Method and device for predicting change trend of condenser fouling resistance, equipment and medium

Abstract

The application provides a condenser fouling thermal resistance change trend prediction method and device, equipment and medium, and the prediction method comprises the following steps: obtaining first historical operation data of a nuclear power unit in a first preset time period, wherein the first historical operation data comprises state parameters of a nuclear power cold end device and a use frequency of a condenser cleaning device; selecting a first target model from a plurality of preset prediction models according to the state parameters and the use frequency, and training the first target model; and predicting the condenser fouling thermal resistance change trend by using the trained first target model. According to the application, the first target model is selected from the plurality of preset prediction models according to the state parameters and the use frequency, and the first target model is trained, so that the prediction model can be selected according to the historical data, and the purpose of improving the prediction accuracy of the condenser fouling thermal resistance change trend is achieved.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and specifically to a method, apparatus, equipment, and medium for predicting the trend of condenser fouling thermal resistance changes. Background Technology

[0002] The condenser is a crucial component of the cold-end system in nuclear power plants, and its operational status directly impacts the overall economic efficiency and safety of the plant. Since nuclear power condensers are cooled by seawater, which contains a large amount of microorganisms, algae, shellfish, small fish and shrimp, silt, and other impurities, as well as salts, chemical reactions can easily occur during heat exchange. Condenser fouling thermal resistance refers to the thermal resistance formed by fouling deposits on the inner walls of the condenser heat exchange tubes. These deposited fouling layers are typically poor conductors of heat, with very low thermal conductivity. Even a thin fouling layer can generate significant thermal resistance, leading to a decrease in condenser vacuum and a reduction in the cycle thermal efficiency of the nuclear power unit, thereby reducing the unit's economic efficiency and availability. Severe fouling thermal resistance on the inner walls of the condenser heat exchange tubes can lead to the shutdown of the nuclear power unit.

[0003] Besides impacting the thermal efficiency of the nuclear power unit itself, regular condenser cleaning also incurs unnecessary cleaning costs, electricity consumption, labor costs, and equipment depreciation. Therefore, calculating and predicting the degree of condenser fouling and the remaining service life before cleaning, and performing cleaning at the appropriate time to ensure unit output meets requirements while reducing unnecessary maintenance costs, is one of the important means to achieve intelligent nuclear power cold end. Furthermore, conventional prediction methods often result in predictions of condenser fouling thermal resistance trends that differ significantly from actual fouling conditions, leading to inaccurate predictions.

[0004] There is currently no effective solution to the problem of accurately predicting changes in the fouling thermal resistance of condensers in the aforementioned related technologies. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, and medium for predicting the trend of condenser fouling thermal resistance changes, thereby overcoming the problem in related technologies that it is difficult to accurately predict the changes in condenser fouling thermal resistance.

[0006] To achieve the above objectives, a first aspect of the present invention provides a method for predicting the trend of condenser fouling thermal resistance variation, comprising:

[0007] Acquire the first historical operating data of the nuclear power unit within a first preset time period. The first historical operating data includes the status parameters of the nuclear power cold end equipment and the number of times the condenser cleaning equipment is used.

[0008] Based on the state parameters and the number of times it is used, a first target model is selected from a plurality of preset prediction models, and the first target model is trained.

[0009] The trend of condenser fouling thermal resistance variation was predicted using the trained first objective model.

[0010] The method for predicting the trend of condenser fouling thermal resistance provided by this invention selects a first target model from multiple preset prediction models for model training based on state parameters and usage frequency. This allows for targeted selection of prediction models based on historical data, thereby improving the accuracy of predicting the trend of condenser fouling thermal resistance.

[0011] Optionally, in one possible implementation of the first aspect, the number of uses includes a first number of operations and a second number of operations, where the first number of operations represents the number of operations of the rubber ball cleaning equipment, and the second number of operations represents the number of operations of the chemical cleaning equipment; the prediction model includes a data fitting model, a nonlinear regression model, and a neural network model;

[0012] The step of selecting a first target model from a set of multiple preset prediction models based on the state parameters and the number of uses includes:

[0013] When both the first and second commissioning counts are zero, and the accumulation time of the state parameters reaches a first preset time threshold, the data fitting model is selected as the first target model.

[0014] When the first number of commissioning operations is greater than zero and the second number of commissioning operations is zero, the nonlinear regression model is selected as the first target model.

[0015] When the second number of commissioning operations is greater than zero, and the accumulation time of the state parameters reaches the second preset time threshold, the neural network model is selected as the first target model.

[0016] The method for predicting the trend of condenser fouling thermal resistance provided by the present invention can automatically select a prediction model that conforms to the actual situation by using the first number of operation, the second number of operation, and state parameters, thereby improving the accuracy of predicting the trend of fouling thermal resistance.

[0017] Optionally, in one possible implementation of the first aspect, the method further includes:

[0018] Determine whether the operating time of the nuclear power unit has reached the timed update time threshold;

[0019] If the runtime reaches the timed update duration threshold, then the second historical operating data of the nuclear power unit within the second preset time period is obtained; the second preset time period is later than the first preset time period.

[0020] Based on the status parameters of the nuclear power cold-end equipment and the number of times the condenser cleaning equipment was used in the second historical operation data, a second target model was selected from the multiple preset prediction models for model training.

[0021] The trained second objective model was used to predict the trend of condenser fouling thermal resistance.

[0022] The method for predicting the trend of condenser fouling thermal resistance provided by the present invention sets a timed update duration threshold. When the operating time of the nuclear power unit reaches the timed update duration threshold, a second target model is selected from multiple preset prediction models for model training. This method can reselect the type of prediction model based on historical data, thereby continuously improving the prediction effect of the prediction model.

[0023] Optionally, in one possible implementation of the first aspect, after selecting the second target model from a plurality of pre-set prediction models for model training, the process includes:

[0024] Determine whether the second target model and the first target model have the same model type;

[0025] If they are the same, the prediction effects of the trained second target model and the trained first target model are evaluated according to the model evaluation index, and the online prediction model used to predict the trend of condenser fouling thermal resistance is selected according to the evaluation results; the online prediction model refers to the model used for commissioning.

[0026] If they are different, the second prediction model is used as the online prediction model.

[0027] Optionally, in one possible implementation of the first aspect, the evaluation index is an evaluation score, and the step of selecting an online prediction model for predicting the trend of condenser fouling thermal resistance based on the evaluation results includes:

[0028] If the evaluation score of the trained second target model is greater than the evaluation score of the trained first target model, then the second target model is used as the online prediction model.

[0029] If the evaluation score of the first target model after training is greater than the evaluation score of the second target model after training, then the first target model is used as the online prediction model.

[0030] The method for predicting the trend of condenser fouling thermal resistance provided by this invention selects an online prediction model by determining whether the first target model and the second target model are the same and by evaluating the prediction effect of the model according to the model evaluation index. This allows the model with the best prediction effect to be selected from multiple models for use, thereby ensuring the accuracy of the prediction of the trend of condenser fouling thermal resistance.

[0031] Optionally, in one possible implementation of the first aspect, after selecting the first target model from a plurality of preset prediction models, the process includes:

[0032] When the first target model is the data fitting model or the nonlinear regression model, the original model file of the first target model is saved, and the original model file is distinguished according to the time identifier;

[0033] When the first target model is the neural network model, the model framework file for the first training of the first target model and the model data file for subsequent training are saved.

[0034] The method for predicting the trend of condenser fouling thermal resistance provided by the present invention can effectively reduce the server memory usage and help improve the prediction efficiency of the model by saving the original model file of the first target model, the model framework file of the first training, and the model data file of the subsequent training.

[0035] Optionally, in one possible implementation of the first aspect, obtaining the first historical operating data of the nuclear power unit within a first preset time period includes:

[0036] Obtain the status parameters of the nuclear power plant cold end equipment within a first preset time period, the status parameters including temperature parameters, pressure parameters and flow parameters;

[0037] Obtain the start and end times of each use of the condenser cleaning equipment within a first preset time period, and calculate the time period of each use based on the start and end times;

[0038] Based on the time period of each use, the number of times the condenser cleaning equipment is used within the first preset time period is determined.

[0039] A second aspect of the present invention provides a device for predicting the trend of condenser fouling thermal resistance variation, comprising:

[0040] The data acquisition module is used to acquire the first historical operating data of the nuclear power unit within a first preset time period. The first historical operating data includes the status parameters of the nuclear power cold end equipment and the number of times the condenser cleaning equipment is used.

[0041] The model training module is used to select a first target model from a plurality of preset prediction models based on the state parameters and the number of times it is used, and to train the first target model.

[0042] The prediction module is used to predict the trend of condenser fouling thermal resistance using the trained first target model.

[0043] A third aspect of the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program executable on the processor, and the processor executes the computer program to implement the steps in the various method embodiments described above.

[0044] A fourth aspect of the present invention provides a readable storage medium storing a computer program, which, when executed by a processor, is used to implement the steps of the method described in the first aspect of the present invention and various possible designs of the first aspect. Attached Figure Description

[0045] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0046] Figure 1 This is a flowchart illustrating the method for predicting the trend of condenser fouling thermal resistance in Embodiment 1 of the present invention.

[0047] Figure 2 This is a schematic diagram of the implementation process corresponding to the prediction method in at least one embodiment of the present invention.

[0048] Figure 3 This is a schematic diagram of the device for predicting the trend of condenser fouling thermal resistance in Embodiment 2 of the present invention.

[0049] Figure 4 This is a structural diagram of the computer device in Embodiment 3 of the present invention. Detailed Implementation

[0050] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0052] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can also refer to the internal connection of two components; and they can refer to a wireless connection or a wired connection. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0053] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0054] Considering that the effects of ball cleaning and chemical cleaning on the formation and removal of condenser fouling thermal resistance are inconsistent, the units may be newly built and lack historical cleaning data, or the units may suddenly change their operating mode, the existing models for predicting the trend of condenser fouling thermal resistance generally have a large difference between the trend prediction results and the actual fouling situation, resulting in inaccurate predictions and seriously affecting the user experience.

[0055] Example 1

[0056] This embodiment provides a method for predicting the trend of condenser fouling thermal resistance changes, such as... Figure 1 , 2 As shown, where Figure 1 This is a flowchart illustrating the prediction method. Figure 2 This is a schematic diagram of the implementation process corresponding to the prediction method; the prediction method includes, but is not limited to, the following steps:

[0057] S100: Obtain the first historical operating data of the nuclear power unit within the first preset time period.

[0058] Specifically, the first preset time period can be understood as a period of time prior to predicting the changing trend of condenser fouling thermal resistance; the first historical operating data (e.g. Figure 2The “nuclear power unit operation data” in the document may include the status parameters of the nuclear power cold end equipment (e.g., Figure 2 The "cold end equipment status data" and the number of times the condenser cleaning equipment is used (e.g.) Figure 2 The “cleaning system commissioning mark” indicates that the condenser cleaning equipment includes ball cleaning equipment and chemical cleaning equipment, which are mainly used to clean the accumulated dirt on the inner wall of the heat exchange tubes of the condenser in the cold end equipment of nuclear power plants.

[0059] The status parameters may include temperature, pressure, and flow parameters of the nuclear power plant cold-end equipment; the number of uses may include the number of uses of the condenser cleaning equipment and the number of uses of the chemical cleaning equipment. The number of uses can be calculated through the following steps: obtain the start and end times of each use of the condenser cleaning equipment, and count the time period of each use based on the start and end times; determine the number of uses of the condenser cleaning equipment within a first preset time period based on the time period of each use. Taking the number of uses of the condenser cleaning equipment as an example: if one month is selected as the first preset time period, and there are three time periods within this month, then the number of uses of the condenser cleaning equipment is determined to be 3 times; the method for determining the number of uses of the chemical cleaning equipment is similar.

[0060] More specifically, the status parameters of the nuclear power cold-end equipment need to be transmitted to the first database in real time, and the status parameters are marked with corresponding time tags; the start and end times of each use of the condenser cleaning equipment are also transmitted to the second database in real time. When predicting the changing trend of the fouling thermal resistance of the condenser, all status parameters with time tags within a first preset time period are retrieved from the first database, and the start and end times of each use of the condenser cleaning equipment within the first preset time period are retrieved from the second database, thereby calculating the number of uses.

[0061] S200: Select the first target model from multiple preset prediction models based on the state parameters and the number of times it is used, and train the first target model.

[0062] Specifically, the pre-defined prediction models include three types: data fitting models based on a combination of mechanistic formulas and data fitting (e.g., Figure 2 "Mechanism formula data fitting" and nonlinear regression models based on data-driven and feature engineering (e.g.) Figure 2 The “nonlinear regression model” in the text), and the long short-term memory neural network model based on time series prediction (e.g. Figure 2 (The "neural network model" in the text). The number of uses can be divided into the first number of operations and the second number of operations. The first number of operations refers to the number of times the rubber ball cleaning equipment is used, and the second number of operations refers to the number of times the chemical cleaning equipment is used.

[0063] Step S200 includes, but is not limited to, steps S210 to S230:

[0064] S210: When both the first and second commissioning counts are zero, and the accumulation time of the state parameters reaches a first preset time threshold, the data fitting model is selected as the first target model, and model training is performed on the data fitting model. The first preset time threshold can be limited according to actual circumstances, and is not specifically limited here.

[0065] Specifically, when selecting a data fitting model as the primary target model for model training, the training steps may include: There exists a functional relationship y = f(x) between the commissioning time of the condenser cleaning equipment and the condenser fouling thermal resistance, where x is the time interval since the last commissioning and y is the fouling thermal resistance calculated based on the mechanistic formula. This functional relationship (the data fitting model) can be fitted using historical operating data of the nuclear power unit. When using the data fitting model for prediction, the predicted demand (i.e., the commissioning time of the condenser cleaning equipment) can be input into the aforementioned functional relationship to output the corresponding fouling thermal resistance.

[0066] S220: When the first commissioning count is greater than zero and the second commissioning count is zero, select the nonlinear regression model as the first target model and train the nonlinear regression model.

[0067] Specifically, when a nonlinear regression model is selected as the primary target model for model training, a regression model of x (i.e., the feature vector) and y (fouling thermal resistance) can be established based on historical commissioning time and fouling thermal resistance characteristic values. When using the nonlinear regression model for prediction, a new commissioning time can be input, and the corresponding features can be extracted according to the training logic, input into the regression model, and the new fouling thermal resistance can be calculated.

[0068] S230: When the second commissioning count is greater than zero, and the accumulation time of the state parameters reaches the second preset time threshold, a neural network model is selected as the first target model, and the neural network model is trained. The second preset time threshold can be limited according to actual circumstances, and is not specifically limited here.

[0069] Specifically, when selecting a neural network model (which could be a long short-term memory neural network model based on time series prediction) as the primary target model for model training, a time series model of X (feature matrix) and y (fouling thermal resistance vector) can be established based on the feature values ​​over continuous time. When using the neural network model for prediction, a new commissioning time can be input, and according to the logic of X during model training, the matrix in the model prediction can be established, input into the time series model, and the corresponding fouling thermal resistance sequence can be output.

[0070] S300: Predict the trend of condenser fouling thermal resistance using the trained first target model.

[0071] Specifically, by inputting the nuclear power unit's operating data for the third preset time period (representing a period before prediction) into the trained first target model, the changing trend of condenser fouling thermal resistance for the fourth preset time period (representing a future period) can be output, along with the resolution of the prediction time range. For example, based on the nuclear power unit's operating data from the previous 60 minutes, the changing trend of fouling thermal resistance within the next 15 minutes can be predicted, i.e., predicting five future fouling thermal resistance points with a time resolution of 3 minutes. The third preset time period, fourth preset time period, prediction time range, and resolution can be set according to actual conditions and are not specifically limited here. The aforementioned third preset time period is not equivalent to the first and second preset time periods. The third time period refers to the time period corresponding to the operating data used to predict the changing trend of condenser fouling thermal resistance, while the first and second preset time periods are mainly used for the time periods corresponding to the operating data used to determine the type of prediction model.

[0072] Preferably, the method further includes:

[0073] Step 1: Determine whether the operating time of the nuclear power unit has reached the timed update time threshold.

[0074] Specifically, runtime can be understood as the total operating time of the nuclear power unit, or the uninterrupted operating time of the nuclear power unit; the timed update duration threshold refers to the duration during which the first target model needs to be updated, which can be set according to the actual situation and is not specifically limited here. Considering that if a prediction model is used to predict the trend of condenser fouling thermal resistance over a long period of time, the historical operating data of the nuclear power unit may change due to the excessive time, and the prediction results may be inaccurate if the original prediction model is used for prediction again, it is necessary to set a timed update duration threshold to update the prediction model according to the timed update duration threshold.

[0075] Step 2: If the runtime reaches the timed update duration threshold, then obtain the second historical operating data of the nuclear power unit within the second preset time period; the second preset time period is later than the first preset time period; the generation time of the second historical operating data is later than the production time of the first historical operating data.

[0076] Step 3: Based on the status parameters of the nuclear power plant cold-end equipment and the number of times the condenser cleaning equipment was used in the second historical operation data, select the second target model from multiple preset prediction models for model training;

[0077] Step 4: Use the trained second objective model to predict the trend of condenser fouling thermal resistance.

[0078] Specifically, let's illustrate steps 1-4 above with an example: Assuming the timed update duration threshold is one month, the first preset time period is June, and the second preset time period is July, on July 1st, the nuclear power unit can select the first target model from multiple preset prediction models based on the operating data of the nuclear power unit in June for model training, and use the trained first target model to predict the trend of condenser fouling thermal resistance changes; when the nuclear power unit operates from July 1st to August 1st, and the operating time of the nuclear power unit reaches the one-month timed update duration threshold, it will select the second target model from multiple preset prediction models based on the operating data of the nuclear power unit in July for model training, and use the trained second target model to predict the trend of condenser fouling thermal resistance changes.

[0079] Preferably, after selecting a second target model from a plurality of pre-set prediction models for model training, the process includes:

[0080] Determine whether the second target model and the first target model are of the same type. If they are the same, evaluate the prediction effect of the trained second target model and the trained first target model according to the model evaluation index, and select the online prediction model to predict the trend of condenser fouling thermal resistance change based on the evaluation results. The online prediction model is the model to be put into use. If they are different, the second prediction model is used as the online prediction model.

[0081] Specifically, when the first target model and the second target model are of the same type, the prediction performance of the trained models of the same type needs to be evaluated using model evaluation metrics, and the model with better prediction performance is used as the online prediction model. When the first target model and the second target model are different, considering that the second target model is trained based on the second historical running data and the first target model is trained based on the first historical running data, and the second historical running data is later than the first historical running data, the prediction performance of the second target model is more in line with the current situation. Therefore, when the first target model and the second target model are different, the second prediction model can be directly used as the online prediction model.

[0082] Preferably, the evaluation index is an evaluation score, and an online prediction model for predicting the trend of condenser fouling thermal resistance is selected based on the evaluation results, including:

[0083] If the evaluation score of the trained second target model is greater than the evaluation score of the trained first target model, then the second target model will be used as the online prediction model.

[0084] If the evaluation score of the first target model after training is greater than the evaluation score of the second target model after training, then the first target model will be used as the online prediction model.

[0085] Specifically, the evaluation score of the trained second target model is compared with the score data of the trained first target model. If the evaluation score of the second target model is greater than the evaluation score of the trained first target model, the second target model is used as an online prediction model and put into practical use; if the evaluation score of the first target model is greater than the evaluation score of the trained second target model, the first target model is used as an online prediction model and put into practical use.

[0086] Preferably, when the target model is a nonlinear regression model based on data-driven and feature engineering, the following regression model evaluation indicators can be used: absolute error (MAE), mean squared error (MSE), coefficient of determination, or goodness of fit.

[0087] Preferably, after selecting the first target model from a plurality of preset prediction models, the process includes:

[0088] When the first target model is a data fitting model or a nonlinear regression model, the original model file of the first target model is saved and the original model file is distinguished according to the time identifier; when the first target model is a neural network model, the model framework file of the first training of the first target model and the model data file of subsequent training are saved.

[0089] Specifically, after training the model, it is necessary to save the training model. For the same type of model, when the model is a data fitting model or a nonlinear regression model, the original file of the model is saved directly and distinguished by time stamp. When the model is a neural network model, the model framework file needs to be saved the first time the model is saved. After subsequent training, only the data file generated after each training session needs to be saved (e.g., the checkpoint file. Checkpoint is an internal event that triggers the database write process to write dirty data blocks in the data buffer to the data file), thereby reducing the memory usage of the server.

[0090] Preferably, acquiring the first historical operating data of the nuclear power unit within a first preset time period includes:

[0091] Obtain the status parameters of the nuclear power cold end equipment within a first preset time period. The status parameters include temperature parameters, pressure parameters, and flow parameters.

[0092] Obtain the start and end times of each use of the condenser cleaning equipment within the first preset time period, and calculate the time period of each use based on the start and end times;

[0093] The number of times the condenser cleaning equipment is used within the first preset time period is determined based on the time period of each use.

[0094] The following example illustrates the prediction of fouling thermal resistance trends in a newly built nuclear power plant cold-end system:

[0095] 1) Determine the framework of the prediction model: According to the actual requirements of the nuclear power plant, it is necessary to predict the change of fouling thermal resistance in the next 15 minutes based on the historical operating data of the condenser. The time resolution is 3 minutes. Therefore, the prediction framework is to predict the change of fouling thermal resistance in the next 15 minutes based on the data of a historical period, that is, to predict the 5 future fouling thermal resistance points.

[0096] 2) Determine the operational status of the cleaning system: Based on the actual situation of the nuclear power plant, there is only data within one cleaning cycle of the glue ball. A distributed gradient enhancement library model can be selected for model training.

[0097] 3) Model training: Based on the prediction framework confirmed in the first step, train the distributed gradient enhancement library model of the multivariate regression model, and save the model file A directly after training, and put model A online for model prediction.

[0098] 4) Scheduled Updates: A one-month period is selected as the model's scheduled update cycle. Based on the nuclear power plant's operating time, the scheduled update strategy is activated when the update time is reached. The data is reassessed, the model type is selected, and model training is performed to obtain Model B. If Model A and Model B are of the same type, an evaluation and selection process is required.

[0099] 5) Model Evaluation: The two models were evaluated using evaluation metrics. Comparing the results, it was found that Model A's evaluation metric score was lower than Model B's. Therefore, Model B was selected as the online model.

[0100] The technical solution of the present invention also has the following technical effects:

[0101] At least one embodiment of the present invention can automatically select a model and train the model based on the number of times the condenser cleaning equipment is used and the state parameters of the nuclear power cold end; and can set a timed prediction model for updating, so that the model prediction effect can be continuously improved as the amount of data accumulates.

[0102] Example 2

[0103] This embodiment provides a device for predicting the trend of condenser fouling thermal resistance variation, such as... Figure 3 As shown, it includes:

[0104] The data acquisition module is used to acquire the first historical operating data of the nuclear power unit within a first preset time period. The first historical operating data includes the status parameters of the nuclear power cold end equipment and the number of times the condenser cleaning equipment has been used.

[0105] The model training module is used to select a first target model from multiple pre-set prediction models based on state parameters and usage frequency, and to train the first target model.

[0106] The prediction module is used to predict the trend of condenser fouling thermal resistance using the trained first target model.

[0107] For details, please refer to the corresponding sections of the above method embodiments, which will not be repeated here.

[0108] Preferably, the number of uses includes a first number of operations and a second number of operations, where the first number of operations represents the number of operations of the rubber ball cleaning equipment and the second number of operations represents the number of operations of the chemical cleaning equipment; the prediction model includes a data fitting model, a nonlinear regression model, and a neural network model;

[0109] The model training module includes:

[0110] The first selection unit is used to select the data fitting model as the first target model when both the first and second commissioning times are zero and the accumulation time of the state parameters reaches a first preset time threshold.

[0111] The second selection unit is used to select a nonlinear regression model as the first target model when the first commissioning count is greater than zero and the second commissioning count is zero.

[0112] The third selection unit is used to select a neural network model as the first target model when the second number of commissioning operations is greater than zero and the accumulation time of the state parameters reaches the second preset time threshold.

[0113] For details, please refer to the corresponding sections of the above method embodiments, which will not be repeated here.

[0114] Preferably, the device for predicting the trend of condenser fouling thermal resistance further includes:

[0115] The judgment module is used to determine whether the running time of the nuclear power unit has reached the timed update duration threshold.

[0116] The second data acquisition module is used to acquire the second historical operating data of the nuclear power unit within a second preset time period if the running time reaches the timed update duration threshold; the second preset time period is later than the first preset time period.

[0117] The second model training module is used to select a second target model from multiple preset prediction models for model training based on the state parameters of the nuclear power cold end equipment and the number of times the condenser cleaning equipment is used in the second historical operation data.

[0118] The second prediction module is used to predict the trend of condenser fouling thermal resistance using the trained second objective model.

[0119] For details, please refer to the corresponding sections of the above method embodiments, which will not be repeated here.

[0120] Preferably, the device for predicting the trend of condenser fouling thermal resistance further includes:

[0121] The type determination unit is used to determine whether the model types of the second target model and the first target model are the same;

[0122] The first judgment unit is used to evaluate the prediction effects of the trained second target model and the trained first target model according to the model evaluation index if they are the same, and select the online prediction model to predict the trend of condenser fouling thermal resistance change based on the evaluation results; the online prediction model represents the model to be put into use.

[0123] The second judgment unit is used to use the second prediction model as the online prediction model if they are different.

[0124] For details, please refer to the corresponding sections of the above method embodiments, which will not be repeated here.

[0125] Preferably, the evaluation index is an evaluation score, and the first judgment unit further includes:

[0126] The first determining subunit is used to use the second target model as the online prediction model if the evaluation score of the trained second target model is greater than the evaluation score of the trained first target model.

[0127] The second step is to determine a sub-unit, which is used to select the first target model as the online prediction model if the evaluation score of the first target model after training is greater than the evaluation score of the second target model after training.

[0128] For details, please refer to the corresponding sections of the above method embodiments, which will not be repeated here.

[0129] Preferably, the device for predicting the trend of condenser fouling thermal resistance further includes:

[0130] The first data storage module is used to save the original model file of the first target model when the first target model is a data fitting model or a nonlinear regression model, and to distinguish the original model file according to the time identifier.

[0131] The second data storage module is used to save the model framework file of the first target model during its initial training and the model data file of subsequent training when the first target model is a neural network model.

[0132] For details, please refer to the corresponding sections of the above method embodiments, which will not be repeated here.

[0133] Preferably, the data acquisition module includes:

[0134] The status parameter acquisition unit is used to acquire the status parameters of the nuclear power cold end equipment within a first preset time period. The status parameters include temperature parameters, pressure parameters, and flow parameters.

[0135] The usage count acquisition unit is used to acquire the start and end times of each use of the condenser cleaning equipment within a first preset time period, and to count the time period of each use based on the start and end times; and to determine the number of times the condenser cleaning equipment is used within the first preset time period based on the time period of each use.

[0136] For details, please refer to the corresponding sections of the above method embodiments, which will not be repeated here.

[0137] Example 3

[0138] The present invention also provides a computer device, such as Figure 4 As shown, it includes a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the computer program, it implements the method for predicting the trend of condenser fouling thermal resistance provided in the various embodiments described above.

[0139] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for predicting the trend of condenser fouling thermal resistance provided in the various embodiments described above.

[0140] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0141] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0142] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0143] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0144] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for predicting the trend of condenser fouling thermal resistance variation, characterized in that, include: Acquire the first historical operating data of the nuclear power unit within a first preset time period. The first historical operating data includes the status parameters of the nuclear power cold end equipment and the number of times the condenser cleaning equipment is used. The number of times the equipment is used includes a first number of times it is put into operation and a second number of times it is put into operation. The first number of times it is put into operation represents the number of times the ball cleaning equipment is put into operation, and the second number of times it is put into operation represents the number of times the chemical cleaning equipment is put into operation. Based on the state parameters and the number of uses, a first target model is selected from a plurality of preset prediction models, and the first target model is trained; the plurality of prediction models include a data fitting model, a nonlinear regression model, and a neural network model; The trend of condenser fouling thermal resistance variation was predicted using the trained first objective model; The step of selecting a first target model from a set of multiple preset prediction models based on the state parameters and the number of uses includes: When both the first and second commissioning counts are zero, and the accumulation time of the state parameters reaches a first preset time threshold, the data fitting model is selected as the first target model. When the first number of commissioning operations is greater than zero and the second number of commissioning operations is zero, the nonlinear regression model is selected as the first target model. When the second number of commissioning operations is greater than zero, and the accumulation time of the state parameters reaches the second preset time threshold, the neural network model is selected as the first target model.

2. The method for predicting the trend of condenser fouling thermal resistance according to claim 1, characterized in that, The method further includes: Determine whether the operating time of the nuclear power unit has reached the timed update time threshold; If the runtime reaches the timed update duration threshold, then the second historical operating data of the nuclear power unit within the second preset time period is obtained; the second preset time period is later than the first preset time period. Based on the status parameters of the nuclear power cold-end equipment and the number of times the condenser cleaning equipment was used in the second historical operation data, a second target model was selected from the multiple preset prediction models for model training. The trained second objective model was used to predict the trend of condenser fouling thermal resistance.

3. The method for predicting the trend of condenser fouling thermal resistance according to claim 2, characterized in that, After selecting the second target model from the preset multiple prediction models for model training, the process includes: Determine whether the second target model and the first target model have the same model type; If they are the same, the prediction effects of the trained second target model and the trained first target model are evaluated according to the model evaluation index, and the online prediction model used to predict the trend of condenser fouling thermal resistance is selected according to the evaluation results; the online prediction model refers to the model used for commissioning. If they are not the same, the second target model is used as the online prediction model.

4. The method for predicting the trend of condenser fouling thermal resistance according to claim 3, characterized in that, The evaluation index is an evaluation score, and the online prediction model selected based on the evaluation results to predict the trend of condenser fouling thermal resistance includes: If the evaluation score of the trained second target model is greater than the evaluation score of the trained first target model, then the second target model is used as the online prediction model. If the evaluation score of the first target model after training is greater than the evaluation score of the second target model after training, then the first target model is used as the online prediction model.

5. The method for predicting the trend of condenser fouling thermal resistance according to claim 1, characterized in that, After selecting the first target model from a set of multiple pre-set prediction models, the process includes: When the first target model is the data fitting model or the nonlinear regression model, the original model file of the first target model is saved, and the original model file is distinguished according to the time identifier; When the first target model is the neural network model, the model framework file for the first training of the first target model and the model data file for subsequent training are saved.

6. The method for predicting the trend of condenser fouling thermal resistance according to claim 1, characterized in that, The acquisition of the first historical operating data of the nuclear power unit within the first preset time period includes: Obtain the status parameters of the nuclear power plant cold end equipment within a first preset time period, the status parameters including temperature parameters, pressure parameters and flow parameters; Obtain the start and end times of each use of the condenser cleaning equipment within a first preset time period, and calculate the time period of each use based on the start and end times; The number of times the condenser cleaning equipment is used within the first preset time period is determined based on the time period of each use.

7. A device for predicting the trend of condenser fouling thermal resistance variation, characterized in that, include: The data acquisition module is used to acquire the first historical operating data of the nuclear power unit within a first preset time period. The first historical operating data includes the status parameters of the nuclear power cold end equipment and the number of times the condenser cleaning equipment is used. The number of times the equipment is used includes a first number of times it is put into operation and a second number of times it is put into operation. The first number of times it is put into operation represents the number of times the ball cleaning equipment is put into operation, and the second number of times it is put into operation represents the number of times the chemical cleaning equipment is put into operation. The model training module is used to select a first target model from a plurality of preset prediction models based on the state parameters and the number of times it is used, and to train the first target model; the plurality of prediction models include a data fitting model, a nonlinear regression model, and a neural network model; The prediction module is used to predict the trend of condenser fouling thermal resistance using the trained first target model. The model training module is used for: When both the first and second commissioning counts are zero, and the accumulation time of the state parameters reaches a first preset time threshold, the data fitting model is selected as the first target model. When the first number of commissioning operations is greater than zero and the second number of commissioning operations is zero, the nonlinear regression model is selected as the first target model. When the second number of commissioning operations is greater than zero, and the accumulation time of the state parameters reaches the second preset time threshold, the neural network model is selected as the first target model.

8. A computer device comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for predicting the trend of condenser fouling thermal resistance according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for predicting the trend of condenser fouling thermal resistance as described in any one of claims 1 to 6.

Images