Modeling method of intelligent monitoring model of thermal power equipment and related device
By utilizing historical data and calibration models from mature thermal power systems, the problem of insufficient data for newly built thermal power systems has been solved, enabling efficient construction of intelligent monitoring models and fault early warning.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HANGZHOU HOLLYSYS AUTOMATION
- Filing Date
- 2026-03-03
- Publication Date
- 2026-06-05
AI Technical Summary
The lack of sufficient historical operating data in newly built thermal power systems leads to long modeling time or high modeling costs for intelligent monitoring models, making it difficult to establish accurate fault prediction models.
By utilizing the rich historical data of mature thermal power systems, a predictive model for newly built thermal power systems is established. The model is then calibrated to learn the differences between similar equipment, including external and internal influencing factors, and the model is corrected to build an intelligent monitoring model.
It shortens modeling time, reduces modeling costs, improves the accuracy of fault early warning, and provides a reliable monitoring mechanism.
Smart Images
Figure CN122154451A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of thermal power equipment fault technology, and more specifically, to a modeling method and related apparatus for an intelligent monitoring panel model of thermal power equipment. Background Technology
[0002] With the deepening of the digital and intelligent transformation of the power industry, intelligent monitoring technology has become a core means to ensure the safe, stable and efficient operation of large-scale power generation systems such as thermal power and nuclear power.
[0003] Currently, building high-performance intelligent monitoring models typically relies on sufficient historical data covering the entire operating cycle and including various operating conditions. However, for newly built thermal power systems (i.e., newly built thermal power generation systems), due to the limited time for commissioning and operation, it is difficult to obtain sufficient historical operating condition data. This data sparsity leads to a decline in the model's generalization ability, making it difficult to establish accurate fault prediction models. Therefore, for newly built thermal power systems, it is usually necessary to collect data over several years or continuously add data to update the model, resulting in a long modeling time span or high modeling costs. Summary of the Invention
[0004] Based on the defects and shortcomings of the existing technology, this application proposes a modeling method and related device for intelligent monitoring models of thermal power equipment, which can solve the problems of long modeling time or high modeling cost of intelligent monitoring models of newly built thermal power generation systems in the existing technology.
[0005] According to a first aspect of this application, a modeling method for an intelligent monitoring panel model of thermal power equipment is provided, the method comprising: Based on the first historical data of parameter measurement points of the first power generation equipment in the first thermal power system under normal operating conditions, a prediction model for the second power generation equipment of the same type in the second thermal power system is established; wherein, the start-up time of the second thermal power system is shorter than that of the first thermal power system; the prediction model is used to predict the parameter values of parameter measurement points at different times under normal operating conditions of the second power generation equipment. The historical observation values of the parameter measurement points during the operation of the second power generation equipment are input into the prediction model to obtain the model prediction value; The model predictions and target impact parameters are used as input data, and the historical observations corresponding to the model predictions are used as the true values to train the calibration model; wherein, the target impact parameters include at least one of the following: external impact parameters, internal impact parameters, design parameters of the second power generation equipment, and generator output power in the second thermal power system; the calibration model is used to correct the predicted values output by the prediction model; Based on the prediction model and the trained calibration model, an intelligent monitoring model for the second power generation equipment is established.
[0006] According to a second aspect of this application, a modeling apparatus for an intelligent monitoring panel model of thermal power equipment is provided, the apparatus comprising: The first model building module is used to build a prediction model for a second power generation device of the same type in a second thermal power system based on the first historical data of parameter measurement points under normal operating conditions of the first power generation device in the first thermal power system; wherein, the start-up time of the second thermal power system is shorter than that of the first thermal power system; the prediction model is used to predict the parameter values of parameter measurement points at different times under normal operating conditions of the second power generation device; The first prediction module is used to input the historical observation values of the parameter measurement points during the operation of the second power generation equipment into the prediction model to obtain the model prediction value; The training module is used to train the calibration model by taking the model's predicted values and target influence parameters as input data and the historical observation values corresponding to the model's predicted values as the true values; wherein, the target influence parameters include at least one of the following: external influence parameters, internal influence parameters, design parameters of the second power generation equipment, and generator output power in the second thermal power system; the calibration model is used to correct the predicted values output by the prediction model; The second model building module is used to build an intelligent monitoring model of the second power generation equipment based on the prediction model and the trained calibration model.
[0007] According to a third aspect of this application, an electronic device is provided, comprising: a memory and a processor; The memory is connected to the processor and is used to store programs; The processor is used to implement the modeling method for the intelligent monitoring model of thermal power equipment as described in the first aspect by running the program in the memory.
[0008] According to a fourth aspect of this application, a storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the modeling method for the intelligent monitoring model of thermal power equipment as described in the first aspect.
[0009] According to a fifth aspect of this application, a computer program product or computer program is provided, the computer program product including the computer program, wherein a processor executing the computer program implements the steps in the modeling method for an intelligent monitoring model of thermal power equipment as described in the first aspect. Optionally, the computer program may be stored in a readable storage medium of a computer device or in the cloud; the processor of the computer device reads the computer program from the readable storage medium or the cloud.
[0010] In the technical solution provided in this application, a predictive model for a second power generation device of the same type in a newly built thermal power system can be constructed based on historical data of parameter measurement points of the first power generation device in a mature thermal power system under normal operating conditions. Then, the historical observations of the second power generation device are input into the predictive model to obtain the model's predicted values. Next, the model's predicted values and model error influence parameters (i.e., target influence parameters) are used as input data, and the historical observations corresponding to the model's predicted values are used as the true values to train the calibration model. Finally, based on the predictive model and the trained calibration model, an intelligent monitoring model for the second power generation device is constructed. Because the calibration model learns the differences between the second power generation device and the first power generation device of the same type during the training process—including differences caused by external influencing factors and differences caused by internal influencing factors—the calibration model can perform more precise correction processing, thereby obtaining more accurate parameter estimates. This allows for monitoring of equipment operating status based on more accurate parameter estimates, improving the accuracy of fault early warning, without the need to continuously add data to update the model, thus reducing modeling costs. This solution utilizes abundant historical data from mature thermal power systems and a small amount of real operational data from newly built thermal power systems to establish an intelligent monitoring model for the new thermal power system. This can shorten the modeling time and provide a reliable monitoring mechanism for the new thermal power system in a timely manner. Attached Figure Description
[0011] Figure 1 This is one of the flowcharts illustrating a modeling method for an intelligent monitoring panel model of thermal power equipment provided in an embodiment of this application.
[0012] Figure 2 This is the second flowchart illustrating the modeling method for the intelligent monitoring panel model of thermal power equipment provided in this application embodiment.
[0013] Figure 3 A scatter plot comparing historical normal data and memory matrix data of turbine bearing vibration measurement points provided in this application embodiment.
[0014] Figure 4 A scatter plot comparing historical normal data and memory matrix data of the turbine thrust bearing metal temperature measurement point provided in the embodiments of this application.
[0015] Figure 5 A scatter plot comparing historical normal data and memory matrix data of turbine bearing metal temperature measuring points provided in this application embodiment.
[0016] Figure 6 A scatter plot comparing historical normal data and memory matrix data of the steam turbine extraction pressure measurement point provided in the embodiments of this application.
[0017] Figure 7This is one of the comparison diagrams of model prediction results for turbine bearing vibration measurement points provided in the embodiments of this application.
[0018] Figure 8 This is one of the comparison diagrams of model prediction results for the metal temperature measuring point of the turbine thrust bearing provided in the embodiments of this application.
[0019] Figure 9 This is one of the comparison charts of model prediction results for turbine bearing metal temperature measuring points provided in the embodiments of this application.
[0020] Figure 10 This is one of the comparison diagrams of model prediction results for the steam turbine extraction pressure measurement points provided in the embodiments of this application.
[0021] Figure 11 The second comparison diagram shows the model prediction results of the turbine bearing vibration measurement points provided in the embodiments of this application.
[0022] Figure 12 The second comparison diagram shows the model prediction results of the turbine thrust bearing metal temperature measurement point provided in the embodiments of this application.
[0023] Figure 13 The second comparison diagram shows the model prediction results of the turbine bearing metal temperature measuring point provided in the embodiments of this application.
[0024] Figure 14 The second comparison diagram shows the model prediction results of the steam turbine extraction pressure measurement points provided in the embodiments of this application.
[0025] Figure 15 A block diagram of a modeling device for an intelligent monitoring panel model of thermal power equipment provided in an embodiment of this application.
[0026] Figure 16 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation
[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0028] Application Overview Currently, for newly built thermal power systems, due to the lack of long-term operational accumulation, the available historical data often only covers a portion of the design operating conditions. The data distribution density in both time and operating condition dimensions is far lower than that of mature thermal power systems. For example, the first-year operating data of some newly built thermal power systems only covers 63% of the design operating conditions, and the distribution density of key parameters is less than 1 / 5 of that of mature thermal power systems. This data sparsity leads to problems such as reduced generalization ability, large estimation errors, and low early warning accuracy in models, making it difficult to meet the accuracy and reliability requirements of engineering applications. To solve this problem, current technical solutions typically require waiting for newly built thermal power systems to accumulate operating data for several years before modeling. However, this results in a serious lag in the construction of intelligent monitoring models, leaving the power generation system without effective intelligent early warning protection in the early stages after commissioning, increasing operational risks and maintenance costs. Current technical solutions can also improve model prediction accuracy by continuously adding data to update the model, but frequent model updates lead to high modeling costs.
[0029] To address this, this application provides a solution that utilizes abundant historical data from mature thermal power systems and limited real-world operational data from newly built thermal power systems to establish a predictive model for the latter. This shortens modeling time and provides a reliable monitoring mechanism for the new thermal power systems. Furthermore, to improve prediction accuracy, a calibration model is added. This calibration model learns from differences between different types of power generation equipment, including differences caused by external and internal influencing factors. This allows the calibration model to more accurately correct the output of the predictive model, resulting in more accurate parameter estimates. These more accurate parameter estimates enable equipment operation status monitoring, improving the accuracy of fault warnings without constantly adding data to update the model, thus reducing modeling costs.
[0030] Exemplary methods This application provides a modeling method for an intelligent monitoring model of thermal power equipment, which is applied to electronic devices. The electronic devices can be terminal devices, such as laptops, desktop computers, tablets, etc.; the electronic devices can also be servers, such as cloud servers, etc.
[0031] The method is described in detail below through some embodiments. The following embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.
[0032] like Figure 1 As shown, the method may include steps 101 to 104, as described below: Step 101: Based on the first historical data of the parameter measurement points of the first power generation equipment in the first thermal power system under normal operating conditions, establish a prediction model for the second power generation equipment of the same type in the second thermal power system.
[0033] The first and second thermal power systems are different thermal power generation systems of the same type, with the second system having been in operation for a shorter period than the first. Specifically, the first thermal power system can be a mature thermal power system, while the second can be a newly built thermal power system. A mature thermal power system can refer to a power generation system that has been operating stably for a sufficiently long time, accumulating massive amounts of historical data including various operating conditions. This historical data meets the basic requirements for constructing high-precision data-driven models (such as intelligent monitoring models) in terms of time span, operating condition coverage, and parameter distribution density.
[0034] The first power generation equipment can be any equipment used for power generation in the first thermal power system. For example, the first power generation equipment can be a boiler, steam turbine, generator, or blower. The second power generation equipment is a power generation equipment of the same type as the first power generation equipment in the second thermal power system. For example, the first power generation equipment is a steam turbine in the first thermal power system, and the second power generation equipment is a steam turbine in the second thermal power system.
[0035] To monitor the operating status of a power generation system, corresponding parameter measurement points can be set for different power generation devices, such as temperature, pressure, flow rate, speed, and current. These parameter measurement points enable monitoring of the equipment's operating status. The specific parameter measurement points can differ for different power generation devices, and can be set according to actual needs.
[0036] In this embodiment, since the first power generation equipment and the second power generation equipment are of the same type, a predictive model for the second power generation equipment can be constructed using historical operating condition data (i.e., first historical data) of the first power generation equipment under normal operating conditions (also known as healthy conditions) in a mature thermal power system. This predictive model is used to predict the parameter values of parameter measurement points of the second power generation equipment under normal operating conditions at different times.
[0037] Optionally, the first historical data of the first generating equipment may include at least one year of historical operating condition data of the first generating equipment. This historical data can be collected by the distributed control system (DCS) of the first thermal power system, and may include, but is not limited to: observed values of parameter measuring points, generator output power (i.e., generator load), etc.
[0038] Step 102: Input the historical observation values of the parameter measurement points during the operation of the second power generation equipment into the prediction model to obtain the model prediction values.
[0039] After obtaining the prediction model of the second power generation equipment, the historical observation values of the parameter measurement points during the operation of the second power generation equipment can be obtained and input into the prediction model. The prediction model then predicts the parameter estimates (i.e., model prediction values) of the same parameter measurement points at the same time under normal operating conditions of the equipment.
[0040] The historical observations mentioned here are a small amount of data acquired over a short period, such as one month's worth of historical observations, compared to the first historical data from the first power generation equipment. These historical observations can be collected through the distributed control system of the second thermal power system.
[0041] Step 103: Use the model predictions and target influence parameters as input data, and the historical observations corresponding to the model predictions as the true values, to train the calibration model.
[0042] Even for the same type of power generation equipment, the equipment itself and the environment in which it is located will be different. For example, the temperature sensing capability of a newly built thermal power system may be more sensitive than that of a mature thermal power system. The temperature and humidity of the environment in which a newly built thermal power system is located will also be different from those of a mature thermal power system. These differences will lead to a large error in the prediction model built using the historical operating data of the first power generation equipment when predicting the parameter values of the parameter measurement points of the second power generation equipment.
[0043] To overcome the above problems, this application embodiment also adds a calibration model to correct the predicted values output by the prediction model.
[0044] To improve the calibration accuracy of the calibration model, it is necessary to train the calibration model first. Considering the causes of model errors, for this calibration model, the model prediction value and the model error influence parameter (i.e., the target influence parameter) can be used as input data, and the historical observation value corresponding to the model prediction value can be used as the true value to train the calibration model. This allows the calibration model to learn the differences between the second power generation equipment and the second power generation equipment, and then calibrate the model prediction value based on these differences.
[0045] The target influencing parameters mentioned here include external influencing parameters of the second power generation equipment (such as ambient temperature, ambient humidity, ambient pressure, etc.), internal influencing parameters, design parameters, and boundary parameters affecting the equipment state (i.e., generator load). In this embodiment, the causes of prediction model errors are considered from multiple aspects, including external, internal, boundary parameters, and design parameters. This allows the calibration model to not only learn the nonlinear characteristics between the parameters of the first and second power generation equipment, but also the delay, inertia, and other characteristics between the influencing factors of model errors and the model parameters, further improving the calibration accuracy of the calibration model. Some influencing parameters cannot be measured; in such cases, design parameters can be used.
[0046] Step 104: Based on the prediction model and the trained calibration model, establish an intelligent monitoring model for the second power generation equipment.
[0047] In this embodiment, the prediction model and the trained calibration model can be combined to form an intelligent monitoring model for the second power generation equipment, so as to monitor the operating status of the second power generation equipment in real time and accurately.
[0048] This application's embodiments utilize abundant historical data from mature thermal power systems and a limited amount of real operational data from newly built thermal power systems to establish an intelligent monitoring model for the latter. While fully leveraging historical data from mature thermal power systems, it also reduces reliance on operational data from newly built systems, shortening modeling time and providing a reliable monitoring mechanism for them. Furthermore, the addition of calibration models and target influence parameters allows for more precise correction of model predictions, resulting in more accurate parameter estimates. These estimates can then be used to monitor equipment operating status, improving the accuracy of fault warnings without constantly adding data to update the model, thus reducing modeling costs.
[0049] It should be noted that, in this embodiment, the newly constructed thermal power system can be divided according to the equipment, and intelligent monitoring models can be constructed for different equipment, with the parameter measurement points corresponding to each model determined. The model input parameters should include the analog measurement points monitored by the distributed control system of the power generation system. Furthermore, the technical solution provided in this embodiment is not only applicable to the construction of intelligent monitoring models for the first thermal power system, but also applicable to thermal power generation systems, wind power generation systems, hydropower generation systems, nuclear power generation systems, or photovoltaic power generation systems, etc.
[0050] After establishing the intelligent monitoring model of the second power generation equipment, the method may further include steps A1 and A3, as described below: Step A1: Input the current observation values of the parameter measurement points of the second power generation equipment into the intelligent monitoring model.
[0051] Step A2: Output the predicted parameter values corresponding to the current observation values through the prediction model in the intelligent monitoring model, and input the predicted parameter values and target influencing factors into the calibration model to obtain the calibration parameter values.
[0052] Step A3: Determine whether to issue a fault warning based on the difference between the current observation value and the calibration parameter value.
[0053] For the constructed intelligent monitoring model, the observed values of parameter measurement points collected in real time or at regular intervals from the second power generation equipment can be input into the intelligent monitoring model. The prediction model in the intelligent monitoring model predicts the parameter values under normal operating conditions at the same time. Then, the predicted parameter values output by the prediction model are input into the calibration model, which corrects the predicted parameter values to obtain the calibration parameter values. Finally, based on the difference between the current observed value and the calibration parameter value, it is determined whether to issue a fault warning. For example, if the absolute value of the difference between the current observed value and the calibration parameter value is greater than or equal to a preset difference threshold, it indicates that the second power generation equipment may have an abnormality, and a fault warning can be issued; if the absolute value of the difference between all current observed values and calibration parameter values is less than the preset difference threshold, it indicates that the second power generation equipment is operating normally and no intervention is required. This achieves intelligent monitoring of the second power generation equipment.
[0054] It should be noted that the electronic device that implements intelligent monitoring modeling can be the same electronic device as the electronic device on which the intelligent monitoring model is installed, or it can be a different electronic device.
[0055] In some alternative embodiments, before constructing a predictive model for the second power generation device using the first historical data of the first power generation device, some data preprocessing operations are required, such as data omission and filling.
[0056] The parameter measurement points planned for newly built thermal power systems may differ from those for mature thermal power systems. For example, some parameter measurement points may be added or some may be removed compared to mature thermal power systems.
[0057] For the reduced parameter measurement points, the data of the corresponding parameter measurement points can be removed from the first historical data of the first power generation equipment.
[0058] For the added parameter measurement points, the data can be supplemented in the following way.
[0059] If the parameter measuring points of the first power generation equipment do not include some parameter measuring points of the second power generation equipment (hereinafter referred to as the first parameter measuring points), the method may further include steps B1 to B3, as described below: Step B1: Obtain the observed values of the first parameter measuring point of the second power generation equipment and the corresponding generator output power.
[0060] Step B2: Fit the observed values of the first parameter measurement point with the corresponding generator output power to obtain the functional relationship between the two.
[0061] Step B3: Based on the obtained functional relationship and the generator output power in the first historical data, obtain the parameter value of the first parameter measurement point.
[0062] In this embodiment, if the first historical data of the first power generation equipment lacks parameter measurement points planned for a second power generation equipment of the same type in a newly built thermal power system, the parameter values of a small number of first parameter measurement points and the corresponding generator output power of the second power generation equipment can be obtained. Then, the parameter values of the first parameter measurement points and the corresponding generator output power are fitted to obtain a functional relationship between them. Finally, the generator output power from the first historical data of the first power generation equipment is substituted into this functional relationship to obtain the parameter values of the first parameter measurement points, thereby completing the data. Afterwards, a prediction model for the second power generation equipment can be established using the first historical data and the completed data.
[0063] In some alternative embodiments, data preprocessing may include not only data filling and missing data, but also the removal of abnormal data, as described below.
[0064] Step 101: Based on the first historical data of parameter measurement points under normal operating conditions of the first power generation equipment in the first thermal power system, establish a prediction model for the second power generation equipment of the same type in the second thermal power system. This may include steps C1 and C2, as described below: Step C1: Based on the generator output power of the first thermal power system and the upper and lower limits of the parameter measurement points of the first power generation equipment, remove abnormal data from the historical data to obtain the second historical data.
[0065] Step C2: Based on the second historical data, establish a prediction model for a second type of power generation equipment.
[0066] The initial historical data of the primary power generation equipment may include outliers. When building a predictive model, it is necessary to remove these outliers to prevent them from affecting the model's accuracy. Outlier removal can be performed based on pre-set upper and lower limits for the unit load (i.e., generator output power) or pre-set upper and lower limits for parameter measurement points.
[0067] It should be noted that some parameters can be set with both upper and lower limits, while others can be set with only an upper or lower limit. The specific settings can be configured according to actual needs.
[0068] It should also be noted that you can first check for missing data and fill in any gaps, and then clean up any abnormal data.
[0069] In some alternative embodiments, a predictive model for the second power generation device can be constructed using the modeling method of the Multivariate State Estimation Technique (MSET) algorithm, as described below.
[0070] Multivariate state estimation algorithms utilize historical data from normal equipment operation to construct a memory matrix (i.e., a prediction model), which is then used to estimate the health values of current operating parameters. This algorithm has relaxed assumptions regarding data distribution, requiring only a physical correlation between parameters. Furthermore, the modeling process is simple, requiring no complex training; the memory matrix serving as the prediction model can be constructed solely from normal operating data.
[0071] Assuming the first power generation equipment is in The observation values of the n parameter measurement points at time n constitute the following observation vector: (1) like Figure 2 As shown, after performing data preprocessing operations such as anomaly cleaning on the historical data of the first power generation equipment, m observation vectors under normal operating conditions can be extracted from the remaining historical data using the equal-interval sampling method to construct the process memory matrix D: (2) Where m is the number of states and n is the number of parameter measurement points.
[0072] For real-time observation vectors MSET can estimate the value of the memory matrix D under normal operating conditions through a linear combination of the memory matrices D. : (3) in, The weight vector reflects The degree of similarity with each state in the memory matrix.
[0073] The weights can be solved using the least squares method, i.e., minimizing the residuals. We can obtain: (4) Substituting formula (4) into formula (3), we get: (5) To avoid To address potential singularity issues, the nonlinear operator ⊗ can be introduced to replace matrix multiplication. Define two vectors... and The Euclidean distance between them is: (6) The operator outputs close to zero when the vectors are similar, and increases when the difference increases. Substituting this into the estimation formula (5), we can obtain the improved MSET estimation expression: (7) By inputting the real-time observation values collected by the second power generation equipment into formula (7), the parameter estimates of the second power generation equipment under normal operating conditions at the same time can be output.
[0074] In some alternative embodiments, the calibration model can be an LSTM-Attention model.
[0075] Accordingly, step 103: using the model predictions and target influence parameters as input data, and the historical observations corresponding to the model predictions as the true values, to train the calibration model, may include: The model is trained by using the model's predicted values and the target influence parameters as input data and the historical observations corresponding to the model's predicted values as output data.
[0076] During the training process, the calibration model can not only learn the nonlinear characteristics between the parameters of the second power generation equipment and the first power generation equipment of the same type, but also learn the characteristics such as delay and inertia between error influencing factors and model parameters, thereby improving the calibration capability of the calibration model.
[0077] Long Short-Term Memory (LSTM) is a special type of recurrent neural network used to solve the gradient vanishing or exploding problem that occurs during the training of traditional recurrent neural networks (RNNs), thereby effectively capturing long-term dependencies in time series.
[0078] Attention mechanisms are a key technique for improving the performance of neural network models by mimicking the selective attention ability of humans. Their core lies in enabling the model to dynamically adjust its focus on input information, rather than processing all parts equally. Specifically, this mechanism can assign differentiated weights to different parts of the input sequence, allowing the model to focus on key information while minimizing interference from secondary content.
[0079] In traditional sequence models, data is processed sequentially through the network, and hidden states are passed sequentially. The final output often depends only on the hidden state of the last step, which can easily lead to the dilution of long-distance dependencies. Attention mechanisms break this fixed pattern, allowing the model to not only refer to the current input when generating the output at each time step, but also flexibly capture other relevant parts of the input sequence, thereby effectively modeling long-distance dependencies and dynamic contextual relationships.
[0080] To address this, this application proposes a calibration method based on the LSTM-Attention mechanism. This method introduces a Long Short-Term Memory (LSTM) network to capture the nonlinear temporal dependencies of equipment operating data and dynamically weights key operating condition features using an attention mechanism, thereby achieving online adaptive calibration of predicted parameter values. Specifically, the LSTM network can learn the nonlinear and dynamic delay characteristics between newly built and mature thermal power systems, while the attention mechanism can highlight important features of the newly built thermal power system under specific operating conditions, thus improving the model's generalization ability to newly built thermal power systems and increasing the accuracy of parameter estimation.
[0081] like Figure 2 As shown, with the prediction model being a memory matrix D, after constructing the intelligent monitoring model for the second power generation equipment, the current observed value x of the parameter measurement points of the second power generation equipment can be obtained. 1obs x 2obs x 3obs ... x nobs Then, these current observations are input into the memory matrix D to obtain the predicted parameter values x of the parameter measurement points under normal operating conditions. 1est x 2est x 3est ... x nest Then, the obtained parameter prediction values and the target influence parameters x1, x2, x3, ..., x can be compared. k The input is given to the LSTM-Attention model, which then uses the target influence parameters to calibrate the predicted parameter values, resulting in more accurate predicted parameter values x. 1newest x 2newest x 3newest …、x nnewest .
[0082] Alternatively, the calibration model can be an LSTM-Attention model, or other network models with attention mechanisms, such as the Transformer model or a convolutional neural network model with attention mechanisms.
[0083] Finally, taking a steam turbine as the second power generation device as an example, the modeling method provided in the embodiments of this application will be further explained.
[0084] When constructing an intelligent monitoring model for a steam turbine, the rational selection of modeling variables is a core aspect of ensuring model performance. This process requires considering the actual layout of on-site monitoring points and selecting key measuring points from numerous operating parameters that not only reflect the dynamic characteristics of the steam turbine but also have high sensitivity to major faults. In this embodiment, by integrating the experience of domain experts and algorithm analysis, the core measuring points in Table 1 are determined as the basis for modeling, as shown in Table 1.
[0085] Table 1 In this embodiment, parameters 1-8 are set as the main parameters of the model, totaling 52 dimensions. The subsequent construction of the prediction model mainly revolves around these core parameters. These parameters reflect the key state variables of the turbine operation and form the basis of the model's predictions.
[0086] In this embodiment, parameters 9-13 are used as auxiliary influencing factors. These parameters are not involved in the construction of the prediction model, but are introduced as input features of the calibration model. These parameters include external influencing factors, internal influencing factors, and design parameters of the operation of the newly built thermal power system, which can provide richer operating condition information for the calibration model, thereby improving the model's adaptability to the newly built thermal power system units. Among them, parameter 13 is a design parameter. The rotor jacking height affects the bearing load, which in turn affects the bearing temperature. This parameter cannot be measured during unit operation, so the design value is used.
[0087] When the prediction model is a memory matrix, the construction of the memory matrix is the core step in steam turbine MSET modeling, and its quality directly determines the estimation accuracy and generalization ability of the model. To ensure that the memory matrix can comprehensively reflect the normal operating status of the equipment while taking into account computational efficiency, historical data of steam turbines in mature thermal power systems can be preprocessed, and then the memory matrix D can be constructed using an equidistant sampling method. The specific process is as follows.
[0088] First, the monitoring data (i.e., observed values) of each parameter measurement point in the historical data were normalized, mapping the actual values to the [0,1] interval. Then, the normalized data was divided into 100 equal parts, and equidistant sampling was performed with a step size of 0.01 to gradually construct the memory matrix D. To verify the sampling effect, several key measurement points were selected, such as bearing vibration, thrust bearing reverse metal temperature, bearing metal temperature, and extraction steam pressure, and their distribution in the memory matrix and the original data was compared. The results are as follows: Figures 3 to 6 As shown, the constructed memory matrix has good representativeness.
[0089] To verify the applicability of the constructed model across different devices, this embodiment conducted a prediction experiment on the constructed memory matrix. Operating data of the steam turbine from April 10th to April 20th, 2025, was collected from the DCS system of a newly built 660MW power generation system in Power Plant B, with a sampling interval of 5 minutes. Based on the model transfer paradigm, a memory matrix D constructed using historical data from the steam turbine in the mature thermal power system of Power Plant A was used to predict key parameters of the steam turbine in Power Plant B using the MSET algorithm. The prediction results were compared with the actual measured values; the prediction results for some parameters are shown below. Figures 7 to 10 As shown.
[0090] The prediction results in the figure show that directly applying the prediction model built from the historical data of the turbines in power plant A to the prediction of turbines in power plant B results in a significant deviation. The main reason is that differences in unit capacity, installation technology, operating conditions, and environment cause variations in the distribution of operating data among different types of equipment. This makes it difficult for the memory matrix built using data from power plant A to characterize the dynamic operating characteristics of the equipment in power plant B, thus leading to a decrease in prediction accuracy.
[0091] Therefore, an LSTM-Attention model is introduced as a calibration model to calibrate the predicted values output by the prediction model.
[0092] The LSTM-Attention model is composed of 52 parameters: internal influencing factors of turbine operation (main engine lubricating oil pressure, main engine lubricating oil tank and main engine lubricating oil cooler outlet temperature), external influencing factors (ambient temperature, which may also include ambient humidity and ambient pressure), boundary parameters (generator output power), and design parameters (rotor jacking height). These parameters are superimposed with the output of the MSET model. The LSTM-Attention model uses 44 turbine-related parameters as its output parameters. The training data consists of turbine operation data collected from the DCS system of the 660MW power plant B from April 10th to April 20th, 2025.
[0093] The LSTM network has one hidden layer with a time delay of 5 steps and 200 neurons. The initial learning rate is 0.01, the learning descent factor is 0.1, and the training algorithm is Adam. The attention module uses a multi-head attention mechanism with 6 attention heads. The number of channels for the key value is 5 times the number of attention heads. The attention module is connected to a fully connected network to achieve dimensionality mapping of the final output.
[0094] After the model training was completed, it was tested using the 660MW power generation system of Plant B, based on the operating data from the DCS system as of April 21, 2025. The predicted results for some parameters are as follows: Figures 11 to 14 As shown.
[0095] After training, the evaluation metrics for each parameter measurement point are shown in Table 2. As can be seen from Table 2, the MSET+LSTM-Attention model significantly improves all evaluation metrics compared to the MSET model, meeting the actual accuracy requirements for intelligent monitoring and early warning.
[0096] Table 2 In Table 2, MSE stands for mean-square error, RMSE stands for root mean-square error, MAE stands for mean absolute error, and MAPE stands for mean absolute percentage error.
[0097] The modeling process for the intelligent monitoring models of other power generation equipment in the power generation system is similar to that of the steam turbine, so examples of other power generation equipment will not be given here.
[0098] In summary, the modeling method for the intelligent monitoring model of thermal power equipment provided in this application can establish a predictive model for a newly built thermal power system by utilizing rich historical data from mature thermal power systems and a small amount of real operating data from newly built thermal power systems. This shortens the modeling time and provides a reliable monitoring mechanism for newly built thermal power systems in a timely manner. To improve the model's predictive accuracy, a calibration model is added. This calibration model can learn the differences between different types of power generation equipment, including differences caused by external and internal influencing factors. This allows the calibration model to more accurately correct the output of the predictive model, resulting in more accurate parameter estimates. Based on these more accurate parameter estimates, equipment operating status monitoring can be performed, improving the accuracy of fault warnings without constantly adding data to update the model, thus reducing modeling costs.
[0099] Exemplary device Accordingly, this application also provides a modeling device for an intelligent monitoring model of thermal power equipment, which is applied to electronic devices. The electronic devices can be terminal devices, such as laptops, desktop computers, tablets, etc.; the electronic devices can also be servers, such as cloud servers, etc.
[0100] like Figure 15 As shown, the device may include: The first model building module 1501 is used to build a prediction model for a second type of power generation equipment in a second thermal power system based on the first historical data of parameter measurement points under normal operating conditions of the first power generation equipment in the first thermal power system.
[0101] The second thermal power system has a shorter start-up time than the first thermal power system; the prediction model is used to predict the parameter values of the parameter measurement points at different times during the normal operation of the second power generation equipment.
[0102] The first prediction module 1502 is used to input the historical observation values of the parameter measurement points during the operation of the second power generation equipment into the prediction model to obtain the model prediction value.
[0103] The training module 1503 is used to train the calibration model by taking the model prediction value and the target influence parameter as input data and the historical observation value corresponding to the model prediction value as the true value.
[0104] The target influence parameters include at least one of the following: external influence parameters, internal influence parameters, design parameters of the second power generation equipment, and generator output power in the second thermal power system; the calibration model is used to correct the predicted values output by the prediction model.
[0105] The second model building module 1504 is used to build an intelligent monitoring model of the second power generation equipment based on the prediction model and the trained calibration model.
[0106] Optionally, the calibration model is an LSTM-Attention model.
[0107] The training module 1503 can be specifically used to: train the calibration model by taking the model prediction value and the target influence parameter as input data and the historical observation value corresponding to the model prediction value as output data.
[0108] Optionally, when the second power generation equipment is a steam turbine, the parameter measurement points of the second power generation equipment include at least one of the following: bearing vibration, intermediate pressure cylinder thermal expansion, steam turbine high pressure cylinder thermal expansion, axial displacement, thrust bearing reverse metal temperature, thrust bearing forward metal temperature, bearing metal temperature, and extraction steam pressure.
[0109] The external influencing parameters include at least one of the following: ambient temperature, ambient humidity, and ambient pressure.
[0110] The internal influencing parameters include at least one of the following: main engine lubricating oil pressure, main engine lubricating oil tank and main engine lubricating oil cooler outlet temperature.
[0111] The design parameters include at least the rotor apex height.
[0112] Optionally, if the parameter measuring points of the first power generation device do not include some of the parameter measuring points of the second power generation device, the device may further include: The acquisition module is used to acquire the observed values of some parameter measurement points of the second power generation equipment and the corresponding generator output power.
[0113] The fitting module is used to fit the observed values of the partial parameter measurement points with the corresponding generator output power to obtain the functional relationship between the two.
[0114] The reasoning module is used to obtain the parameter values of the partial parameter measurement points based on the functional relationship and the generator output power in the first historical data.
[0115] Optionally, the first module establishment module 1501 can be specifically used to: remove abnormal data from the first historical data based on the generator output power of the first thermal power system and the upper and lower limits of the parameter measurement points of the first power generation equipment to obtain second historical data; and establish a prediction model for the second power generation equipment of the same type based on the second historical data.
[0116] Optionally, the device may further include: The input module is used to input the current observation values of the parameter measurement points of the second power generation equipment into the intelligent monitoring model.
[0117] The second prediction module is used to output the prediction parameter value corresponding to the current observation value through the prediction model in the intelligent monitoring model, and to input the prediction parameter value and the target influence parameter into the calibration model to obtain the calibration parameter value.
[0118] The monitoring module is used to determine whether to issue a fault warning based on the difference between the current observed value and the calibration parameter value.
[0119] The modeling device for the intelligent monitoring model of thermal power equipment provided in this embodiment belongs to the same application concept as the modeling method for the intelligent monitoring model of thermal power equipment provided in the above embodiments of this application. It can execute the modeling method for the intelligent monitoring model of thermal power equipment provided in any of the above embodiments of this application, and has the corresponding functional modules and beneficial effects of the execution method. Technical details not described in detail in this embodiment can be found in the specific processing content of the modeling method for the intelligent monitoring model of thermal power equipment provided in the above embodiments of this application, and will not be repeated here.
[0120] It should be understood that the modules in the above map-using device can be implemented by a processor calling software. For example, the device includes a processor connected to a memory containing instructions. The processor calls the instructions stored in the memory to implement any of the above methods or to implement the functions of each unit of the device. The processor can be a general-purpose processor, such as a CPU or microprocessor, and the memory can be internal or external to the device. Alternatively, the units in the device can be implemented as hardware circuits. By designing the hardware circuits, some or all of the unit functions can be implemented. The hardware circuit can be understood as one or more processors. For example, in one implementation, the hardware circuit is an ASIC, and the functions of some or all of the above units are implemented by designing the logical relationships between the components within the circuit. In another implementation, the hardware circuit can be implemented by a PLD, such as an FPGA, which can include a large number of logic gates. The connection relationships between the logic gates are configured through configuration files to implement the functions of some or all of the above units. All units of the above device can be implemented entirely by a processor calling software, entirely by hardware circuits, or partially by a processor calling software with the remaining parts implemented by hardware circuits.
[0121] In this application embodiment, a processor is a circuit with signal processing capabilities. In one implementation, the processor can be a circuit with instruction reading and execution capabilities, such as a CPU, microprocessor, GPU, or DSP. In another implementation, the processor can implement certain functions through the logical relationships of hardware circuits. These logical relationships are fixed or reconfigurable. For example, the processor may be a hardware circuit implemented as an ASIC or PLD, such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the processor loading instructions to implement the functions of some or all of the above units. Furthermore, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as an NPU, TPU, or DPU.
[0122] As can be seen, each unit in the above device can be one or more processors (or processing circuits) configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.
[0123] Furthermore, the units in the above devices can be integrated in whole or in part, or they can be implemented independently. In one implementation, these units are integrated together and implemented in the form of a System-on-Chip (SoC). The SoC may include at least one processor for implementing any of the above methods or implementing the functions of the units in the device. The at least one processor may be of different types, such as CPU and FPGA, CPU and artificial intelligence processor, CPU and GPU, etc.
[0124] Exemplary electronic devices This application also provides an electronic device, such as... Figure 16 As shown, the electronic device includes a memory 1600 and a processor 1610.
[0125] The memory 1600 is connected to the processor 1610 and is used to store programs.
[0126] The processor 1610 is used to implement the modeling method of the intelligent monitoring model of thermal power equipment in the above embodiments by running the program stored in the memory 1600.
[0127] Specifically, the aforementioned electronic device may also include: a communication interface 1620, an input device 1630, an output device 1640, and a bus 1650.
[0128] The processor 1610, memory 1600, communication interface 1620, input device 1630, and output device 1640 are interconnected via a bus. Among them: Bus 1650 may include a pathway for transmitting information between various components of a computer system.
[0129] The processor 1610 can be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present invention. It can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.
[0130] The processor 1610 may include a main processor, as well as a baseband chip, modem, etc.
[0131] The memory 1600 stores a program for executing the technical solution of this invention, and may also store an operating system and other key business functions. Specifically, the program may include program code, which includes computer operation instructions. More specifically, the memory 1600 may include read-only memory (ROM), other types of static storage devices capable of storing static information and instructions, random access memory (RAM), other types of dynamic storage devices capable of storing information and instructions, disk storage, flash memory, etc.
[0132] Input device 1630 may include a device for receiving user input data and information, such as a keyboard, mouse, camera, scanner, light pen, voice input device, touch screen, pedometer, or gravity sensor.
[0133] Output device 1640 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.
[0134] The communication interface 1620 may include a device that uses any transceiver to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.
[0135] The processor 1610 executes the program stored in the memory 1600 and calls other devices, which can be used to implement the various steps of the modeling method for the intelligent monitoring model of thermal power equipment provided in the above embodiments of this application.
[0136] Exemplary computer program products and storage media In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the modeling method for the intelligent monitoring model of thermal power equipment described in the embodiments of this application.
[0137] The aforementioned computer program product can be implemented through hardware, software, or a combination thereof. In one optional embodiment, the computer program product is specifically embodied in a computer storage medium; in another optional embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.
[0138] The computer program product can be written in any combination of one or more programming languages to perform the operations of the embodiments of this application. The programming languages include object-oriented programming languages such as Java and C++, as well as conventional procedural programming languages such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0139] In addition, embodiments of this application may also be storage media storing computer programs, which are executed by a processor in the modeling method of the intelligent monitoring model of thermal power equipment described in the embodiments of this application.
[0140] In addition, embodiments of this application may also be chips, which include processors and data interfaces. The processor reads instructions stored in the memory through the data interface to execute the steps in the modeling method of the intelligent monitoring model of thermal power equipment described in the embodiments of this application.
[0141] For the foregoing method embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.
[0142] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.
[0143] The steps in the methods of the various embodiments of this application can be adjusted, merged, or deleted in order according to actual needs, and the technical features described in each embodiment can be replaced or combined.
[0144] The modules and sub-modules in the devices and terminals in the various embodiments of this application can be merged, divided, and deleted according to actual needs.
[0145] It should be understood that the disclosed terminals, devices, and methods can be implemented in other ways, given the several embodiments provided in this application. For example, the terminal embodiments described above are merely illustrative. For instance, the division of modules or sub-modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple sub-modules or modules may be combined or integrated into another module, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.
[0146] The modules or submodules described as separate components may or may not be physically separate. The components that constitute a module or submodule may or may not be physical modules or submodules; that is, they may be located in one place or distributed across multiple network modules or submodules. Some or all of the modules or submodules can be selected to achieve the purpose of this embodiment's solution, depending on actual needs.
[0147] Furthermore, the functional modules or sub-modules in the various embodiments of this application can be integrated into one processing module, or each module or sub-module can exist physically separately, or two or more modules or sub-modules can be integrated into one module. The integrated modules or sub-modules described above can be implemented in hardware or in the form of software functional modules or sub-modules.
[0148] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0149] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software unit executed by a processor, or a combination of both. The software unit can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.
[0150] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
Claims
1. A modeling method for an intelligent monitoring panel model of thermal power equipment, characterized in that, The method includes: Based on the first historical data of parameter measurement points of the first power generation equipment in the first thermal power system under normal operating conditions, a prediction model for the second power generation equipment of the same type in the second thermal power system is established; wherein, the start-up time of the second thermal power system is shorter than that of the first thermal power system; the prediction model is used to predict the parameter values of parameter measurement points at different times under normal operating conditions of the second power generation equipment. The historical observation values of the parameter measurement points during the operation of the second power generation equipment are input into the prediction model to obtain the model prediction value; The model predictions and target impact parameters are used as input data, and the historical observations corresponding to the model predictions are used as the true values to train the calibration model; wherein, the target impact parameters include at least one of the following: external impact parameters, internal impact parameters, design parameters of the second power generation equipment, and generator output power in the second thermal power system; the calibration model is used to correct the predicted values output by the prediction model; Based on the prediction model and the trained calibration model, an intelligent monitoring model for the second power generation equipment is established.
2. The modeling method for the intelligent monitoring panel model of thermal power equipment according to claim 1, characterized in that, The calibration model is an LSTM-Attention model; The step of training the calibration model by using the model's predicted values and target influence parameters as input data, and using the historical observations corresponding to the model's predicted values as the true values, includes: The model predictions and target influence parameters are used as input data, and the historical observations corresponding to the model predictions are used as output data to train the calibration model.
3. The modeling method for the intelligent monitoring panel model of thermal power equipment according to claim 1, characterized in that, When the second power generation equipment is a steam turbine, the parameter measurement points of the second power generation equipment include at least one of the following: bearing vibration, intermediate pressure cylinder thermal expansion, steam turbine high pressure cylinder thermal expansion, axial displacement, thrust bearing reverse metal temperature, thrust bearing forward metal temperature, bearing metal temperature, and extraction steam pressure. The external influencing parameters include at least one of the following: ambient temperature, ambient humidity, and ambient pressure; The internal influencing parameters include at least one of the following: main engine lubricating oil pressure, main engine lubricating oil tank and main engine lubricating oil cooler outlet temperature; The design parameters include the rotor jacking height.
4. The modeling method for the intelligent monitoring panel model of thermal power equipment according to claim 1, characterized in that, When the parameter measuring points of the first power generation equipment do not include some of the parameter measuring points of the second power generation equipment, the method further includes: Obtain the observed values of some parameter measuring points of the second power generation equipment and the corresponding generator output power; The observed values of the aforementioned parameter measurement points are fitted with the corresponding generator output power to obtain their functional relationship; Based on the functional relationship and the generator output power in the first historical data, the parameter values of the partial parameter measurement points are obtained.
5. The modeling method for the intelligent monitoring panel model of thermal power equipment according to claim 1 or 4, characterized in that, The step of establishing a prediction model for a second type of power generation equipment in a second thermal power system based on the first historical data of parameter measurement points under normal operating conditions of the first power generation equipment in the first thermal power system includes: Based on the generator output power of the first thermal power system and the upper and lower limits of the parameter measurement points of the first power generation equipment, abnormal data in the first historical data are removed to obtain the second historical data. Based on the second historical data, a predictive model for the second type of power generation equipment is established.
6. The modeling method for the intelligent monitoring model of thermal power equipment according to claim 1 or 3, characterized in that, After establishing the intelligent monitoring model of the second power generation equipment, the method further includes: Input the current observed values of the parameter measuring points of the second power generation equipment into the intelligent monitoring model; The intelligent monitoring model outputs a predicted parameter value corresponding to the current observation value, and the predicted parameter value and the target influence parameter are input into the calibration model to obtain a calibration parameter value; Based on the difference between the current observation value and the calibration parameter value, determine whether to issue a fault warning.
7. A modeling device for an intelligent monitoring panel model of thermal power equipment, characterized in that, The device includes: The first model building module is used to build a prediction model for a second power generation device of the same type in a second thermal power system based on the first historical data of parameter measurement points under normal operating conditions of the first power generation device in the first thermal power system; wherein, the start-up time of the second thermal power system is shorter than that of the first thermal power system; the prediction model is used to predict the parameter values of parameter measurement points at different times under normal operating conditions of the second power generation device; The first prediction module is used to input the historical observation values of the parameter measurement points during the operation of the second power generation equipment into the prediction model to obtain the model prediction value; The training module is used to train the calibration model by taking the model's predicted values and target influence parameters as input data and the historical observation values corresponding to the model's predicted values as the true values; wherein, the target influence parameters include at least one of the following: external influence parameters, internal influence parameters, design parameters of the second power generation equipment, and generator output power in the second thermal power system; the calibration model is used to correct the predicted values output by the prediction model; The second model building module is used to build an intelligent monitoring model of the second power generation equipment based on the prediction model and the trained calibration model.
8. The modeling device for the intelligent monitoring model of thermal power equipment according to claim 7, characterized in that, The device further includes: The input module is used to input the current observation values of the parameter measurement points of the second power generation equipment into the intelligent monitoring model; The second prediction module is used to output the prediction parameter value corresponding to the current observation value through the prediction model in the intelligent monitoring model, and to input the prediction parameter value and the target influence parameter into the calibration model to obtain the calibration parameter value; The monitoring module is used to determine whether to issue a fault warning based on the difference between the current observed value and the calibration parameter value.
9. An electronic device, characterized in that, include: Memory and processor; The memory is connected to the processor and is used to store programs; The processor is used to implement the modeling method for the intelligent monitoring model of thermal power equipment as described in any one of claims 1 to 6 by running the program in the memory.
10. A computer program product, characterized in that, The computer program product stores a computer program, which, when executed by a processor, implements the modeling method for the intelligent monitoring model of thermal power equipment as described in any one of claims 1 to 6.