Early warning method and device for scale deposition fault of centrifugal compressor, medium and product
The method uses surface roughness and static pressure data in a trained model to assess centrifugal compressor fouling, providing accurate early warnings and enhancing maintenance efficiency.
Patent Information
- Application Number
- CN202411194683.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-29
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-08-29
AI Technical Summary
Centrifugal compressors in turbochargers suffer from fouling due to particle adhesion, leading to performance degradation, increased fuel consumption, and maintenance costs, with existing methods lacking effective early warning systems.
An early warning method using surface roughness and static pressure data input into a trained operating status prediction model, based on new evaluation indices derived from predicted and actual flow field data, to accurately assess fouling failure.
Enables precise early detection of centrifugal compressor fouling, reducing false alarms and improving maintenance efficiency by quantifying performance changes.
Smart Images

Figure CN119084341A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of fault detection, and in particular to an early warning method, device, medium and product for a fouling fault in a centrifugal compressor. Background Art
[0002] As a common power device, the engine has been widely used in ships, automobiles, power generation and other fields. With the increase in the demand for power output of ship engines, turbochargers have become the core components of diesel engines. However, during the operation of marine diesel engines, particles such as dust and salt spray contained in the atmosphere gradually adhere to the surface of the compressor blades to form fouling, causing engine performance degradation. As one of the main components of the turbocharger, the centrifugal compressor is inevitably affected by these risks during operation, resulting in performance degradation, increased fuel consumption, increased additional maintenance costs, and a narrowing of the stable operating range.
[0003] The increase in blade surface roughness has been shown to have a significant impact on the aerodynamic performance of the compressor. Studies have shown that due to the ingestion of corrosive substances and foreign matter such as sand and stone particles, the surface roughness of the compressor blades increases, thus affecting the geometry of the blades and resulting in a reduction in the effective flow area of the flow channel. This change causes the compressor to deviate from its design state, thereby reducing its overall performance. Summary of the invention
[0004] The purpose of this application is to provide an early warning method, equipment, medium and product for centrifugal compressor fouling failure, which can achieve accurate early warning of centrifugal compressor fouling failure.
[0005] To achieve the above objectives, this application provides the following solutions:
[0006] In a first aspect, the present application provides an early warning method for a centrifugal compressor fouling failure, comprising:
[0007] The surface roughness and static pressure of components in a target centrifugal compressor are obtained; the surface roughness of the components includes the surface roughness of the impeller and the surface roughness of the diffuser.
[0008] The surface roughness and static pressure of the components are input into a trained operating status prediction model to obtain the fouling failure degree of the target centrifugal compressor; the trained operating status prediction model is a prediction model based on a new evaluation index of the equipment operating status; the new evaluation index of the equipment operating status is an evaluation index constructed based on predicted flow field data and actual flow field data; the flow field data includes a corrected volume flow rate, a pressure ratio and a compression efficiency.
[0009] Optionally, the training process of the running status prediction model is:
[0010] The surface roughness of several sample components and sample static pressure are designed through CFD simulation, and the sample flow field data corresponding to the surface roughness of the sample components and the sample static pressure are obtained; the surface roughness of the sample components includes the surface roughness of the sample impeller and the surface roughness of the sample diffuser; the sample flow field data includes the sample corrected volume flow rate, the sample pressure ratio and the sample isentropic efficiency.
[0011] The deep forest model is trained with sample component surface roughness and sample static pressure as input and sample flow field data corresponding to the sample component surface roughness and sample static pressure as labels to obtain a trained optimal deep forest model; the trained optimal deep forest model is used to output the flow field data of the target centrifugal compressor.
[0012] The new evaluation index of the equipment operation status and the preset fault threshold are imported into the trained optimal deep forest model to obtain a trained operation status prediction model; the trained operation status prediction model is used to: obtain the current operation status value of the target centrifugal compressor based on the flow field data of the target centrifugal compressor output by the trained optimal deep forest model and the new evaluation index of the equipment operation status; judge the fouling fault degree of the equipment based on the preset fault threshold according to the current operation status value of the equipment; the current operation status value is represented by the pressure ratio-corrected volume flow curve composed of the pressure ratio, efficiency and corrected volume flow in the flow field data, and the area enclosed by the pressure ratio axis.
[0013] Optionally, the formula expression of the new evaluation index of the equipment operation status is:
[0014]
[0015] Among them, ΔS is a new evaluation index for equipment operation status, S 预测 is the area enclosed by the pressure ratio-corrected volume flow curve and the pressure ratio axis obtained based on the predicted flow field data; 实际 It is the area enclosed by the pressure ratio-corrected volume flow curve obtained based on the actual flow field data and the pressure ratio axis.
[0016] Optionally, the S 实际 The solution expression is:
[0017]
[0018] Among them, n represents the number of vertices, x i ,y i is the x-coordinate and y-coordinate of the i-th vertex, x i+1 ,y i+1 is the x-coordinate and y-coordinate of the i+1th vertex, x n ,y nare the x and y coordinates of the nth vertex.
[0019] Optionally, judging the fouling fault degree of the equipment according to the current operating state value of the equipment and based on a preset fault threshold value specifically includes:
[0020] When the current operating state value of the equipment is greater than the first fault threshold, it is determined that the target centrifugal compressor is in a normal operating state.
[0021] When the current operating state value of the equipment is greater than or equal to the second fault threshold and less than the first fault threshold, it is determined that the target centrifugal compressor is in a minor fault state.
[0022] When the current operating state value of the equipment is greater than or equal to the third fault threshold and less than the second fault threshold, it is determined that the target centrifugal compressor is in a common fault state.
[0023] When the current operating state value of the equipment is less than the third fault threshold, it is determined that the target centrifugal compressor is in a serious fault state.
[0024] Optionally, the pressure ratio is a total-total pressure ratio, which refers to the ratio of the total pressure at the device outlet to the total pressure at the device inlet.
[0025] Optionally, the roughness setting range of the sample impeller surface roughness among the plurality of sample component surface roughnesses is 1 to 150 μm; the roughness setting range of the sample diffuser surface roughness is 1 to 150 μm.
[0026] In a second aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any one of the above-mentioned early warning methods for centrifugal compressor fouling failures.
[0027] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any one of the above-mentioned early warning methods for fouling failure of a centrifugal compressor.
[0028] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any one of the above-mentioned early warning methods for fouling failure of a centrifugal compressor.
[0029] According to the specific embodiments provided in this application, this application discloses the following technical effects:
[0030] The present application provides an early warning method, equipment, medium and product for fouling failure of centrifugal compressors, which obtains the surface roughness and static pressure values of the components of the target centrifugal compressor; wherein the surface roughness of the components covers the surface conditions of the impeller and diffuser. Subsequently, these surface roughness data and static pressure values are input into a trained operating status prediction model to evaluate the degree of fouling failure of the target centrifugal compressor. The operating status prediction model is constructed based on new equipment operating status evaluation indicators, and these new evaluation indicators are formed by comparing the predicted flow field data with the actual flow field data. The flow field data includes key parameters such as corrected volume flow rate, pressure ratio and isentropic efficiency. The present application uses these new evaluation indicators to achieve accurate early warning of fouling failures in centrifugal compressors. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0032] Figure 1 This is a diagram of the application environment of an early warning method for centrifugal compressor fouling failure in one embodiment of the present application.
[0033] Figure 2 A schematic flow chart of a method for early warning of fouling failure in a centrifugal compressor provided in accordance with an embodiment of the present application.
[0034] Figure 3 A pressure ratio-corrected volume flow diagram provided for an embodiment of the present application.
[0035] Figure 4 An efficiency-corrected volume flow diagram provided for an embodiment of the present application.
[0036] Figure 5 A schematic diagram of solving a new evaluation index for equipment operating status provided in one embodiment of the present application.
[0037] Figure 6 A framework diagram of a compressor fouling fault warning process flow for implementing a method based on a combination of a deep forest model and a new evaluation index provided in an embodiment of the present application.
[0038] Figure 7 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION
[0039] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0040] The purpose of this application is to provide an early warning method, equipment, medium and product for centrifugal compressor fouling failure, which can achieve accurate early warning of centrifugal compressor fouling failure.
[0041] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0042] The early warning method for centrifugal compressor fouling failure provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send the surface roughness and static pressure of the parts to be processed to the server 104. After the server 104 receives the surface roughness and static pressure of the parts to be processed, the server 104 obtains the surface roughness and static pressure of the parts in the target centrifugal compressor for the surface roughness and static pressure of the parts to be processed; the surface roughness of the parts includes the surface roughness of the impeller and the surface roughness of the diffuser; the surface roughness and static pressure of the parts are input into the trained operating state prediction model to obtain the fouling fault degree of the target centrifugal compressor; the trained operating state prediction model is a prediction model based on a new evaluation index of the equipment operating state; the new evaluation index of the equipment operating state is an evaluation index constructed based on the predicted flow field data and the actual flow field data; the flow field data includes the corrected volume flow rate, the pressure ratio and the isentropic efficiency. The server 104 can feed back the obtained fouling fault degree of the target centrifugal compressor to the terminal 102. In addition, in some embodiments, the early warning method for the fouling fault of the centrifugal compressor can also be implemented by the server 104 or the terminal 102 alone, such as the terminal 102 can directly perform data processing on the surface roughness and static pressure of the parts to be processed, or the server 104 can obtain the surface roughness and static pressure of the parts to be processed from the data storage system and perform data processing on the surface roughness and static pressure of the parts to be processed.
[0043] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart vehicle-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.
[0044] In an exemplary embodiment, Figure 2 As shown, a method for early warning of fouling failure of a centrifugal compressor is provided. The method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used for explanation, and the steps include the following steps 201 to 202.
[0045] in:
[0046] Step 201, obtaining the surface roughness and static pressure of components in a target centrifugal compressor; the component surface roughness includes the impeller surface roughness and the diffuser surface roughness.
[0047] Step 202, input the surface roughness and static pressure of the components into the trained operating status prediction model to obtain the fouling failure degree of the target centrifugal compressor; the trained operating status prediction model is a prediction model based on a new evaluation index of the equipment operating status; the new evaluation index of the equipment operating status is an evaluation index constructed based on the predicted flow field data and the actual flow field data; the flow field data includes a corrected volume flow rate, a pressure ratio and an isentropic efficiency.
[0048] The training process of the running status prediction model is as follows:
[0049] Step 301: Design several sample component surface roughnesses and sample static pressures through CFD simulation, and obtain sample flow field data corresponding to the sample component surface roughnesses and sample static pressures; the sample component surface roughnesses include sample impeller surface roughnesses and sample diffuser surface roughnesses; the sample flow field data include sample corrected volume flow rate, sample pressure ratio and sample isentropic efficiency.
[0050] Step 302: Build a deep forest model.
[0051] Step 303: Taking the surface roughness of sample parts and sample static pressure as input and the sample flow field data corresponding to the surface roughness of sample parts and sample static pressure as labels, the deep forest model is trained to obtain a trained optimal deep forest model; the trained optimal deep forest model is used to output the flow field data of the target centrifugal compressor.
[0052] Step 304: Import the new evaluation index of the equipment operating status and the preset fault threshold into the trained optimal deep forest model to obtain a trained operating status prediction model; the trained operating status prediction model is used to: obtain the current operating status value of the target centrifugal compressor based on the flow field data of the target centrifugal compressor output by the trained optimal deep forest model and the new evaluation index of the equipment operating status; judge the fouling fault degree of the equipment based on the preset fault threshold according to the current operating status value of the equipment; the current operating status value is represented by the pressure ratio-corrected volume flow curve composed of the pressure ratio, corrected volume flow and efficiency in the flow field data, and the area enclosed by the pressure ratio axis.
[0053] Among them, the formula expression of the new evaluation index of the equipment operation status is:
[0054]
[0055] Among them, ΔS is a new evaluation index for equipment operation status, S 预测 is the area enclosed by the pressure ratio-corrected volume flow curve and the pressure ratio axis obtained based on the predicted flow field data; 实际 It is the area enclosed by the pressure ratio-corrected volume flow curve obtained based on the actual flow field data and the pressure ratio axis.
[0056] Specifically, the S 实际 The solution expression is:
[0057]
[0058] Among them, n represents the number of vertices, x i ,y i is the x-coordinate and y-coordinate of the i-th vertex, x i+1 ,y i+1 is the x-coordinate and y-coordinate of the i+1th vertex, x n ,y n are the x and y coordinates of the nth vertex.
[0059] After executing step 301, the method further includes:
[0060] The data is preprocessed, including outlier detection, data normalization, feature engineering, etc., to ensure data quality and the effectiveness of model training. The preprocessed data will be divided into training set, validation set and test set in a ratio of 8:2 for subsequent model training and evaluation.
[0061] Specifically, different surface roughness and static pressure of the impeller and diffuser are used as input data (samples), and sample corrected volume flow rate, sample pressure ratio and sample isentropic efficiency are used as output data. The pressure ratio is the total-total pressure ratio, which refers to the ratio of the total pressure at the device outlet to the total pressure at the device inlet.
[0062] In the samples obtained, the roughness settings of the impeller and diffuser are both in the range of 1μm to 150μm, with a total of 100 groups of examples. The roughness settings of the impeller are specifically set to 1, 5, 10, 15, 25, 35, 50, 75, 100 and 150; the roughness settings of the diffuser are the same as the impeller. The details are shown in the following table:
[0063] Table 1 Roughness settings
[0064]
[0065]
[0066] When executing step 302, the specific steps may be as follows:
[0067] A deep forest model is constructed and trained using grid search and cross validation methods to systematically adjust the model's hyperparameters to obtain the model with the highest prediction accuracy. Grid search will traverse a series of preset hyperparameter combinations, while cross validation will evaluate the performance of each parameter combination by dividing the training set and validation set multiple times, and finally select the optimal parameter settings.
[0068] Specifically, during construction, determine the model structure and hyperparameter range, select the appropriate deep forest model structure, and determine the hyperparameters that need to be optimized. Data preprocessing, standardize or normalize data. Grid search traverses all preset hyperparameter combinations, trains the model and evaluates performance. Cross-validation re-divides the training set and validation set, and takes the average of multiple results. Select the optimal parameter combination, train the final model and use the test set to evaluate the generalization performance.
[0069] When the model structure is determined, the model structure selects the adaptive deep forest model, determines the appropriate number of layers of the cascade structure, and the number of random forests contained in each layer. The maximum number of layers of the cascade structure is 2. The initial number of random forests is 130 layers. After each layer after the initial layer, the number of trees will increase by "2*number of layers", that is, the number of trees in the random forest will increase by 2 for each additional layer. For example: the number of trees in the first layer is 130; the number of trees in the second layer is 130+2*1=132, and so on.
[0070] After executing step 302, data preprocessing is also included, which can be as follows:
[0071] In data preprocessing, in order to eliminate the dimensional differences between different feature data, standardization is selected to preprocess the data. It adjusts the scale of the data to make each feature comparable in statistical processing, thereby ensuring the equal weight contribution of different features in the model. Standardization not only helps to optimize the numerical stability during model training, but also accelerates the convergence of the model. The standardization formula is as follows:
[0072] X`=(X-μ) / σ.
[0073] Among them, X represents the original data, X` represents the standardized data, μ is the mean of X, and σ is the standard deviation of X.
[0074] When executing step 303, grid search and cross validation are used to train the deep forest model, as follows:
[0075] Determine the hyperparameters as the number of trees in each random forest, the maximum depth of each tree, the minimum number of samples in leaf nodes, and the number of forests in each cascade structure, and set a reasonable search range for each hyperparameter. And use the mean square error to evaluate the performance index corresponding to each parameter combination. The mean square error formula is as follows:
[0076]
[0077] Where n is the number of samples; y i Represents the true value of the i-th sample; Represents the predicted value of the i-th sample.
[0078] When performing cross-validation, the cross-validation method uses k-fold cross-validation, and k is set to 5. The specific steps are: randomly divide the data set into k subsets of equal size; each time select a subset as the validation set, and the remaining k-1 subsets as the training set. Repeat k times, using a different subset as the validation set each time. The final performance evaluation result is the average of the k validation results.
[0079] The roughness at the impeller, the roughness at the diffuser, and the static pressure were selected as input data; the corrected volume flow rate, pressure ratio, and efficiency were selected as output data. The optimal model was verified using the validation set data, and the accuracy of each output data was verified using the mean square error.
[0080] The further validation set data construction is shown in Table 2 below:
[0081] Table 2 Validation set data table
[0082]
[0083] Furthermore, the prediction results of the validation set data are as follows Figure 3 and Figure 4 As shown:
[0084] According to the comparison between the actual results and the predicted results of verification cases 1, 2, and 3, the corrected volume flow rate, pressure ratio, and efficiency prediction errors of the three verification cases are shown in Table 3 below:
[0085] Table 3 Corrected volume flow rate, pressure ratio and efficiency prediction error table
[0086] MSE Corrected volume flow Pressure ratio efficiency Verification Case 1 0.000588 3.0e-06 0.0115 Verification Case 2 0.000746 6.94e-06 0.0142 Verification Case 3 0.00143 2.46e-06 0.005
[0087] Among them, Figure 3 On the PV (pressure ratio-corrected volume flow) diagram in the figure, select a point with an efficiency of at least 60% to the highest efficiency point of the equipment to obtain the pressure ratio-corrected volume flow curve. The area enclosed by the curve and the Y axis (here represents the pressure ratio axis) can be used to construct a new physical quantity S (the current operating status value of the equipment). Figure 5 As shown in Figure 1, a challenge often faced when performing model evaluation is how to accurately assess the overall prediction accuracy of the model from multiple output variables. For example, if a model predicts three different outputs, such as modified volume flow, pressure ratio, and efficiency, each output will have its own mean squared error (MSE). These errors measure the difference between the model prediction and the actual data, but considering these errors separately can complicate the evaluation process and may lead to misinterpretation of model performance because the magnitude and importance of each metric may be different.
[0088] To solve this problem, this embodiment adopts a unified new evaluation index ΔS to combine the prediction errors of multiple outputs into a comprehensive index. The purpose of this is to provide a more comprehensive perspective to evaluate the overall performance of the model, while reducing the interference and misleading nature of a single output error index.
[0089]
[0090] Furthermore, since the actual compression ratio-efficiency curve is not a complete arc, but is composed of coordinate points, this patent uses the Gaussian area formula to solve S. The Gaussian area formula is commonly known as the shoelace formula and can be used to calculate the area of a polygon. For example, given the vertex coordinates of a polygon (x0, y0), (x1, y2), ..., (x n ,y n ),(x n+1 ,y n+1 ), connect these vertices in order to form an irregular closed polygon, and the shoelace formula can calculate the actual area S of the polygon:
[0091]
[0092] Among them, n represents the number of vertices, x i ,y i is the x-coordinate and y-coordinate of the i-th vertex, x i+1 ,y i+1 is the x-coordinate and y-coordinate of the i+1th vertex, x n ,y n is the x-coordinate and y-coordinate of the nth vertex. The prediction results obtained based on the new evaluation index are shown in Table 4 below:
[0093] Table 4 Prediction results
[0094] <![CDATA[S 实际 ]]> <![CDATA[S 预测 ]]> ΔS Verification Case 1 7.88 7.83 0.00635 Verification Case 2 8.77 8.83 0.00684 Verification Case 3 9.43 9.45 0.00212
[0095] Wherein, when executing step 304, judging the fouling fault degree of the equipment according to the current running state value of the equipment and based on the preset fault threshold, it specifically includes:
[0096] When the current operating state value of the equipment is greater than the first fault threshold, it is determined that the target centrifugal compressor is in a normal operating state.
[0097] When the current operating state value of the equipment is greater than or equal to the second fault threshold and less than the first fault threshold, it is determined that the target centrifugal compressor is in a minor fault state.
[0098] When the current operating state value of the equipment is greater than or equal to the third fault threshold and less than the second fault threshold, it is determined that the target centrifugal compressor is in a common fault state.
[0099] When the current operating state value of the equipment is less than the third fault threshold, it is determined that the target centrifugal compressor is in a serious fault state.
[0100] Specifically, when the roughness at the impeller is 15 μm and the roughness at the diffuser is 15 μm, S is 9.81;
[0101] When the roughness at the impeller is 50 μm and the roughness at the diffuser is 50 μm, S is 8.35;
[0102] When the roughness at the impeller is 75 μm and the roughness at the diffuser is 75 μm, S is 7.87.
[0103] Therefore, when S is set to be greater than 9.81: the compressor is in normal condition.
[0104] When S is greater than or equal to 8.35 and less than 9.81: it indicates a minor fault.
[0105] When S is greater than or equal to 7.87 and less than 8.35: compressor failure.
[0106] When S is less than or equal to 7.87: indicates a serious fault.
[0107] Specifically, the new physical quantity S can be expressed as shown in Table 5 below:
[0108] Table 5 Fault Interpretation Table
[0109]
[0110] By evaluating the impact of roughness on the performance of the compressor impeller and diffuser, different operating conditions can be classified according to the preset roughness threshold. According to this framework, the roughness index of verification case 1 indicates that it is in a faulty state, which is defined by the standard that S exceeds 7.87 but does not reach 8.35. At the same time, the S of verification cases 2 and 3 fall in the range of 8.35 to 9.81, respectively, indicating that both cases are in a slight fault state. This classification is consistent with the actual observations, confirming the effectiveness of using the new evaluation index as a method to judge the operating state of the compressor.
[0111] In some exemplary embodiments, Figure 6 As shown, a method based on the deep forest model combined with the new evaluation index is provided to realize compressor fouling fault warning. The specific steps are as follows:
[0112] A1. Input of original data:
[0113] Load and read the raw data table. The input raw data D = [X, Y, Z] is a set of three-dimensional data sequences, representing the impeller roughness, diffuser roughness and static pressure respectively. Ensure that the data is loaded correctly and prepare for subsequent analysis.
[0114] A2. Data preprocessing:
[0115] The original data D is cleaned. The specific steps include removing duplicate data rows, filling missing values, and identifying and processing outliers. The purpose of data cleaning is to ensure the integrity and consistency of the data and provide high-quality data input for the model.
[0116] A3. Construction of deep forest model:
[0117] According to the characteristics of the data, the adaptive deep forest model is constructed. First, the number of layers in the cascade structure and the number of random forests contained in each layer are determined. These structural parameters will determine the complexity and adaptability of the model.
[0118] A4. Grid search and cross validation:
[0119] Select hyperparameters for the model, including the number of trees per random forest, the maximum depth of the tree, the minimum number of samples per leaf node, and the number of forests in each cascade structure. Use grid search to find the best combination within the set parameter range, and evaluate model performance through 5-fold cross validation. The average of the cross validation results will be used as the measure of the final performance of the model.
[0120] A5. Optimal model output:
[0121] The model is optimized and adjusted based on the results of the validation set, and the final model performance is evaluated using the test set to ensure that it meets the expected indicators. The final output model is the training result based on the optimal parameters and structure.
[0122] B1. Introduction of new evaluation indicators:
[0123] New evaluation indicators are introduced to comprehensively consider the modified volume flow rate, pressure ratio and efficiency to provide a more comprehensive assessment of system performance.
[0124] A6. Prediction result output:
[0125] Based on the new evaluation index, a new physical quantity S is output for evaluating the current operating status of the equipment.
[0126] B2. Introduction of fault threshold:
[0127] The preset fault thresholds are imported into the model, and the prediction results S are matched with these thresholds to determine whether the equipment is in normal operation.
[0128] A7. Fault warning:
[0129] Compare the current operating status S of the device with the set fault threshold. If it exceeds a certain fault range, the fault warning level of the device at this moment is output to indicate the possible fault risk of the device.
[0130] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 7 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store fault judgment results. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, an early warning method for fouling failure of a centrifugal compressor is implemented.
[0131] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0132] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.
[0133] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.
[0134] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.
[0135] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.
[0136] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
[0137] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.
[0138] In summary, this application has the following technical effects:
[0139] 1) The fault warning method provided in this application can objectively evaluate the operating status of the compressor through quantified roughness values, reduce the interference of subjective judgment, and improve the accuracy of the evaluation.
[0140] 2) The fault warning method provided in this application can capture slight changes or anomalies in system performance earlier, thereby helping to identify potential faults or performance degradation in advance.
[0141] 3) The new evaluation indicators provided in this application are based on specific physical or engineering principles, which can better explain the model or system behavior. And they can integrate multiple factors to provide a more comprehensive evaluation of system performance.
[0142] 4) The new evaluation indicators provided in this application can improve the accuracy of fault warning, reduce false alarms and missed alarms, and help maintenance teams make more effective decisions.
[0143] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0144] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.
Claims
1. A method for early warning of fouling failure of a centrifugal compressor, characterized in that: The early warning method for the fouling failure of a centrifugal compressor comprises: Obtaining the surface roughness and static pressure of components in a target centrifugal compressor; the component surface roughness includes the impeller surface roughness and the diffuser surface roughness; The surface roughness and static pressure of the components are input into a trained operating status prediction model to obtain the fouling failure degree of the target centrifugal compressor; the trained operating status prediction model is a prediction model based on a new evaluation index of the equipment operating status; the new evaluation index of the equipment operating status is an evaluation index constructed based on predicted flow field data and actual flow field data; the flow field data includes a corrected volume flow rate, a pressure ratio and a compression efficiency.
2. The early warning method for centrifugal compressor fouling failure according to claim 1, characterized in that: The training process of the running status prediction model is as follows: Designing the surface roughness and sample static pressure of several sample parts through CFD simulation, and obtaining sample flow field data corresponding to the surface roughness and sample static pressure of the sample parts; the surface roughness of the sample parts includes the surface roughness of the sample impeller and the surface roughness of the sample diffuser; The sample flow field data includes sample corrected volume flow rate, sample pressure ratio and sample isentropic efficiency; Taking the surface roughness of sample parts and sample static pressure as input, and taking the sample flow field data corresponding to the surface roughness of sample parts and sample static pressure as labels, the deep forest model is trained to obtain a trained optimal deep forest model; the trained optimal deep forest model is used to output the flow field data of the target centrifugal compressor; The new evaluation index of the equipment operation state and the preset fault threshold are introduced into the trained optimal deep forest model to obtain a trained operation state prediction model; the trained operation state prediction model is used to: obtain the current operation state value of the target centrifugal compressor based on the flow field data of the target centrifugal compressor output by the trained optimal deep forest model and the new evaluation index of the equipment operation state; and judge the fouling fault degree of the equipment based on the preset fault threshold according to the current operation state value of the equipment; The current operating state value is represented by the area enclosed by the pressure ratio-corrected volume flow curve composed of the pressure ratio, efficiency and corrected volume flow in the flow field data and the pressure ratio axis.
3. The early warning method for centrifugal compressor fouling failure according to claim 2, characterized in that: The formula expression of the new evaluation index of the equipment operation status is: Among them, ΔS is a new evaluation index for equipment operation status, S 预测 is the area enclosed by the pressure ratio-corrected volume flow curve and the pressure ratio axis obtained based on the predicted flow field data; the S 实际 It is the area enclosed by the pressure ratio-corrected volume flow curve obtained based on the actual flow field data and the pressure ratio axis.
4. The early warning method for centrifugal compressor fouling failure according to claim 3, characterized in that: The S 实际 The solution expression is: Among them, n represents the number of vertices, x i ,y i is the x-coordinate and y-coordinate of the i-th vertex, x i+1 ,y i+1 is the x-coordinate and y-coordinate of the i+1th vertex, x n ,y n are the x and y coordinates of the nth vertex.
5. The early warning method for centrifugal compressor fouling failure according to claim 4, characterized in that: According to the current operating state value of the equipment and based on the preset fault threshold, the fouling fault degree of the equipment is judged, specifically including: When the current operating state value of the device is greater than the first fault threshold, it is determined that the target centrifugal compressor is in a normal operating state; When the current operating state value of the equipment is greater than or equal to the second fault threshold and less than the first fault threshold, the target centrifugal compressor is judged to be in a minor fault state; When the current operating state value of the equipment is greater than or equal to the third fault threshold and less than the second fault threshold, it is determined that the target centrifugal compressor is in a common fault state; When the current operating state value of the equipment is less than the third fault threshold, it is determined that the target centrifugal compressor is in a serious fault state.
6. The early warning method for centrifugal compressor fouling failure according to claim 5, characterized in that: The pressure ratio is a total-total pressure ratio, which refers to the ratio of the total pressure at the device outlet to the total pressure at the device inlet.
7. The early warning method for centrifugal compressor fouling failure according to claim 6, characterized in that: The roughness setting range of the sample impeller surface roughness among the surface roughness of the plurality of sample parts is 1 to 150 μm; the roughness setting range of the sample diffuser surface roughness is 1 to 150 μm.
8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement an early warning method for fouling failure of a centrifugal compressor as described in any one of claims 1-7.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, an early warning method for fouling failure of a centrifugal compressor according to any one of claims 1 to 7 is implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, an early warning method for fouling failure of a centrifugal compressor according to any one of claims 1 to 7 is implemented.
Citation Information
Patent Citations
Gas turbine gas compressor blade scale deposition state monitoring method and device
CN113847107A
Method for studying influence of blade scale deposition on performance of multistage axial flow compressor
CN117057063A
Gas turbine compressor blade scale deposition degree grading diagnosis method and device
CN118293088A
Multi-scale modeling method for compressor blade scale deposition
CN118504133A
On-line prediction method of surface roughness of parts based on SDAE-DBN algorithm
US20210287098A1
Cited By
Wall adhesion quantitative evaluation method and system for spray drying tower
CN120448987A
