Marine turboset operation state evaluation method
By screening key indicator parameters and analyzing the severity of faults, the problem of inaccurate evaluation results in existing technologies has been solved, enabling accurate evaluation and safety assurance of steam turbine units.
Patent Information
- Application Number
- CN202510869443.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-26
- Publication Date
- 2025-10-17
Smart Images
Figure CN120804910A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of marine steam turbine unit state evaluation. BACKGROUND
[0002] As a large rotating power machine, steam turbine is widely used in ship power devices, and its stable operation is related to the safety of the whole ship power system, Figure 2 The structure principle diagram of the marine steam turbine unit is given. There is a complex coupling relationship between each device in the steam turbine system, and once a fault occurs, it will affect the overall operation of the system, thereby causing economic losses. Therefore, real-time detection of steam turbine unit abnormalities and real-time monitoring of its health status have become a research focus.
[0003] The early fault detection and health evaluation technology of steam turbine unit mainly evaluates the single state of single device, and does not consider the system perspective research, and lacks analysis and modeling of fault evolution and propagation process. With the improvement of steam turbine manufacturing quality, the timeliness of maintenance is difficult to guarantee the production demand. In recent years, with the progress of sensor and data acquisition technology, as well as the development of data processing, analysis and intelligent algorithm, the data-driven steam turbine unit fault diagnosis and health evaluation technology (Prognostics and Health Management, PHM) has gradually become the research focus of scholars.
[0004] The purpose of PHM is to ensure the reliable operation of the equipment and prevent performance degradation or safety accidents caused by sudden failure. With the development of artificial intelligence and big data technology, fault diagnosis and health management technology driven by various data methods has attracted widespread attention and rapid development, forming a complete technology system including trend prediction, anomaly detection, state evaluation and fault diagnosis. Compared with the traditional maintenance method, fault prediction and health management promote the transformation of equipment maintenance from passive maintenance to regular inspection, and then to active prevention. The research on steam turbine unit fault prediction and health management technology has important strategic significance for improving the stability and reliability of the steam power system, and thus realizing accurate maintenance. In the state evaluation task, the selection and weighting of feature parameters are crucial.
[0005] On the one hand, the existing state evaluation method only considers the deviation degree of the system state (the influence of the index parameter on the whole system state), and does not care whether a fault occurs and the type of the fault, and does not consider the influence difference of different fault types. There are differences in the danger degree of different fault types, and there are also slight or serious differences in the same fault. If only the deviation of the system state is considered, the accuracy and practicality of the evaluation result are low. Therefore, the above problems need to be solved. SUMMARY
[0006] In view of the problem that the existing state evaluation method only considers the offset degree of the system state, resulting in poor accuracy of the evaluation result, the application provides a marine steam turbine unit operation state evaluation method.
[0007] The marine steam turbine unit operation state evaluation method comprises the following steps:
[0008] S1, data preparation: important index parameters are obtained by performing feature screening on historical operation data of the steam turbine unit, and influence factors of various fault types are obtained according to historical operation data corresponding to the N important index parameters;
[0009] S2, the running data corresponding to the N important index parameters of the steam turbine unit collected in real time are preprocessed to obtain the to-be-evaluated data;
[0010] S3, fault diagnosis is performed on the to-be-evaluated data, the fault type is predicted, and meanwhile, state similarity analysis is performed on the to-be-evaluated data and standard fault data corresponding to the predicted fault type, to obtain the fault severity of the fault type;
[0011] S4, the influence factors corresponding to the predicted fault type and the fault severity corresponding to the fault type are calculated to obtain the steam turbine unit operation state evaluation result.
[0012] Preferably, the implementation mode of important index parameters obtained by performing importance screening on the historical operation data of the steam turbine unit in step S1 is as follows:
[0013] S1-11, thermal economic analysis is performed on the steam turbine unit, and M index parameters are preliminarily screened out;
[0014] S1-12, the steam turbine unit is operated under different fault types to obtain historical operation data corresponding to the M index parameters;
[0015] S1-13, parameter importance analysis is performed on all historical operation data corresponding to the M index parameters under all faults by using a parameter importance analysis method, to obtain an importance ranking result of the M index parameters, and each index parameter in the result is ranked from high to low in importance;
[0016] S1-14, the first N index parameters are screened out from the importance ranking result as important index parameters.
[0017] Preferably, the parameter importance analysis method is implemented by using a random forest parameter importance analysis method.
[0018] Preferably, in step S1, the implementation mode of the historical operation data corresponding to the N important index parameters, to obtain the influence factors of various fault types, comprises the following steps:
[0019] S1-21. Apply the entropy weight method to process the historical operating data corresponding to N important indicator parameters to obtain the weight corresponding to each important indicator parameter;
[0020] S1-22. Sum the weights of one or more important indicator parameters corresponding to the equipment affected by each fault type, and use the summed result as the influencing factor of the fault type.
[0021] Preferably, in step S2, preprocessing the operating data corresponding to the N important index parameters of the steam turbine group collected in real time to obtain the data to be evaluated is achieved by:
[0022] After normalizing the operating data corresponding to N important indicator parameters of the steam turbine unit collected in real time, the data to be evaluated are obtained.
[0023] Preferably, in step S3, fault diagnosis of the data to be evaluated is performed using a pre-trained diagnostic model.
[0024] Preferably, the pre-trained diagnostic model is implemented using a CNN neural network.
[0025] Preferably, in step S4, the method for obtaining the steam turbine unit operating status assessment result by performing calculations based on the impact factor corresponding to the predicted fault type and the fault severity corresponding to the fault type is as follows: ;
[0026] is the result of the engine unit operation status assessment, For the The influencing factors of the fault types are For the The fault severity corresponding to each fault type.
[0027] Preferably, the N important indicator parameters include high-pressure cylinder exhaust pressure, high-pressure cylinder exhaust temperature, high-pressure cylinder inlet temperature, high-pressure cylinder inlet pressure, high-pressure cylinder steam flow, low-pressure cylinder exhaust pressure, low-pressure cylinder exhaust temperature, low-pressure cylinder inlet temperature, low-pressure cylinder inlet pressure, low-pressure cylinder steam flow, main condensation area circulating water outlet temperature, main condensation area tube bundle wall temperature, main condensation area circulating water inlet temperature, main condensation area circulating water flow, main condensation area condensate flow, air cooling area inlet gas flow, air cooling area extraction pressure, air cooling area extraction flow, hot well condensate flow, hot well condensate outlet flow, hot well water level and hot well water level.
[0028] The steam turbine unit operating status evaluation device includes a storage device, a processor, and a computer program stored in the storage device and executable on the processor. The processor executes the computer program to implement the marine steam turbine unit operating status evaluation method.
[0029] The beneficial effects of the present application are:
[0030] The present application obtains the influence factor of each fault type according to the historical operation data corresponding to the screened N important index parameters, considers the influence difference of different fault types, introduces into the state assessment, calculates the fault severity of each fault type according to similarity, and combines the influence difference of each fault type and the fault severity of the fault type to realize quantitative evaluation of the steam turbine for different faults, thereby meeting the practical needs of engineering.
[0031] The method can accurately and effectively evaluate and analyze the operation state of the steam turbine unit, guarantee the safe and stable operation of the marine steam turbine unit, thereby meeting the practical needs of engineering. BRIEF DESCRIPTION OF DRAWINGS
[0032] Figure 1 is a flow chart of the marine steam turbine unit operation state evaluation method described in the present application;
[0033] Figure 2 is a schematic diagram of the structure principle of the marine steam turbine unit;
[0034] Figure 3 is a schematic diagram before data normalization;
[0035] Figure 4 is a schematic diagram after data normalization;
[0036] Figure 5 is a schematic diagram of importance sorting of index parameters by machine learning random forest method. DETAILED DESCRIPTION
[0037] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.
[0038] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.
[0039] The present application will be further described below in combination with the drawings and specific embodiments, but not as a limitation of the present application.
[0040] Specific embodiment one, in combination Figure 1 The marine steam turbine unit operation state evaluation method described in the present embodiment includes the following steps:
[0041] S1, data preparation: feature screening of historical operation data of the steam turbine unit to obtain N important index parameters, and according to the historical operation data corresponding to the N important index parameters, the influence factors of each fault type are obtained;
[0042] S2, preprocessing the operation data corresponding to the N important index parameters of the steam turbine unit collected in real time to obtain the evaluation data;
[0043] S3, fault diagnosis is performed on the evaluation data, the fault type is predicted, and at the same time, the evaluation data is analyzed for state similarity with the standard fault data corresponding to the predicted fault type, to obtain the fault severity of the fault type;
[0044] S4, according to the influence factors corresponding to the predicted fault type and the fault severity corresponding to the fault type, the operation state evaluation result of the steam turbine unit is obtained.
[0045] In specific application, the fault diagnosis of the evaluation data is realized by using a pre-trained diagnosis model, and the pre-trained diagnosis model can be obtained by using the existing technology. The diagnosis model is pre-trained in advance. As an example, the pre-trained diagnosis model is realized by using a CNN neural network.
[0046] The application obtains the influence factors of each fault type according to the historical operation data corresponding to the N important index parameters screened out, considers the influence difference of different fault types, introduces it into the state evaluation, calculates the fault severity of each fault type according to the similarity, and realizes the quantitative evaluation of the steam turbine for different faults by combining the influence difference of each fault type with the fault severity of the fault type, so as to meet the practical needs of engineering.
[0047] Further, the implementation of step S1, important screening of the historical operation data of the steam turbine unit to obtain N important index parameters, is as follows:
[0048] S1-11, thermal economic analysis is performed on the steam turbine unit, and M index parameters are preliminarily screened out;
[0049] S1-12, the steam turbine unit is operated under different fault types to obtain historical operation data corresponding to the M index parameters;
[0050] S1-13, parameter importance analysis method is used to analyze the importance of all historical operation data corresponding to the M index parameters under all faults, to obtain the importance ranking result of the M index parameters, and the importance of each index parameter in the result is ranked from high to low. In specific application, the parameter importance analysis method is realized by using the random forest parameter importance analysis method. Referring to Figure 5 , an important ranking diagram of index parameters by using the random forest method is given.
[0051] S1-14, screening the first N index parameters from the importance ranking result as important index parameters.
[0052] In the preferred embodiment, to ensure that the selected parameters can comprehensively and accurately reflect the true state of the system or object, the selection of important index parameters needs to consider relevant standards such as the characteristics of the evaluation object and the evaluation purpose; the index parameter screening is performed using the thermoeconomic analysis and parameter importance analysis method, which reflects the relative importance of different indexes in the overall evaluation and the contribution to the system, ensuring the comprehensiveness of the system evaluation.
[0053] Further, in step S1, the historical operation data corresponding to the N important index parameters include the following implementation manners of obtaining the influence factor of each fault type:
[0054] S1-21, applying the entropy weight method to process the historical operation data corresponding to the N important index parameters to obtain the weight corresponding to each important index parameter;
[0055] S1-22, summing the weights of one or more important index parameters corresponding to the equipment affected by each fault type, and taking the summation result as the influence factor of the fault type.
[0056] Further, referring to Figure 3 and Figure 4 , in step S2, the implementation manner of pre-processing the operation data corresponding to the N important index parameters of the real-time collected steam turbine unit to obtain the to-be-evaluated data is as follows:
[0057] After normalizing the operation data corresponding to the N important index parameters of the real-time collected steam turbine unit, the to-be-evaluated data is obtained.
[0058] Figure 3 and Figure 4 are schematic diagrams before and after data normalization; the normalization process does not change the basic structure and change trend of the data, but only changes the dimension. Through mathematical transformation, the feature data of different dimensions and different ranges are converted into a unified standard range, so that the data obtained eliminates the influence of dimension difference on the model, and improves the convergence speed and accuracy of the algorithm. As shown in Figure 4 , the numerical range is controlled between 0 and 1 through normalization.
[0059] Further, in step S4, the implementation manner of obtaining the steam turbine unit operation state evaluation result by performing operation according to the influence factor corresponding to the predicted fault type and the fault severity corresponding to the fault type is as follows: ;
[0060] is the steam turbine unit operation state evaluation result, an influence factor of a first type of fault, a fault severity corresponding to a first type of fault.
[0061] Further, the N important indicator parameters include high-pressure cylinder exhaust pressure, high-pressure cylinder exhaust temperature, high-pressure cylinder inlet temperature, high-pressure cylinder inlet pressure, high-pressure cylinder steam flow, low-pressure cylinder exhaust pressure, low-pressure cylinder exhaust temperature, low-pressure cylinder inlet temperature, low-pressure cylinder inlet pressure, low-pressure cylinder steam flow, main condenser area circulating water outlet temperature, main condenser area tube bundle wall temperature, main condenser area circulating water inlet temperature, main condenser area circulating water flow, main condenser area condensate flow, air cooling area inlet gas flow, air cooling area extraction pressure, air cooling area extraction flow, hot well condensate flow, hot well condensate outlet flow, hot well water level, and hot well water level.
[0062] In a second embodiment, the steam turbine unit operation state evaluation device includes a storage device, a processor, and a computer program stored in the storage device and executable on the processor, and the processor executes the computer program to implement the steam turbine unit operation state evaluation method according to any one of claims 1 to 9.
[0063] While the application has been described with reference to particular embodiments, it will be understood that the examples are merely for illustration and that many modifications can be made by persons of ordinary skill in the art. Other arrangements can be devised without departing from the spirit or scope of the application as defined by the appended claims. It will be understood that where the application is referred to as comprising particular features, embodiments of the application can consist of at least one of those features. It will also be understood that features described as being for one embodiment can be incorporated into another embodiment. It will be further understood that the features described for the individual embodiments can be used in other embodiments.
Claims
1. A method for evaluating the operating status of a marine steam turbine unit, characterized in that: The steps include: S1. Data preparation: Feature screening is performed on the historical operating data of the steam turbine unit to obtain N important indicator parameters. Based on the historical operating data corresponding to the N important indicator parameters, the influencing factors of each fault type are obtained; S2. Preprocessing the operating data corresponding to N important indicator parameters of the steam turbine unit collected in real time to obtain data to be evaluated; S3. Perform fault diagnosis on the data to be evaluated and predict the fault type. At the same time, perform state similarity analysis on the data to be evaluated and the standard fault data corresponding to the predicted fault type to obtain the fault severity of the fault type. S4. Calculate the impact factor corresponding to the predicted fault type and the fault severity corresponding to the fault type to obtain an evaluation result of the steam turbine unit operation status.
2. The method for evaluating the operating status of a marine steam turbine unit according to claim 1, wherein: In step S1, the implementation method of screening the historical operating data of the steam turbine unit according to importance to obtain N important index parameters is as follows: S1-11. Conduct thermoeconomic analysis on the steam turbine unit and preliminarily select M index parameters; S1-12, operating the steam turbine unit under different fault types to obtain historical operating data corresponding to M indicator parameters; S1-13. Perform parameter importance analysis on all historical operating data corresponding to the M indicator parameters under all faults using a parameter importance analysis method to obtain importance ranking results of the M indicator parameters, in which the importance of each indicator parameter is ranked from high to low; S1-14. Filter out the top N indicator parameters from the importance ranking results as important indicator parameters.
3. The method for evaluating the operating status of a marine steam turbine unit according to claim 2, wherein: The parameter importance analysis method is implemented using the random forest parameter importance analysis method.
4. The method for evaluating the operating status of a marine steam turbine unit according to claim 1, wherein: In step S1, the historical operating data corresponding to N important indicator parameters are used to obtain the influencing factors of various fault types in the following ways: S1-21. Apply the entropy weight method to process the historical operating data corresponding to N important indicator parameters to obtain the weight corresponding to each important indicator parameter; S1-22. Sum the weights of one or more important indicator parameters corresponding to the equipment affected by each fault type, and use the summed result as the influencing factor of the fault type.
5. The method for evaluating the operating status of a marine steam turbine unit according to claim 1, wherein: In step S2, the operation data corresponding to the N important index parameters of the steam turbine group collected in real time are preprocessed to obtain the data to be evaluated in the following manner: After normalizing the operating data corresponding to N important index parameters of the steam turbine unit collected in real time, the data to be evaluated are obtained.
6. The method for evaluating the operating status of a marine steam turbine unit according to claim 1, wherein: In step S3, fault diagnosis of the data to be evaluated is performed using a pre-trained diagnosis model.
7. The method for evaluating the operating status of a marine steam turbine unit according to claim 6, wherein: The pre-trained diagnostic model is implemented using a CNN neural network.
8. The method for evaluating the operating status of a marine steam turbine unit according to claim 1, wherein: In step S4, the impact factor corresponding to the predicted fault type and the fault severity corresponding to the fault type are calculated to obtain the steam turbine unit operation status assessment result in the following manner: ; is the result of the engine unit operation status assessment, For the The influencing factors of the fault types are For the The fault severity corresponding to each fault type.
9. The method for evaluating the operating status of a marine steam turbine unit according to claim 1, wherein: N important indicator parameters include high-pressure cylinder exhaust pressure, high-pressure cylinder exhaust temperature, high-pressure cylinder inlet temperature, high-pressure cylinder inlet pressure, high-pressure cylinder steam flow, low-pressure cylinder exhaust pressure, low-pressure cylinder exhaust temperature, low-pressure cylinder inlet temperature, low-pressure cylinder inlet pressure, low-pressure cylinder steam flow, main condensing area circulating water outlet temperature, main condensing area tube bundle wall temperature, main condensing area circulating water inlet temperature, main condensing area circulating water flow, main condensing area condensate flow, air cooling area inlet gas flow, air cooling area extraction pressure, air cooling area extraction flow, hot well condensate flow, hot well condensate outlet flow, hot well water level and hot well water level.
10. A steam turbine unit operating status evaluation device, comprising a storage device, a processor, and a computer program stored in the storage device and executable on the processor, characterized in that: The processor executes the computer program to implement the method for evaluating the operating status of a marine steam turbine unit according to any one of claims 1 to 9.