parameters corresponding to each level are extracted, and the parameter information is organized to generate a hierarchical equipment parameter set. Based on the hierarchical equipment parameter set, source domain data of similar hydropower station equipment operation is obtained. The hierarchical equipment parameters are compared and matched with the parameter types in the source domain data one by one, and the matching results are recorded to obtain the parameter type matching results. Based on the parameter type matching results, source domain data is filtered according to preset matching standards, and source domain data that meets the standards is retained to generate transfer learning model adaptation data sources. The transfer learning model adaptation data sources include source domain data for hydro turbine generator sets and source domain data for main transformers.

[0015] In one embodiment of the present invention, the step of combining the transfer learning model to adapt the data source and real-time operating parameters, obtaining the warning thresholds for each level, comparing the real-time correlation coefficient with the warning thresholds for each level, and generating a device status warning judgment result includes: To adapt the transfer learning model to the data source, and combine the real-time parameters in the basic parameter set of the core equipment of the hydropower station, obtain the warning thresholds of each evaluation index level corresponding to the governor, and integrate the warning thresholds of each level to generate a set of warning thresholds for the governor evaluation level. Based on the transfer learning model, the data source and the real-time parameters in the set of basic parameters of the core equipment of the hydropower station are adapted to the data source. The correlation coefficients corresponding to the real-time parameters are extracted and the correlation coefficients are sorted to obtain the set of real-time parameter correlation coefficients. Based on the real-time parameter correlation coefficient set and the governor evaluation level early warning threshold set, the correlation coefficients corresponding to the real-time parameters are aligned and compared with the corresponding early warning thresholds of the governor level one by one to generate the hydropower station equipment status early warning judgment results. The hydropower station equipment status early warning judgment results include the early warning judgment conclusion of the turbine generator set based on the governor threshold, the early warning judgment conclusion of the main transformer, and the early warning matching result of the real-time parameter correlation coefficient.

[0016] In one embodiment of the present invention, the step of comparing the correlation coefficients of the hydro-generator unit with the corresponding reference coefficients and the correlation coefficients of the main transformer with the corresponding reference coefficients, respectively, based on the equipment characteristic operating condition correlation coefficient set and the equipment rated reference coefficient set, to determine the parameter combination attribution and obtain the parameter combination attribution determination result includes: When comparing each group, first match the parameter types, compare the real-time correlation coefficient of the rated power of the hydro-generator unit with the reference coefficient of the rated power, calculate the absolute value of the difference between the two, and set the absolute value of the difference. To match the qualified range, determine which parameter combination belongs to the qualified category; set up absolute value of difference For minor deviations, the real-time correlation coefficient of the rated head of the hydro-generator unit is compared with the reference coefficient of the rated head. If the absolute value of the difference is within this range, it is determined to belong to the minor deviation category. For the main transformer, the real-time correlation coefficients of rated capacity, rated voltage, and rated current are compared with the corresponding benchmark coefficients. The transformer is classified as either qualified or slightly deviated based on the range of the absolute value of the difference. The comparison and classification of all parameter combinations are completed one by one to obtain the classification result of the parameter combination.

[0017] To achieve the above objectives, a second aspect of the present invention provides a hydropower station equipment status assessment and early warning device based on transfer learning, comprising: The parameter processing module is used to acquire the operating parameters of the core equipment of the hydropower station, verify the completeness of the acquisition of the operating parameters, and integrate them to generate a set of basic parameters for the core equipment. The feature acquisition module is used to calculate the correlation between equipment characteristics and operating conditions based on the core equipment basic parameter set and according to preset correlation calculation rules, and generate a feature operating condition correlation coefficient set. The hierarchical division module is used to compare the set of correlation coefficients of the characteristic working conditions with the benchmark coefficients of the rated parameters to establish the hierarchical division results of the evaluation indicators; The model matching module is used to match similar source domain data according to the hierarchical division results of the evaluation index and generate a transfer learning model adapted data source. The status determination module is used to combine the transfer learning model with the data source and real-time operating parameters to obtain the warning thresholds at each level, compare the real-time correlation coefficient with the warning thresholds at each level, and generate the equipment status warning determination result.

[0018] The present invention provides a method and apparatus for assessing and providing early warning of the status of hydropower station equipment based on transfer learning, which has the following beneficial effects: By accurately selecting the core operating parameters of turbines, generators, and main transformers, classifying and organizing them according to equipment type, and verifying their completeness, a structured set of core equipment basic parameters is generated. This effectively avoids data loss or redundancy, ensures the comprehensiveness and reliability of basic data, and provides high-quality data support for subsequent analysis. Based on the basic parameter set, targeted calculations are performed to construct a set of equipment characteristic-operating condition correlation coefficients, which realizes a deep correlation between equipment parameters and operating conditions, making the data more representative and able to accurately reflect the equipment operating status. By combining the baseline coefficients of the equipment's rated parameters and determining the attribute of parameter combinations through group comparison, a hierarchical evaluation index system is established, which makes the evaluation index highly compatible with the actual operating characteristics of the equipment and significantly improves the pertinence of the evaluation. Based on the hierarchical extraction of parameters from the evaluation indicators, matching source domain data of similar hydropower stations and screening suitable data sources, the adaptability of the transfer learning model to the target equipment is enhanced, and the accuracy of model application is effectively improved. By combining real-time parameters with the warning thresholds of each level of the speed governor and comparing them one by one, accurate warning judgment results are generated, which realizes accurate and efficient early warning of equipment status, provides a more reliable basis for equipment maintenance decisions, ensures stable equipment operation, and thus improves the overall production efficiency of the hydropower station.

[0019] To achieve the above objectives, a third aspect of this application provides a computer device, including a processor and a memory; wherein the processor reads executable program code stored in the memory to run a program corresponding to the executable program code, for implementing a method for assessing and warning the status of hydropower station equipment based on transfer learning as described in the first aspect embodiment.

[0020] To achieve the above objectives, the fourth aspect of this application proposes a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a method for assessing and warning the status of hydropower station equipment based on transfer learning as described in the first aspect embodiment.

[0021] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0022] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart of a hydropower station equipment status assessment and early warning method based on transfer learning according to an embodiment of the present invention; Figure 2 This is an architecture diagram of a hydropower station equipment status assessment and early warning method based on transfer learning according to an embodiment of the present invention; Figure 3 This is a structural diagram of a hydropower station equipment status assessment and early warning device based on transfer learning according to an embodiment of the present invention; Figure 4 It is a computer device according to an embodiment of the present invention. Detailed Implementation

[0023] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0025] The following description, with reference to the accompanying drawings, describes a method and apparatus for assessing and providing early warning of the status of hydropower station equipment based on transfer learning, according to an embodiment of the present invention.

[0026] Example 1 Please see Figure 1 and Figure 2 This invention provides a technical solution: a method for assessing and issuing early warnings of the status of hydropower station equipment based on transfer learning, comprising the following steps: S1. Obtain the operating parameters of the core equipment of the hydropower station, verify the completeness of the collected operating parameters, and integrate them to generate a set of basic parameters for the core equipment.

[0027] The turbine's runner diameter, blade reference angle, runner inlet and outlet pressure, and speed fluctuation value are obtained. The generator's stator winding resistance, stator winding temperature, rotor vibration amplitude, and output current stability are also obtained. The parameters of the two types of equipment are classified and organized according to equipment type. The completeness of parameter collection is checked. The parameters after classification and verification are integrated to generate a set of basic parameters for the core equipment of the hydropower station.

[0028] S2, based on the core equipment basic parameter set, calculate the correlation between equipment characteristics and operating conditions according to the preset correlation calculation rules, and generate a characteristic operating condition correlation coefficient set.

[0029] Based on the basic parameter set of the core equipment of the hydropower station, the ratio of the pressure difference between the inlet and outlet of the turbine runner to the blade reference angle is calculated. This ratio is then multiplied by the speed fluctuation value. The ratio of the generator stator winding temperature to the stator winding resistance is calculated. This ratio is then compared with the rotor vibration amplitude. The calculation results of the two types of equipment are summarized to generate a set of equipment characteristic-operating condition correlation coefficients.

[0030] S3, compare the set of characteristic working condition correlation coefficients with the rated parameter benchmark coefficients to establish the evaluation index hierarchy division results.

[0031] The equipment characteristic-operating condition correlation coefficient set is called to obtain the corresponding rated parameter benchmark coefficients for the hydro-generator unit and the main transformer. The characteristic-operating condition correlation coefficient of the hydro-generator unit is compared with the rated benchmark coefficient of the hydro-generator unit group by group, and the characteristic-operating condition correlation coefficient of the main transformer is compared with the rated benchmark coefficient of the main transformer group by group. The parameter combination corresponding to different correlation coefficients is determined, and the hierarchical classification result of hydropower station equipment evaluation index is established based on the determination result.

[0032] S4. Match similar source domain data according to the hierarchical division results of the evaluation indicators to generate a transfer learning model adapted to the data source.

[0033] Based on the hierarchical classification of hydropower station equipment evaluation indicators, the equipment parameters corresponding to each level are extracted, source domain data of similar hydropower station equipment operation are obtained, and the extracted equipment parameters of each level are matched one by one with the parameter types in the source domain data. Source domain data with matching degree that meets the preset standard are selected to generate transfer learning model adaptation data sources.

[0034] S5. Combining the transfer learning model with the adapted data source and real-time operating parameters, obtain the warning thresholds for each level, compare the real-time correlation coefficient with the warning thresholds for each level, and generate the device status warning judgment result.

[0035] To adapt the transfer learning model to the data source, and in conjunction with the real-time parameters in the basic parameter set of the core equipment of the hydropower station, the warning thresholds of each evaluation index level corresponding to the governor are obtained. The correlation coefficients corresponding to the real-time parameters are aligned and compared with the warning thresholds of the corresponding levels of the governor one by one to generate the warning judgment results of the hydropower station equipment status.

[0036] The set of basic parameters for core equipment of a hydropower station includes a subset of core parameters for turbines and a subset of core parameters for generators. The set of equipment characteristic-operating condition correlation coefficients includes a subset of characteristics-operating condition correlation coefficients for turbines and a subset of characteristics-operating condition correlation coefficients for generators. The hierarchical classification results of hydropower station equipment evaluation indicators include the hierarchical evaluation indicators for turbine-generator units and the hierarchical evaluation indicators for main transformers. The data sources for the transfer learning model adaptation include the adaptation source domain data for turbine-generator units and the adaptation source domain data for main transformers. The early warning judgment results for hydropower station equipment status include the early warning judgment conclusions for parameters at each level of the governor and the early warning matching results for real-time parameter correlation coefficients.

[0037] The steps for obtaining the basic parameter set of core equipment of a hydropower station are as follows: S101: Obtain the turbine runner diameter, blade reference angle, runner inlet and outlet pressure, and speed fluctuation value; obtain the generator stator winding resistance, stator winding temperature, rotor vibration amplitude, and output current stability; integrate the original parameters of the two types of equipment to generate a summary of the original parameters of the equipment. S102: Based on the summary of the original parameters of the equipment, the parameters are split according to the type of water turbine and generator, and the order of similar parameters is sorted to obtain the equipment classification parameter set; S103: For the equipment classification parameter set, verify the matching of the parameters that should be collected for each type of equipment with the actual parameters collected, determine the missing parameters, and integrate the verified parameters to generate the basic parameter set of the core equipment of the hydropower station.

[0038] The steps for obtaining the equipment characteristic-operating condition correlation coefficient set are as follows: S201: Based on the basic parameter set of the core equipment of the hydropower station, extract the inlet and outlet pressure of the turbine runner, the blade reference angle, and the speed fluctuation value. Calculate the ratio of the pressure difference between the runner inlet and outlet to the blade reference angle. Multiply this ratio with the speed fluctuation value to generate the associated parameters of the turbine characteristic operating conditions. S202: Based on the basic parameter set of core equipment of hydropower station, extract the generator stator winding temperature, stator winding resistance and rotor vibration amplitude, calculate the ratio of stator winding temperature to stator winding resistance, perform difference calculation on the ratio and rotor vibration amplitude, and generate generator characteristic operating condition related parameters. The generator stator winding temperature, stator winding resistance, and rotor vibration amplitude are measured in real-time using resistance thermometers embedded in the stator slots. The bridge test circuit is invoked to obtain the current DC resistance measurement value of the stator winding. Stator winding temperature As a molecule, and considering the stator winding resistance The temperature resistance ratio is obtained by performing a division operation using the denominator. The combined vibration displacement of the rotor in the X and Y axes was collected using an eddy current sensor, and the peak-to-peak value was extracted as the rotor vibration amplitude data. The temperature resistance ratio With rotor vibration amplitude The difference obtained by performing subtraction is: Call the generator operation standard weighting coefficient Its setting refers to the ratio of generator insulation class to mechanical vibration limit, and the calculation formula is set as follows: in For insulation withstand temperature and Reference resistor The conclusion is The difference results With standard weighting coefficient Perform normalized mapping operation, and set a judgment interval if the operation result is within... If the result is within the range, it is determined that the generator's electromechanical coordination is normal. If the value is within the specified range, it is determined to be electrical heating or mechanical vibration deviation. Each set of dynamic values ​​obtained is labeled and summarized to form the generator's state mapping index under the current operating load and the generator characteristic condition related parameters.

[0039] S203: Based on the correlation parameters of turbine characteristic operating conditions and generator characteristic operating conditions, summarize the calculation results of the two types of equipment and generate a set of equipment characteristic-operating condition correlation coefficients.

[0040] The correlation parameters of turbine characteristic operating conditions and generator characteristic operating conditions are used to extract the product results generated in the turbine calculation process. and pressure angle ratio Simultaneously, extract the difference results generated during the generator calculation process. and temperature resistance ratio The associated values ​​on the turbine side and the associated values ​​on the generator side are paired according to the unit number, and the equipment operation synchronization coefficient is called. Its setting is based on the transmission efficiency of the connection between the turbine and generator shaft systems, taking... The set operation is performed to weight and combine the turbine parameters and generator parameters. The calculation formula is as follows: , Derive the sum coefficient ; The sum coefficient is compared with a preset device coupling determination threshold, and a range is set. A range is set for the high-efficiency energy conversion zone of the unit. For the energy loss increasing region, all intermediate variables, weighting factors and final merged coefficient values ​​involved in the calculation are stored according to equipment components. The dynamic correlation indicators of the turbine unit and the power generation system are logically integrated through database write operations, forming a set of equipment characteristic-operating condition correlation coefficients.

[0041] The steps for obtaining the hierarchical classification results of hydropower station equipment evaluation indicators are as follows: S301: Call the equipment characteristics-operating condition correlation coefficient set, obtain the rated parameter reference coefficients of the hydro-generator set and the rated parameter reference coefficients of the main transformer, and integrate the two types of reference coefficients to generate the equipment rated reference coefficient set; First, extract the associated data of rated parameters of the hydro-generator unit from the coefficient set, clarifying that the types of rated parameter reference coefficients to be obtained are rated power reference coefficient, rated speed reference coefficient, and rated head reference coefficient. Then, extract the associated data of rated parameters of the main transformer, determining that the types of rated parameter reference coefficients to be obtained are rated capacity reference coefficient, rated voltage reference coefficient, and rated current reference coefficient. During the retrieval process, the associated data of the corresponding equipment is accurately matched by the equipment number. For the hydro-generator unit, the rated power reference coefficient is set with reference to the rated power standard value of the corresponding model in GB / T7894-2019 "Technical Conditions for Hydro-generator Units". Taking a 300MW hydro-generator unit, its rated power standard value is 300MW, and the rated power reference coefficient K is set as rated power standard value / 1000, which calculates to K=300 / 1000=0.3; the rated speed reference coefficient K is set with reference to the rated speed of the unit, 500r / min. The formula is K = rated speed / 1000, so K = 500 / 1000 = 0.5; the rated head reference coefficient K is based on the unit's rated head of 100m, and the formula is K = rated head / 200, so K = 100 / 200 = 0.5. This generates the reference coefficient set for the rated parameters of the hydro-generator unit {0.3, 0.5, 0.5}. For the main transformer, the rated capacity reference coefficient K refers to GB / T1094.1-2013 "Power Transformers - Part 1" Part: In the General Provisions, the standard value of the rated capacity of a 220kV / 300MVA transformer is 300MVA. The calculation formula is K = standard value of rated capacity / 1000, which gives K = 300 / 1000 = 0.3. The rated voltage reference coefficient K is based on the rated high voltage side voltage of 220kV, and the calculation formula is K = rated voltage / 500, which gives K = 220 / 500 = 0.44. The rated current reference coefficient K is calculated from the rated capacity and rated voltage, first using the formula... Calculate the rated current, where S is the rated capacity of 300MVA and U is the rated voltage of 220kV, and obtain... Then, setting K=I / 2000, we get K=787.8 / 2000≈0.39, generating the main transformer rated parameter reference coefficient set {0.3,0.44,0.39}. The two reference coefficient sets are sorted and integrated according to equipment type, with the hydro generator set reference coefficient first and the main transformer reference coefficient second, generating the equipment rated reference coefficient set {0.3,0.5,0.5,0.3,0.44,0.39}.

[0042] S302: Based on the set of equipment characteristic-operating condition correlation coefficients and the set of equipment rated reference coefficients, the correlation coefficients of the hydro-generator unit and the corresponding reference coefficients, and the correlation coefficients of the main transformer and the corresponding reference coefficients are compared group by group to determine the parameter combination attribution and obtain the parameter combination attribution determination result. The real-time correlation coefficients of the hydro-generator unit under different operating conditions, such as the real-time correlation coefficient K for rated power, rated speed, and rated head, and the real-time correlation coefficients of the main transformer under different operating conditions, such as the real-time correlation coefficient K for rated capacity, rated voltage, and rated current, are given by the equipment rated reference coefficient set {0.3, 0.5, 0.5, 0.3, 0.44, 0.39}. During the comparison, the parameter types are matched first. The real-time correlation coefficient K of the rated power of the hydro-generator unit is compared with the reference coefficient K of the rated power. Taking K=0.28 under a certain operating condition, the absolute value of the difference between the two is calculated as |0.28-0.3|=0.02. The absolute value of the difference is set to ≤0.05 as the qualified matching range, and 0.02≤0.05 is determined to be in the qualified category. Then, the real-time correlation coefficient K=0.52 of the rated speed of the hydro-generator unit is compared with the reference coefficient K=0.5 of the rated speed. The absolute value of the difference is |0.52-0.5|=0.02, and ≤0.05 is determined to be in the qualified range. 0.05, classified as qualified; next, the real-time correlation coefficient of the rated head of the hydro-generator unit K=0.56 is compared with the reference coefficient of the rated head K=0.5. The absolute value of the difference |0.56-0.5|=0.06. Setting 0.05 < absolute value of difference ≤ 0.1 as a slight deviation category, 0.06 falls within this range and is classified as a slight deviation category; for the main transformer, the real-time correlation coefficient of the rated capacity K=0.35 under a certain operating condition is compared with the reference coefficient of the rated capacity K=0.3. The absolute value of the difference |0.35-0.3|=0.05, ≤ 0. 0.05, classified as qualified; Comparing the real-time correlation coefficient of rated voltage K=0.49 with the reference coefficient of rated voltage K=0.44, the absolute value of the difference |0.49-0.44|=0.05, ≤0.05, classified as qualified; Comparing the real-time correlation coefficient of rated current K=0.45 with the reference coefficient of rated current K=0.39, the absolute value of the difference |0.45-0.39|=0.06, within the range of slight deviation, classified as slight deviation, the comparison and classification of all parameter combinations are completed group by group to obtain the parameter combination classification result.

[0043] S303: Based on the parameter combination attribution determination results, sort out the parameter combinations corresponding to different attributions and establish the hierarchical classification results of hydropower station equipment evaluation indicators; First, we sort out the parameter combinations corresponding to the qualified category, extract the rated power parameter combination (K=0.28, K=0.3) and rated speed parameter combination (K=0.52, K=0.5) of the hydro-generator unit, and the rated capacity parameter combination (K=0.35, K=0.3) and rated voltage parameter combination (K=0.49, K=0.44) of the main transformer. We then classify these parameter combinations into a first-level category according to equipment type and label it as the "Core Rated Parameter Qualified Combination Layer". Next, we sort out the parameter combinations corresponding to the minor deviation category, extract the rated head parameter combination (K=0.56, K=0.5) of the hydro-generator unit and the rated current parameter combination (K=0.45, K=0.39) of the main transformer. We then classify these parameter combinations into a second-level category according to equipment type and label it as the "Minor Rated Parameter Minor Deviation Combination Layer". We clarify the parameter combination composition and corresponding equipment type of the two levels and establish a hierarchical classification result of hydropower station equipment evaluation indicators that includes level number, level name, equipment, and parameter combination list.

[0044] The steps for obtaining data sources to adapt a transfer learning model are as follows: S401: Based on the hierarchical classification of hydropower station equipment evaluation indicators, extract the equipment parameters corresponding to each level, organize the parameter information, and generate a hierarchical equipment parameter set. S402: Based on the hierarchical equipment parameter set, obtain source domain data of similar hydropower station equipment operation, compare and match the hierarchical equipment parameters with the parameter types in the source domain data one by one, record the matching situation and obtain the parameter type matching result; S403: Based on the parameter type matching results, filter source domain data according to the preset matching criteria, and retain source domain data that meets the criteria to generate a transfer learning model adaptation data source.

[0045] Steps for obtaining early warning results of hydropower station equipment status: S501: Adapt the data source to the transfer learning model, combine the real-time parameters in the basic parameter set of the core equipment of the hydropower station, obtain the warning thresholds of each evaluation index level corresponding to the governor, and integrate the warning thresholds of each level to generate the governor evaluation level warning threshold set. S502: Based on the transfer learning model, adapt the data source and the real-time parameters in the basic parameter set of the core equipment of the hydropower station, extract the correlation coefficients corresponding to the real-time parameters, and organize the correlation coefficients to obtain the real-time parameter correlation coefficient set. S503: Based on the real-time parameter correlation coefficient set and the governor evaluation level early warning threshold set, the correlation coefficients corresponding to the real-time parameters are aligned and compared with the corresponding early warning thresholds of the governor level one by one to generate the hydropower station equipment status early warning judgment result.

[0046] The present invention provides a method for assessing and warning the status of hydropower station equipment based on transfer learning, which can ensure the comprehensiveness and reliability of basic data, realize the deep correlation between equipment parameters and operating conditions, improve the adaptability of transfer learning models and the pertinence of assessment indicators, and thus achieve accurate and efficient equipment status warning.

[0047] Example 2 To achieve the above embodiments, such as Figure 3 As shown, this embodiment also provides a hydropower station equipment status assessment and early warning device 10 based on transfer learning. The device 10 includes a parameter processing module 100, a feature acquisition module 200, a hierarchical division module 300, a model matching module 400, and a status determination module 500.

[0048] The parameter processing module 100 is used to acquire the operating parameters of the core equipment of the hydropower station, verify the completeness of the acquisition of the operating parameters, and integrate them to generate a set of basic parameters for the core equipment. The feature acquisition module 200 is used to calculate the correlation between equipment characteristics and operating conditions based on the core equipment basic parameter set and according to preset correlation calculation rules, and generate a feature operating condition correlation coefficient set. The hierarchical division module 300 is used to compare the set of characteristic working condition correlation coefficients with the rated parameter benchmark coefficients to establish the hierarchical division result of the evaluation index; The model matching module 400 is used to match similar source domain data according to the hierarchical division results of the evaluation index and generate a transfer learning model adapted data source. The status determination module 500 is used to combine the transfer learning model with the data source and real-time operating parameters to obtain the warning thresholds at each level, compare the real-time correlation coefficient with the warning thresholds at each level, and generate the equipment status warning determination result.

[0049] Furthermore, the parameter processing module 100 described above is used for: The system obtains the turbine runner diameter, blade reference angle, runner inlet and outlet pressure, and speed fluctuation value; the generator stator winding resistance, stator winding temperature, rotor vibration amplitude, and output current stability; and the main transformer rated capacity, rated voltage, and rated current parameters. It then integrates the original parameters of various equipment to generate a summary of the original parameters of the equipment. Based on the summary of the original parameters of the equipment, the parameters are broken down according to the type of hydro-generator set and main transformer, the order of similar parameters is sorted out, the matching of the parameters that should be collected for each type of equipment with the actual collected parameters is checked, and the status of missing parameters is determined. The parameters that have been verified and confirmed to be correct are integrated to generate a set of basic parameters for the core equipment of the hydropower station. The set of basic parameters for the core equipment of the hydropower station includes a subset of core parameters for the turbine, a subset of core parameters for the generator, and a subset of core parameters for the main transformer.

[0050] Furthermore, the aforementioned feature acquisition module 200 is used for: Based on the core equipment basic parameter set, the turbine runner inlet and outlet pressures, blade reference angles, and speed fluctuation values ​​are extracted. The ratio of the runner inlet and outlet pressure difference to the blade reference angle is calculated. This ratio is then multiplied by the speed fluctuation value to generate turbine characteristic operating condition related parameters. Based on the core equipment basic parameter set, the generator stator winding temperature, stator winding resistance, and rotor vibration amplitude are extracted. The ratio of stator winding temperature to stator winding resistance is calculated. The difference between this ratio and the rotor vibration amplitude is calculated to generate generator characteristic operating condition related parameters. Based on the associated parameters of the turbine characteristic operating conditions and the associated parameters of the generator characteristic operating conditions, the calculation results of the two types of equipment are summarized to generate a set of associated coefficients of equipment characteristic operating conditions. The set of associated coefficients of equipment characteristic operating conditions includes a subset of associated coefficients of turbine generator set characteristic operating conditions and a subset of associated coefficients of main transformer characteristic operating conditions.

[0051] Furthermore, the hierarchy division module 300 is used for: The system retrieves the set of correlation coefficients for equipment characteristic operating conditions, obtains the reference coefficients for rated parameters of the hydro-generator unit and the reference coefficients for rated parameters of the main transformer, and integrates the two types of reference coefficients to generate a set of equipment rated reference coefficients. Based on the set of correlation coefficients for equipment characteristic operating conditions and the set of equipment rated reference coefficients, the correlation coefficients of the hydro-generator unit and the corresponding reference coefficients, and the correlation coefficients of the main transformer and the corresponding reference coefficients are compared group by group to determine the parameter combination attribution and obtain the parameter combination attribution determination result.

[0052] Furthermore, the model matching module 400 is used for: Based on the hierarchical classification of hydropower station equipment evaluation indicators, the equipment parameters corresponding to each level are extracted, and the parameter information is organized to generate a hierarchical equipment parameter set. Based on the hierarchical equipment parameter set, source domain data of similar hydropower station equipment operation is obtained. The hierarchical equipment parameters are compared and matched with the parameter types in the source domain data one by one, and the matching results are recorded to obtain the parameter type matching results. Based on the parameter type matching results, source domain data is filtered according to preset matching standards, and source domain data that meets the standards is retained to generate transfer learning model adaptation data sources. The transfer learning model adaptation data sources include source domain data for hydro turbine generator sets and source domain data for main transformers.

[0053] Furthermore, the state determination module 500 is used for: To adapt the transfer learning model to the data source, and combine the real-time parameters in the basic parameter set of the core equipment of the hydropower station, obtain the warning thresholds of each evaluation index level corresponding to the governor, and integrate the warning thresholds of each level to generate a set of warning thresholds for the governor evaluation level. Based on the transfer learning model, the data source and the real-time parameters in the set of basic parameters of the core equipment of the hydropower station are adapted to the data source. The correlation coefficients corresponding to the real-time parameters are extracted and the correlation coefficients are sorted to obtain the set of real-time parameter correlation coefficients. Based on the real-time parameter correlation coefficient set and the governor evaluation level early warning threshold set, the correlation coefficients corresponding to the real-time parameters are aligned and compared with the corresponding early warning thresholds of the governor level one by one to generate the hydropower station equipment status early warning judgment results. The hydropower station equipment status early warning judgment results include the early warning judgment conclusion of the turbine generator set based on the governor threshold, the early warning judgment conclusion of the main transformer, and the early warning matching result of the real-time parameter correlation coefficient.

[0054] Furthermore, the hierarchy division module 300 is also used for: When comparing each group, first match the parameter types, compare the real-time correlation coefficient of the rated power of the hydro-generator unit with the reference coefficient of the rated power, calculate the absolute value of the difference between the two, and set the absolute value of the difference. To match the qualified range, determine which parameter combination belongs to the qualified category; set up absolute value of difference For minor deviations, the real-time correlation coefficient of the rated head of the hydro-generator unit is compared with the reference coefficient of the rated head. If the absolute value of the difference is within this range, it is determined to belong to the minor deviation category. For the main transformer, the real-time correlation coefficients of rated capacity, rated voltage, and rated current are compared with the corresponding benchmark coefficients. The transformer is classified as either qualified or slightly deviated based on the range of the absolute value of the difference. The comparison and classification of all parameter combinations are completed one by one to obtain the classification result of the parameter combination.

[0055] The present invention discloses a hydropower station equipment status assessment and early warning device based on transfer learning, which can ensure comprehensive and reliable basic data, realize deep correlation between equipment parameters and operating conditions, improve the adaptability of transfer learning models and the pertinence of assessment indicators, and thus achieve accurate and efficient equipment status early warning.

[0056] To implement the methods of the above embodiments, the present invention also provides a computer device, such as... Figure 4 As shown, the computer device 600 includes a memory 601 and a processor 602; wherein, the processor 602 reads executable program code stored in the memory 601 to run a program corresponding to the executable program code, so as to implement the various steps of the method described above.

[0057] To implement the above embodiments, this application also proposes a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in the foregoing embodiments.

[0058] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0059] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for evaluating and early warning of the state of hydropower equipment based on transfer learning, characterized in that, include: Obtain the operating parameters of the core equipment of the hydropower station, verify the completeness of the collected operating parameters, and integrate them to generate a set of basic parameters for the core equipment; Based on the core equipment basic parameter set, the correlation between equipment characteristics and operating conditions is calculated according to the preset correlation calculation rules, and a characteristic operating condition correlation coefficient set is generated. The set of correlation coefficients for the characteristic working conditions is compared with the benchmark coefficients of the rated parameters to establish the hierarchical division results of the evaluation indicators; Based on the hierarchical division results of the evaluation indicators, similar source domain data are matched to generate a transfer learning model that adapts to the data source. By combining the transfer learning model with the data source and real-time operating parameters, the warning thresholds for each level are obtained. The real-time correlation coefficients are compared with the warning thresholds for each level to generate the device status warning judgment result.

2. The method of claim 1, wherein, The process of acquiring the operating parameters of the core equipment of the hydropower station, verifying the completeness of the collected operating parameters, and integrating them to generate a basic parameter set for the core equipment includes: The system obtains the turbine runner diameter, blade reference angle, runner inlet and outlet pressure, and speed fluctuation value; the generator stator winding resistance, stator winding temperature, rotor vibration amplitude, and output current stability; and the main transformer rated capacity, rated voltage, and rated current parameters. It then integrates the original parameters of various equipment to generate a summary of the original parameters of the equipment. Based on the summary of the original parameters of the equipment, the parameters are broken down according to the type of hydro-generator set and main transformer, the order of similar parameters is sorted out, the matching of the parameters that should be collected for each type of equipment with the actual collected parameters is checked, and the status of missing parameters is determined. The parameters that have been verified and confirmed to be correct are integrated to generate a set of basic parameters for the core equipment of the hydropower station. The set of basic parameters for the core equipment of the hydropower station includes a subset of core parameters for the turbine, a subset of core parameters for the generator, and a subset of core parameters for the main transformer.

3. The method of claim 1, wherein, The process involves calculating the correlation between equipment characteristics and operating conditions based on the core equipment's basic parameter set, according to preset correlation calculation rules, to generate a set of characteristic operating condition correlation coefficients, including: Based on the core equipment basic parameter set, the turbine runner inlet and outlet pressures, blade reference angles, and speed fluctuation values ​​are extracted. The ratio of the runner inlet and outlet pressure difference to the blade reference angle is calculated. This ratio is then multiplied by the speed fluctuation value to generate turbine characteristic operating condition related parameters. Based on the core equipment basic parameter set, the generator stator winding temperature, stator winding resistance, and rotor vibration amplitude are extracted. The ratio of stator winding temperature to stator winding resistance is calculated. The difference between this ratio and the rotor vibration amplitude is calculated to generate generator characteristic operating condition related parameters. Based on the associated parameters of the turbine characteristic operating conditions and the associated parameters of the generator characteristic operating conditions, the calculation results of the two types of equipment are summarized to generate a set of associated coefficients of equipment characteristic operating conditions. The set of associated coefficients of equipment characteristic operating conditions includes a subset of associated coefficients of turbine generator set characteristic operating conditions and a subset of associated coefficients of main transformer characteristic operating conditions.

4. The method of claim 1, wherein, The step of comparing the set of correlation coefficients for characteristic operating conditions with the baseline coefficients of rated parameters to establish the hierarchical division results of evaluation indicators includes: Call the set of correlation coefficients for equipment characteristic operating conditions, obtain the reference coefficients for rated parameters of the hydro-generator unit and the reference coefficients for rated parameters of the main transformer, and integrate the two types of reference coefficients to generate a set of rated reference coefficients for the equipment. Based on the set of correlation coefficients for equipment characteristic operating conditions and the set of equipment rated reference coefficients, the correlation coefficients of the hydro-generator unit and the corresponding reference coefficients, and the correlation coefficients of the main transformer and the corresponding reference coefficients are compared group by group to determine the parameter combination attribution and obtain the parameter combination attribution determination result.

5. The method of claim 1, wherein, The step of matching similar source domain data according to the hierarchical division results of the evaluation indicators to generate a data source suitable for the transfer learning model includes: Based on the hierarchical classification of hydropower station equipment evaluation indicators, the equipment parameters corresponding to each level are extracted, and the parameter information is organized to generate a hierarchical equipment parameter set. Based on the hierarchical equipment parameter set, source domain data of similar hydropower station equipment operation is obtained. The hierarchical equipment parameters are compared and matched with the parameter types in the source domain data one by one, and the matching results are recorded to obtain the parameter type matching results. Based on the parameter type matching results, source domain data is filtered according to preset matching standards, and source domain data that meets the standards is retained to generate transfer learning model adaptation data sources. The transfer learning model adaptation data sources include source domain data for hydro turbine generator sets and source domain data for main transformers.

6. The method of claim 1, wherein, The process involves combining the transfer learning model with the data source and real-time operating parameters to obtain warning thresholds for each level, comparing the real-time correlation coefficient with the warning thresholds for each level, and generating a device status warning judgment result, including: To adapt the transfer learning model to the data source, and combine the real-time parameters in the basic parameter set of the core equipment of the hydropower station, obtain the warning thresholds of each evaluation index level corresponding to the governor, and integrate the warning thresholds of each level to generate a set of warning thresholds for the governor evaluation level. Based on the transfer learning model, the data source and the real-time parameters in the set of basic parameters of the core equipment of the hydropower station are adapted to the data source. The correlation coefficients corresponding to the real-time parameters are extracted and the correlation coefficients are sorted to obtain the set of real-time parameter correlation coefficients. Based on the real-time parameter correlation coefficient set and the governor evaluation level early warning threshold set, the correlation coefficients corresponding to the real-time parameters are aligned and compared with the corresponding early warning thresholds of the governor level one by one to generate the hydropower station equipment status early warning judgment results. The hydropower station equipment status early warning judgment results include the early warning judgment conclusion of the turbine generator set based on the governor threshold, the early warning judgment conclusion of the main transformer, and the early warning matching result of the real-time parameter correlation coefficient.

7. The method of claim 4, wherein, The method, based on the set of correlation coefficients for equipment characteristic operating conditions and the set of equipment rated reference coefficients, compares the correlation coefficients of the hydro-generator unit with their corresponding reference coefficients and the correlation coefficients of the main transformer with their corresponding reference coefficients group by group to determine the parameter combination attribution and obtain the parameter combination attribution determination result, including: When comparing groups, the parameter type is matched first, the real-time correlation coefficient of the rated power of the hydro-generator unit is compared with the rated power reference coefficient, the absolute value of the difference between the two is calculated, and the absolute value of the difference is set as a threshold value For the eligible interval, it is determined that the parameter combination belongs to the eligible category. Setting Difference absolute value For the slight deviation category, the real-time correlation coefficient of the rated head of the hydroelectric generating set is compared with the reference coefficient of the rated head. If the absolute value of the difference is in the interval, it is determined to belong to the slight deviation category. For the main transformer, the real-time correlation coefficients of rated capacity, rated voltage, and rated current are compared with the corresponding benchmark coefficients. The transformer is classified as either qualified or slightly deviated based on the range of the absolute value of the difference. The comparison and classification of all parameter combinations are completed one by one to obtain the classification result of the parameter combination.

8. A device for evaluating and warning the state of a hydropower station equipment based on transfer learning, characterized in that, include: The parameter processing module is used to acquire the operating parameters of the core equipment of the hydropower station, verify the completeness of the acquisition of the operating parameters, and integrate them to generate a set of basic parameters for the core equipment. The feature acquisition module is used to calculate the correlation between equipment characteristics and operating conditions based on the core equipment basic parameter set and according to preset correlation calculation rules, and generate a feature operating condition correlation coefficient set. The hierarchical division module is used to compare the set of correlation coefficients of the characteristic working conditions with the benchmark coefficients of the rated parameters to establish the hierarchical division results of the evaluation indicators; The model matching module is used to match similar source domain data based on the hierarchical division results of the evaluation indicators and generate a transfer learning model adapted to the data source. The status determination module is used to combine the transfer learning model with the data source and real-time operating parameters to obtain the warning thresholds at each level, compare the real-time correlation coefficient with the warning thresholds at each level, and generate the equipment status warning determination result.

9. A computer device, characterized in that, Including processor and memory; The processor reads executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the hydropower station equipment status assessment and early warning method based on transfer learning as described in any one of claims 1-7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements a method for assessing and warning the status of hydropower station equipment based on transfer learning as described in any one of claims 1-7.

Images