Membrane component data management system and method based on artificial intelligence

Abstract

The invention discloses a membrane component data management system and method based on artificial intelligence, and relates to the technical field of membrane component data management.The method comprises the steps that historical operation data of a membrane component is obtained, and the historical operation data of the membrane component is preprocessed; analyzing data features based on the preprocessed time series of each operation parameter of the membrane component, building a multi-dimensional feature vector, analyzing a reconstruction error of the multi-dimensional feature vector, and building a health model based on the reconstruction error; constructing a service life analysis data pair, and constructing a corresponding service life prediction model based on the service life analysis data pair; and obtaining life prediction model parameters of related historical membrane components of the to-be-predicted membrane component, constructing a real-time life prediction model of the to-be-predicted membrane component, and performing early warning analysis according to the real-time residual life and the health index of the to-be-predicted membrane component. Dynamic updating of parameters is realized, and the accuracy of real-time residual life prediction is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of membrane component data management, and specifically to an artificial intelligence-based membrane component data management system and method. Background Art

[0002] In the thermoelectric industry, the demineralized water process is an important link to ensure the normal operation of equipment and extend its service life. The demineralized water process mainly removes salts and other impurities in water through membrane separation technologies such as reverse osmosis (RO) membranes and nanofiltration (NF) membranes. As the core component of the demineralized water process, the performance of the membrane component directly affects the efficiency and quality of water treatment. The performance of the membrane component will decline with the increase of the use time, which may lead to unqualified water quality or equipment failure.

[0003] Therefore, timely diagnosing the state of the membrane component, performing maintenance and replacement are the keys to ensuring the normal operation of the demineralized water process. The degradation of the membrane component is affected by multiple factors, and the degradation process is usually non-linear, which cannot be accurately described by the existing models. The existing models cannot update parameters in real time, making it difficult to adapt to the dynamic changes of the equipment state, resulting in low prediction accuracy.

[0004] Therefore, the present invention discloses an artificial intelligence-based membrane component data management system and method to solve the above problems. Summary of the Invention

[0005] The purpose of the present invention is to provide an artificial intelligence-based membrane component data management system and method to solve the problems raised in the prior art.

[0006] To achieve the above purpose, the present invention provides the following technical solution: An artificial intelligence-based membrane component data management method, which includes the following steps:

[0007] S1: Obtain the historical operation data of the membrane component, preprocess the historical operation data of the membrane component, achieve spatio-temporal alignment of the operation parameters with different collection cycles, and perform spatial interpolation;

[0008] S2: Analyze the data characteristics of the preprocessed time series of each operation parameter of the membrane component, form a multi-dimensional feature vector, analyze the reconstruction error of the multi-dimensional feature vector, and build a health model based on the reconstruction error;

[0009] S3: Construct a life analysis data pair based on the historical operation data of the whole process of the membrane component from the start of use state to the failure state, and build a corresponding life prediction model based on the life analysis data pair;

[0010] S4: Obtain the life prediction model parameters of the historical membrane component related to the membrane component to be predicted, build a real-time life prediction model of the membrane component to be predicted, and perform early warning analysis according to the real-time remaining life and health index of the membrane component to be predicted.

[0011] According to the above solution, in S1, the following content is included:

[0012] S101: Obtain the historical operation data of the membrane component, generate the time series of each operation parameter, where the operation parameters include online monitoring data and offline detection data; the online monitoring data includes transmembrane pressure difference, water production quality, flow rate, and inlet water temperature; the water production quality is equal to the conductivity of the produced water divided by the turbidity of the produced water; the offline detection data includes the membrane surface topography image and the pollutant component concentration; the acquisition periods of each operation parameter are different and preset by the system; the offline detection data is laboratory detection data;

[0013] S102: Binarize the membrane surface topography image, perform full coverage of the foreground area based on the preset window set of the system, and count the number of windows of each size that complete the full coverage of the foreground area; form data pairs based on the size of each window and the corresponding number of windows, perform linear fitting based on the data pairs, and record the slope of the fitting line as the image feature value;

[0014] S103: Extract the time series of any two operation parameters, denoted as X = {x i |i ∈ [1, I]} and Y = {y j |j ∈ [1, J]}, where x i represents the i-th operation parameter value in the time series X, y j represents the j-th operation parameter value in the time series Y, I represents the number of values in the time series X, and J represents the number of values in the time series Y; the acquisition frequency of the operation parameters in the time series Y is lower than that of the operation parameters in the time series X; construct an accumulated distance matrix D of size I × J, and initialize all elements in the accumulated distance matrix D to 0; for each matrix element D(i, j), calculate the corresponding distance d(i, j) and fill the accumulated distance matrix, and terminate the calculation if the accumulated distance is greater than the preset threshold; where D(i, j) represents the minimum distance from the start point of the sequence to x i and y j , and d(i, j) is the Euclidean distance between x i and y j ; start from D(N, M), backtrack along the recursive path to (1, 1), and record the shortest path; according to the obtained shortest path, interpolate the operation parameter values in the time series Y on the time scale of the time series X to make them aligned, and use the Kriging method for spatial interpolation; traverse all types of operation parameters to form the extended time series of each operation parameter.

[0015] This application preprocesses the time series of operating parameters to achieve time alignment, solves the problem of inconsistent time scales between sensor data and off-line detection data, eliminates the need to unify the acquisition periods of individual sensors, and reduces overall detection resources. When filling the cumulative distance matrix, if the cumulative distance is greater than a preset threshold, the calculation is terminated, improving the overall analysis efficiency.

[0016] According to the above solution, in S2, the following content is included:

[0017] S201: Obtain the extended time series of transmembrane pressure difference, analyze the coefficient of variation CV of the transmembrane pressure difference, where the coefficient of variation is equal to the standard deviation of the extended time series of transmembrane pressure difference divided by the mean of the extended time series of transmembrane pressure difference; analyze the theoretical value of membrane flux based on the extended time series of transmembrane pressure difference and Darcy's law, and generate the residual feature ε by analyzing the theoretical value and actual value of membrane flux; obtain the extended time series of membrane flux and use Fourier transform to analyze the spectral energy E and the main frequency component f dominant ;

[0018] S202: Obtain the extended time series of water production quality, flow rate, and inlet temperature, and analyze the mean and standard deviation of each extended time series; based on the coefficient of variation CV of the transmembrane pressure difference, the mean and standard deviation of each extended time series, the residual feature ε, the spectral energy E, and the main frequency component f dominant Construct a multi-dimensional feature vector A, A = [CV, μ T , σ T , μ Q , σ Q , μ C , σ C , ε, E, f dominant ; where μ T represents the mean of the extended time series of inlet temperature, and σ T represents the standard deviation of the extended time series of inlet temperature; μ Q represents the mean of the extended time series of flow rate, and σ Q represents the standard deviation of the extended time series of flow rate; μ C represents the mean of the extended time series of water production quality, and σ C represents the standard deviation of the extended time series of water production quality;

[0019] The performance of the membrane component is the result of the combined action of multiple operating parameters. By performing Fourier transform only on the comprehensive performance index, the periodic characteristics of the combined action of multiple operating parameters can be captured, ensuring the analysis effect while reducing the computational complexity, avoiding the redundancy and errors caused by analyzing individual operating parameters separately. The frequency domain features (such as spectral energy and main frequency component) extracted by one Fourier transform have clear physical meanings and can effectively support the health status assessment and fault diagnosis of the membrane component.

[0020] S203: Compress the multi-dimensional feature vector A to form a latent space representation, and then reconstruct the latent space representation into the original data; the encoder expression corresponding to compressing the multi-dimensional feature vector A to form a latent space representation is h = Ψ(W e A + b e ); where h represents the latent space representation, W e represents the encoder weight matrix, b e represents the encoder bias vector, and Ψ() represents the activation function ReLU; the decoder expression corresponding to reconstructing the latent space representation into the original data is A' = Ψ(W d h + b d ); where A' represents the reconstructed feature vector; W d represents the decoder weight matrix; b d represents the decoder bias vector; the weight matrices W e and W d are initialized by randomly sampling from a normal distribution with a mean of 0 and a variance of 2 / N, where N is the number of neurons in the previous layer; the bias vectors b e and b d are initialized as zero vectors;

[0021] Optimize the parameters of the encoder and decoder through backpropagation to minimize the reconstruction error; where the loss function is the mean squared error; construct a health model HI = 1 - L AE / L AE max ; where L AE max represents the maximum reconstruction error of the training set, and L AE represents the reconstruction error of the current sample.

[0022] This application compresses high-dimensional input data into a low-dimensional latent representation through an encoder, extracts key features in the data; learns the essential features of the data, and removes noise in the sensor data; the denoised data can be used for analysis to enhance the accuracy of remaining useful life prediction;

[0023] According to the above solution, in S3, it includes the following content:

[0024] S301: Obtain the historical operation data of the membrane component from the start-of-use state to the failure state, record the remaining useful life corresponding to the failure state as 0, and construct a life analysis data pair (t, HI t , RUL t ) based on the historical operation data, where t represents time, HI t represents the health index corresponding to time t, and RUL trepresents the remaining useful life corresponding to time t; construct a life prediction model based on life analysis data: dRUL t / dt = -k × exp(β × HI t ); where k represents the degradation rate coefficient, β represents the health index influence coefficient, and exp() represents the exponential function with the natural number as the base;

[0025] S302: Discretize the life prediction model into a difference equation: RUL t+1 = RUL t - Δt × k × exp(β × HI t ), where Δt represents the time step, and the initial values of k and β are preset by the system; substitute the life analysis data pairs corresponding to the historical operation data into the discretized difference equation, calculate the prediction error corresponding to each time node, the prediction error is equal to the predicted remaining useful life minus the true useful life, analyze the total error based on the prediction error, and the total error is equal to the sum of the squares of the prediction errors at all time points; use the gradient descent method to update the degradation rate coefficient and the health index influence coefficient until the total error reaches the minimum value or the maximum number of iterations is reached; bind the latest degradation rate coefficient and health index influence coefficient to the corresponding membrane component.

[0026] This application constructs life analysis data pairs based on historical operation data and constructs a life prediction model, combining the advantages of physical models and data-driven models to reduce prediction errors; in the case of insufficient data, the physical model provides reliable support and reduces the dependence on data.

[0027] According to the above solution, in S4, it includes the following content:

[0028] S401: Extract the degradation rate coefficient and health index influence coefficient corresponding to a preset number of historical membrane components, calculate the average value of the degradation rate coefficient and the average value of the health index influence coefficient of the historical membrane components, and use them as the initial values of the degradation rate coefficient and the health index influence coefficient of the membrane component to be predicted; the production coordinates corresponding to the extracted preset number of historical membrane components are the same as those of the membrane component to be predicted;

[0029] S402: Construct a real-time life prediction model dRUL / dt = -k × exp(β × HI) based on the initial values of the degradation rate coefficient and the health index influence coefficient of the membrane component to be predicted; use the Kalman filter to update the degradation rate coefficient and the health index influence coefficient of the membrane component to be predicted in real time; predict the real-time remaining useful life based on the real-time operation data of the membrane component to be predicted;

[0030] S403: If the real-time remaining life of the membrane component to be predicted is less than the corresponding threshold and the health index of the membrane component to be predicted is less than the corresponding threshold, issue a first-level warning signal; if the real-time remaining life of the membrane component to be predicted is less than the corresponding threshold and the health index of the membrane component to be predicted is greater than or equal to the corresponding threshold, issue a second-level warning signal; if the real-time remaining life of the membrane component to be predicted is greater than or equal to the corresponding threshold and the health index of the membrane component to be predicted is less than the corresponding threshold, issue a third-level warning signal; if the real-time remaining life of the membrane component to be predicted is greater than or equal to the corresponding threshold and the health index of the membrane component to be predicted is greater than or equal to the corresponding threshold, do not issue a warning signal;

[0031] When the system identifies a first-level warning signal, send an instruction to notify the administrator to replace the membrane component; when the system identifies a second-level warning signal, send an instruction to notify the administrator to perform routine maintenance on the membrane component; when the system identifies a third-level warning signal, send an instruction to notify the administrator to perform routine maintenance or repair on the membrane component.

[0032] This application can improve the accuracy of prediction by borrowing the model coefficients of historical membrane components to replace the preset model coefficients of the membrane components to be predicted, dynamically update the parameters, and further improve the accuracy of real-time remaining life prediction.

[0033] Another aspect of this application provides an artificial intelligence-based membrane component data management system. The system is implemented by applying the above-mentioned artificial intelligence-based membrane component data management method. The system includes a data preprocessing module, a health analysis module, a life prediction model construction module, and a real-time warning module;

[0034] The data preprocessing module is used to obtain the historical operation data of the membrane component, preprocess the historical operation data of the membrane component, achieve spatio-temporal alignment of the operation parameters in different acquisition cycles, and perform spatial interpolation;

[0035] The health analysis module analyzes the data characteristics of the time series after preprocessing of each operation parameter of the membrane component, forms a multi-dimensional feature vector, analyzes the reconstruction error according to the multi-dimensional feature vector, and constructs a health model based on the reconstruction error;

[0036] The life prediction model construction module constructs a life analysis data pair according to the historical operation data of the whole process of the membrane component from the start-of-use state to the failure state, and constructs a corresponding life prediction model based on the life analysis data pair;

[0037] The real-time warning module obtains the life prediction model parameters of the historical membrane components related to the membrane component to be predicted, constructs a real-time life prediction model of the membrane component to be predicted, and performs warning analysis according to the real-time remaining life and health index of the membrane component to be predicted.

[0038] According to the above solution, the data preprocessing module includes a data acquisition unit and a spatio-temporal alignment and augmentation unit;

[0039] The data acquisition unit is used to obtain the historical operation data of the membrane component, generate time series of each operation parameter; perform image binarization on the membrane surface topography image, fully cover the foreground area based on a preset window set of the system, and count the number of windows of each size that complete the full coverage of the foreground area; form data pairs based on the size of each window and the corresponding number of windows, perform linear fitting on the data pairs, and record the slope of the fitting line as the image feature value;

[0040] The spatio-temporal alignment and augmentation unit is used to extract the time series of any two operation parameters, analyze the shortest path of each operation parameter value in the two time series, perform interpolation based on the shortest path, and perform spatial interpolation using the Kriging method; traverse all types of operation parameters to form augmented time series of each operation parameter.

[0041] According to the above solution, the health analysis module includes a feature analysis unit and a health model construction unit;

[0042] The feature analysis unit is used to obtain the augmented time series of the transmembrane pressure difference, and analyze the coefficient of variation of the transmembrane pressure difference; analyze the theoretical value of the membrane flux according to the augmented time series of the transmembrane pressure difference and Darcy's law, and analyze the residual characteristics by comparing the theoretical value and the actual value of the membrane flux; perform Fourier transform on the augmented time series of the membrane flux, and analyze the spectral energy and main frequency components; obtain the augmented time series of the produced water quality, flow rate, and inlet water temperature, and analyze the mean and standard deviation of each augmented time series;

[0043] The health model construction unit constructs a multi-dimensional feature vector based on the coefficient of variation of the transmembrane pressure difference, the mean and standard deviation of each augmented time series, the residual characteristics, the spectral energy, and the main frequency components; compresses the multi-dimensional feature vector to form a latent space representation, and then reconstructs the latent space representation into the original data; optimizes the parameters of the encoder and decoder through backpropagation to minimize the reconstruction error; constructs a health model based on the reconstruction error.

[0044] According to the above solution, the life prediction model construction module includes a life prediction model construction unit and a parameter optimization unit;

[0045] The life prediction model construction unit obtains the historical operation data of the whole process of the membrane component from the start of use to the failure state, constructs life analysis data pairs based on the historical operation data, and constructs a life prediction model according to the life analysis data pairs;

[0046] The parameter optimization unit is used to discretize the life prediction model into a difference equation; update the degradation rate coefficient and the health index influence coefficient using the gradient descent method until the total error reaches the minimum value or the maximum number of iterations is reached; and bind the latest degradation rate coefficient, the health index influence coefficient, and the affiliated membrane component.

[0047] According to the above solution, the real-time warning module includes a real-time model construction unit and a warning unit;

[0048] The real-time model construction unit is used to extract the degradation rate coefficient and the health index influence coefficient corresponding to a preset number of historical membrane components, calculate the average value of the degradation rate coefficient and the average value of the health index influence coefficient of the historical membrane components, and use them as the initial values of the degradation rate coefficient and the health index influence coefficient of the membrane component to be predicted; the production coordinates extracted corresponding to the preset number of historical membrane components are the same as those of the membrane component to be predicted; construct a real-time life prediction model based on the initial values of the degradation rate coefficient and the health index influence coefficient of the membrane component to be predicted; and perform real-time remaining life prediction;

[0049] The warning unit classifies warning signals according to the real-time remaining life and the health index, and gives system prompts according to the warning signal level.

[0050] Compared with the prior art, the beneficial effects of the present invention are as follows: By preprocessing the time series of operating parameters, this application achieves time alignment; solves the problem of inconsistent time scales between sensor data and off-line detection data, does not require uniform sampling periods for each sensor, and reduces overall detection resources; when filling the cumulative distance matrix, if the cumulative distance is greater than the preset threshold, the calculation is terminated; improves the overall analysis efficiency; by performing Fourier transform only on the comprehensive performance index, the periodic characteristics of the combined action of multiple operating parameters can be captured, reducing the computational complexity while ensuring the analysis effect; avoiding redundancy and errors caused by analyzing operating parameters alone; the frequency domain features extracted by one Fourier transform have clear physical meanings and can effectively support the health state assessment of membrane components; this application compresses high-dimensional input data into low-dimensional latent representations through an encoder, extracts key features in the data; learns the essential features of the data and removes noise in the sensor data; the denoised data can be used for analysis, enhancing the accuracy of remaining life prediction; this application constructs a life analysis data pair based on historical operating data and constructs a life prediction model, combining the advantages of physical models and data-driven models to reduce prediction errors; in the case of insufficient data, the physical model provides reliable support and reduces the dependence on data. This application can improve the prediction accuracy by borrowing the model coefficients of historical membrane components to replace the preset model coefficients of the membrane component to be predicted, dynamically update the parameters, and further improve the accuracy of real-time remaining life prediction. Description of the Drawings

[0051] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, but do not constitute a limitation to the present invention. In the accompanying drawings:

[0052] Figure 1 It is a schematic flowchart of a method for managing membrane component data based on artificial intelligence according to the present invention;

[0053] Figure 2 It is a schematic structural diagram of a system for managing membrane component data based on artificial intelligence according to the present invention. Detailed implementation manners

[0054] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0055] Please refer to Figure 1 , the present invention provides a technical solution: a method for managing membrane component data based on artificial intelligence, the method includes the following steps:

[0056] S1: Obtain the historical operation data of the membrane component, preprocess the historical operation data of the membrane component, achieve spatio-temporal alignment of the operation parameters in different acquisition cycles, and perform spatial interpolation;

[0057] In S1, the following contents are included:

[0058] S101: Obtain the historical operation data of the membrane component, generate time series of each operation parameter, and the operation parameters include on-line monitoring data and off-line detection data; the on-line monitoring data includes transmembrane pressure difference, water production quality, flow rate, and inlet water temperature; the water production quality is equal to the conductivity of the produced water divided by the turbidity of the produced water; the off-line detection data includes the membrane surface topography image and the pollutant component concentration; the acquisition cycles of each operation parameter are different and are preset by the system;

[0059] S102: Binarize the membrane surface topography image, fully cover the foreground area based on the preset window set of the system, and count the number of windows of each size that complete the full coverage of the foreground area; form data pairs based on the size of each window and the corresponding number of windows, perform linear fitting based on the data pairs, and record the slope of the fitting line as the image feature value;

[0060] S103: Extract the time series of any two operation parameters, denoted as X = {x i |i ∈ [1, I]} and Y = {y j|j ∈ [1, J]}, where x i represents the value of the i-th operating parameter in the time series X, and y j represents the value of the j-th operating parameter in the time series Y. I represents the number of values in the time series X, and J represents the number of values in the time series Y; the acquisition frequency of the operating parameters of the time series Y is lower than that of the time series X; construct an accumulated distance matrix D of size I × J, and the elements in the accumulated distance matrix D are all initialized to 0; for each matrix element D(i, j), calculate the corresponding distance d(i, j) and fill the accumulated distance matrix. If the accumulated distance is greater than the preset threshold, terminate the calculation; where D(i, j) represents the minimum distance from the start point of the sequence to x i and y j , and d(i, j) is the Euclidean distance between x i and y j ; starting from D(N, M), backtrack along the recurrence path to (1, 1) and record the shortest path; according to the obtained shortest path, interpolate the values of the operating parameters in the time series Y on the time scale of the time series X to make them aligned, and use the Kriging method for spatial interpolation; traverse all types of operating parameters to form the extended time series of each operating parameter.

[0061] S2: Based on the preprocessed time series analysis data features of each operating parameter of the membrane component, form a multi-dimensional feature vector, analyze the reconstruction error of the multi-dimensional feature vector, and construct a health model based on the reconstruction error;

[0062] In S2, the following contents are included:

[0063] S201: Obtain the extended time series of the transmembrane pressure difference, analyze the coefficient of variation CV of the transmembrane pressure difference. The coefficient of variation is equal to the standard deviation of the extended time series of the transmembrane pressure difference divided by the average value of the extended time series of the transmembrane pressure difference; analyze the theoretical value of the membrane flux according to the extended time series of the transmembrane pressure difference and Darcy's law, and analyze the residual feature ε generated by the theoretical value and the actual value of the membrane flux; obtain the extended time series of the membrane flux and use Fourier transform to analyze the spectral energy E and the main frequency component f dominant ;

[0064] S202: Obtain the extended time series of the water production quality, flow rate, and inlet temperature, and analyze the mean and standard deviation of each extended time series; based on the coefficient of variation CV of the transmembrane pressure difference, the mean and standard deviation of each extended time series, the residual feature ε, the spectral energy E, and the main frequency component f dominant construct a multi-dimensional feature vector A, A = [CV, μ T , σ T , μ Q , σ Q , μ C , σC , ε, E, f dominant ; where μ T represents the mean of the extended time series of the influent temperature, and σ T represents the standard deviation of the extended time series of the influent temperature; μ Q represents the mean of the extended time series of the flow rate, and σ Q represents the standard deviation of the extended time series of the flow rate; μ C represents the mean of the extended time series of the produced water quality, and σ C represents the standard deviation of the extended time series of the produced water quality;

[0065] S203: Compress the multi-dimensional feature vector A to form a latent space representation, and then reconstruct the latent space representation into the original data; the encoder expression for compressing the multi-dimensional feature vector A to form the latent space representation is h = Ψ(W e A + b e ); where h represents the latent space representation, W e represents the encoder weight matrix, b e represents the encoder bias vector, and Ψ() represents the activation function ReLU; the decoder expression for reconstructing the latent space representation into the original data is A' = Ψ(W d h + b d ); where A' represents the reconstructed feature vector; W d represents the decoder weight matrix; b d represents the decoder bias vector; the weight matrices W e and W d are initialized by randomly sampling from a normal distribution with a mean of 0 and a variance of 2 / N, where N is the number of neurons in the previous layer; the bias vectors b e and b d are initialized as zero vectors;

[0066] Optimize the parameters of the encoder and decoder through backpropagation to minimize the reconstruction error; where the loss function is the mean squared error; construct a health model HI = 1 - L AE / L AE max ; where L AE max represents the maximum reconstruction error of the training set, and L AE represents the reconstruction error of the current sample.

[0067] S3: Construct a life analysis data pair based on the historical operation data of the whole process of the membrane component from the start-of-use state to the failure state, and construct a corresponding life prediction model based on the life analysis data pair;

[0068] In S3, the following content is included:

[0069] S301: Obtain the historical operation data of the membrane component throughout the process from the start-of-use state to the failure state, record the remaining life corresponding to the failure state as 0, and construct a life analysis data pair (t, HI t , RUL t ) based on the historical operation data, where t represents time, HI t represents the health index corresponding to time t, and RUL t represents the remaining life corresponding to time t; construct a life prediction model according to the life analysis data pair: d RUL t / dt = -k × exp(β × HI t ); where k represents the degradation rate coefficient, β represents the health index influence coefficient, and exp() represents the exponential function with the natural number as the base;

[0070] S302: Discretize the life prediction model into a difference equation: RUL t+1 = RUL t - Δt × k × exp(β × HI t ), where Δt represents the time step, and the initial values of k and β are preset by the system; substitute the life analysis data pair corresponding to the historical operation data into the discretized difference equation, calculate the prediction error corresponding to each time node, the prediction error is equal to the predicted remaining life minus the true service life, analyze the total error based on the prediction error, and the total error is equal to the sum of the squares of the prediction errors at all time points; use the gradient descent method to update the degradation rate coefficient and the health index influence coefficient until the total error reaches the minimum value or the maximum number of iterations is reached; bind the latest degradation rate coefficient, health index influence coefficient and the corresponding membrane component.

[0071] Example 1: In this example, the life analysis data pairs are (0, 1.0, 1000), (100, 0.9, 900), (200, 0.8, 800), (300, 0.7, 700) and (400, 0.6, 600); where the units of time t and the remaining life RUL t are both hours;

[0072] In this example, the time step Δt = 100h, the initial parameters k = 0.01, β = 0.1;

[0073] Therefore, the discretized formula: RUL t+1 = RUL t - 100 × 0.01 × exp(0.1 × HI t );

[0074] For each time point t, calculate the prediction error e t = RUL true，T+1 - [RUL t-exp(0.1×HI t )];

[0075] Total error Use the gradient descent method for iterative optimization, and finally obtain k = 0.1 and β = 0.5;

[0076] S4: Obtain the life prediction model parameters of the historical membrane components related to the membrane component to be predicted, construct the real-time life prediction model of the membrane component to be predicted, and perform early warning analysis based on the real-time remaining life and health index of the membrane component to be predicted.

[0077] In S4, it includes the following content:

[0078] S401: Extract the degradation rate coefficients and health index influence coefficients corresponding to a preset number of historical membrane components, calculate the average values of the degradation rate coefficients and health index influence coefficients of the historical membrane components, and use them as the initial values of the degradation rate coefficients and health index influence coefficients of the membrane component to be predicted; extract the production coordinates corresponding to a preset number of historical membrane components to be the same as those of the membrane component to be predicted;

[0079] S402: Based on the initial values of the degradation rate coefficient and health index influence coefficient of the membrane component to be predicted, construct the real-time life prediction model dRUL / dt = -k×exp(β×HI); use the Kalman filter to update the degradation rate coefficient and health index influence coefficient of the membrane component to be predicted in real time; predict the real-time remaining life based on the real-time operation data of the membrane component to be predicted;

[0080] S403: If the real-time remaining life of the membrane component to be predicted is less than the corresponding threshold and the health index of the membrane component to be predicted is less than the corresponding threshold, issue a first-level warning signal; if the real-time remaining life of the membrane component to be predicted is less than the corresponding threshold and the health index of the membrane component to be predicted is greater than or equal to the corresponding threshold, issue a second-level warning signal; if the real-time remaining life of the membrane component to be predicted is greater than or equal to the corresponding threshold and the health index of the membrane component to be predicted is less than the corresponding threshold, issue a third-level warning signal; if the real-time remaining life of the membrane component to be predicted is greater than or equal to the corresponding threshold and the health index of the membrane component to be predicted is greater than or equal to the corresponding threshold, do not issue a warning signal;

[0081] When the system identifies a first-level warning signal, send an instruction to notify the administrator to replace the membrane component; when the system identifies a second-level warning signal, send an instruction to notify the administrator to perform routine maintenance on the membrane component; when the system identifies a third-level warning signal, send an instruction to notify the administrator to perform routine maintenance or repair on the membrane component.

[0082] Please refer to Figure 2, the present invention provides a technical solution: an artificial intelligence-based membrane component data management system, which includes a data preprocessing module, a health analysis module, a life prediction model construction module, and a real-time warning module;

[0083] The data preprocessing module is used to obtain the historical operation data of the membrane component, preprocess the historical operation data of the membrane component, achieve spatio-temporal alignment of operation parameters with different acquisition cycles, and perform spatial interpolation;

[0084] The health analysis module analyzes the data characteristics of the time series after preprocessing of each operation parameter of the membrane component, forms a multi-dimensional feature vector, analyzes the reconstruction error based on the multi-dimensional feature vector, and constructs a health model based on the reconstruction error;

[0085] The life prediction model construction module constructs a life analysis data pair according to the historical operation data of the whole process of the membrane component from the start of use to the failure state, and constructs a corresponding life prediction model based on the life analysis data pair;

[0086] The real-time warning module obtains the life prediction model parameters of the historical membrane components related to the membrane component to be predicted, constructs a real-time life prediction model of the membrane component to be predicted, and performs warning analysis according to the real-time remaining life and health indicators of the membrane component to be predicted.

[0087] The data preprocessing module includes a data acquisition unit and a spatio-temporal alignment and expansion unit;

[0088] The data acquisition unit is used to obtain the historical operation data of the membrane component and generate a time series of each operation parameter; perform image binarization on the membrane surface topography image, fully cover the foreground area based on the preset window set of the system, and count the number of windows of each size that complete the full coverage of the foreground area; form a data pair based on the size of each window and the corresponding number of windows, perform linear fitting on the data pair, and record the slope of the fitting line as the image feature value;

[0089] The spatio-temporal alignment and expansion unit is used to extract the time series of any two operation parameters, analyze the shortest path of the numerical values of each operation parameter in the two time series, perform interpolation based on the shortest path, and perform spatial interpolation using the Kriging method; traverse all types of operation parameters to form an extended time series of each operation parameter.

[0090] The health analysis module includes a feature analysis unit and a health model construction unit;

[0091] The feature analysis unit is used to obtain the extended time series of the transmembrane pressure difference and analyze the coefficient of variation of the transmembrane pressure difference; analyze the theoretical value of the membrane flux according to the extended time series of the transmembrane pressure difference and Darcy's law, and generate residual features by analyzing the theoretical value and the actual value of the membrane flux; obtain the extended time series of the membrane flux and use Fourier transform to analyze the spectral energy and the main frequency component; obtain the extended time series of the produced water quality, flow rate, and inlet water temperature, and analyze the mean and standard deviation of each extended time series.

[0092] The health model construction unit constructs a multi-dimensional feature vector based on the coefficient of variation of the transmembrane pressure difference, the mean and standard deviation of each extended time series, the residual features, the spectral energy, and the main frequency component; compresses the multi-dimensional feature vector to form a latent space representation, and then reconstructs the latent space representation into the original data; optimizes the parameters of the encoder and decoder through backpropagation to minimize the reconstruction error; constructs a health model based on the reconstruction error.

[0093] The service life prediction model construction module includes a service life prediction model construction unit and a parameter optimization unit.

[0094] The service life prediction model construction unit obtains the historical operation data of the whole process of the membrane component from the starting use state to the failure state, constructs a service life analysis data pair based on the historical operation data, and constructs a service life prediction model according to the service life analysis data pair.

[0095] The parameter optimization unit is used to discretize the service life prediction model into a difference equation; update the degradation rate coefficient and the health index influence coefficient using the gradient descent method until the total error reaches the minimum value or the maximum number of iterations is reached; bind the latest degradation rate coefficient, the health index influence coefficient, and the corresponding membrane component.

[0096] The real-time warning module includes a real-time model construction unit and a warning unit.

[0097] The real-time model construction unit is used to extract the degradation rate coefficient and the health index influence coefficient corresponding to a preset number of historical membrane components, calculate the average value of the degradation rate coefficient and the average value of the health index influence coefficient of the historical membrane components, and use them as the initial values of the degradation rate coefficient and the health index influence coefficient of the membrane component to be predicted; extract the production coordinates corresponding to a preset number of historical membrane components to be the same as those of the membrane component to be predicted; construct a real-time service life prediction model based on the initial values of the degradation rate coefficient and the health index influence coefficient of the membrane component to be predicted; and perform real-time remaining service life prediction.

[0098] The warning unit classifies the warning signals according to the real-time remaining service life and the health index, and gives system prompts according to the warning signal levels.

[0099] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0100] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention. Any reference signs in the claims should not be construed as limiting the claims concerned.

Claims

1. An artificial intelligence-based method for managing membrane component data, characterized in that, The method includes the following steps: S1: Obtain the historical operation data of the membrane component, preprocess the historical operation data of the membrane component, achieve spatio-temporal alignment of the operation parameters in different acquisition cycles, and perform spatial interpolation; S2: Based on the data feature analysis of the preprocessed time series of each operation parameter of the membrane component, form a multi-dimensional feature vector, analyze the reconstruction error of the multi-dimensional feature vector, and build a health model based on the reconstruction error; S3: Construct a life analysis data pair according to the historical operation data of the whole process of the membrane component from the start-of-use state to the failure state, and build a corresponding life prediction model based on the life analysis data pair; S4: Obtain the life prediction model parameters of the historical membrane component related to the membrane component to be predicted, construct a real-time life prediction model of the membrane component to be predicted, and perform early warning analysis according to the real-time remaining life and health index of the membrane component to be predicted.

2. The method for managing membrane component data based on artificial intelligence according to claim 1, characterized in that: In S-1, the following is included: S101: Obtain the historical operation data of the membrane component, generate the time series of each operation parameter, where the operation parameters include on-line monitoring data and off-line detection data; the on-line monitoring data includes transmembrane pressure difference, product water quality, flow rate, and inlet water temperature; the product water quality is equal to the product water conductivity divided by the product water turbidity; the off-line detection data includes the membrane surface topography image and the pollutant component concentration; the acquisition cycles of each operation parameter are different and preset by the system; S102: Binarize the membrane surface topography image, fully cover the foreground area based on the preset window set of the system, and count the number of windows of each size that complete the full coverage of the foreground area; Form a data pair based on the size of each window and the corresponding number of windows, perform linear fitting based on the data pair, and record the slope of the fitting line as the image feature value; S103: Extract the time series of any two operating parameters, denoted as X = {x i | i ∈ [1, I]} and Y = {y j | j ∈ [1, J]}, where x i represents the value of the i-th operating parameter in the time series X, y j represents the value of the j-th operating parameter in the time series Y, I represents the number of values in the time series X, and J represents the number of values in the time series Y; the acquisition frequency of the operating parameters in the time series Y is lower than that of the operating parameters in the time series X; construct an accumulated distance matrix D of size I × J, and the elements in the accumulated distance matrix D are all initialized to 0; For each matrix element D(i, j), calculate the corresponding distance d(i, j) and fill the cumulative distance matrix. If the cumulative distance is greater than the preset threshold, terminate the calculation; where D(i, j) represents the minimum distance from the starting point of the sequence to x i and y j and d(i, j) is the Euclidean distance between x i and y j Starting from D(N, M), backtrack along the recurrence path to (1, 1) to record the shortest path; according to the obtained shortest path, interpolate the numerical values of the operating parameters in time series Y on the time scale of time series X to align the two, and use the Kriging method for spatial interpolation; traverse all types of operating parameters to form the extended time series of each operating parameter.

3. The method for managing membrane component data based on artificial intelligence according to claim 2, characterized in that: In S2, the following is included: S201: Obtain the extended time series of the transmembrane pressure difference, analyze the coefficient of variation CV of the transmembrane pressure difference, where the coefficient of variation is equal to the standard deviation of the extended time series of the transmembrane pressure difference divided by the average value of the extended time series of the transmembrane pressure difference; analyze the theoretical value of the membrane flux according to the extended time series of the transmembrane pressure difference and Darcy's law, and generate the residual feature ε by analyzing the theoretical value and the actual value of the membrane flux; obtain the extended time series of the membrane flux and use Fourier transform to analyze the spectral energy E and the main frequency component f dominant ; S202: Obtain the extended time series of the produced water quality, flow rate, and inlet water temperature, and analyze the mean and standard deviation of each extended time series; based on the coefficient of variation CV of the transmembrane pressure difference, the mean and standard deviation of each extended time series, the residual characteristic ε, the spectral energy E, and the main frequency component f dominant Construct a multi-dimensional feature vector A, A = [CV, μ T , σ T , μ Q , σ Q , μ C , σ C , ε, E, f dominant ; where μ T represents the mean of the extended time series of the inlet water temperature, and σ T represents the standard deviation of the extended time series of the inlet water temperature; μ Q represents the mean of the extended time series of the flow rate, and σ Q represents the standard deviation of the extended time series of the flow rate; μ C represents the mean of the extended time series of the produced water quality, and σ C represents the standard deviation of the extended time series of the produced water quality; S203: Compress the multi-dimensional feature vector A to form a latent space representation, and then reconstruct the latent space representation into the original data; Optimize the parameters of the encoder and decoder through backpropagation to minimize the reconstruction error; where the loss function is the mean square error; Build a health model based on the reconstruction error.

4. The method for managing membrane component data based on artificial intelligence according to claim 3, characterized in that: In S3, the following is included: S301: Obtain the historical operation data of the membrane component throughout the whole process from the start-of-use state to the failure state, record the remaining life corresponding to the failure state as 0, and construct a life analysis data pair (t, HI t , RUL t ), where t represents time, HI t represents the health index corresponding to time t, and RUL t represents the remaining life corresponding to time t; construct a life prediction model based on the life analysis data pair; S302: Discretize the life prediction model into a difference equation, substitute the life analysis data pair corresponding to the historical operation data into the discretized difference equation, calculate the prediction error corresponding to each time node, where the prediction error is equal to the predicted remaining life minus the true service life, analyze the total error based on the prediction error, and the total error is equal to the sum of the squares of the prediction errors at all time points; Use the gradient descent method to update the degradation rate coefficient and the health index influence coefficient until the total error reaches the minimum value or the maximum number of iterations is reached; Bind the latest degradation rate coefficient and health index influence coefficient to the corresponding membrane component.

5. The method for managing membrane component data based on artificial intelligence according to claim 4, characterized in that: In S4, the following is included: S401: Extract the degradation rate coefficients and health index impact coefficients corresponding to a preset number of historical membrane components, calculate the average values of the degradation rate coefficients and health index impact coefficients of the historical membrane components, and use them as the initial values of the degradation rate coefficients and health index impact coefficients of the membrane component to be predicted; the production coordinates corresponding to the preset number of historical membrane components extracted are the same as those of the membrane component to be predicted; S402: Build a real-time life prediction model based on the initial values of the degradation rate coefficients and health index impact coefficients of the membrane component to be predicted; use the Kalman filter to update the degradation rate coefficients and health index impact coefficients of the membrane component to be predicted in real time; predict the real-time remaining life based on the real-time operation data of the membrane component to be predicted; S403: If the real-time remaining life of the membrane component to be predicted is less than the corresponding threshold and the health index of the membrane component to be predicted is less than the corresponding threshold, send a first-level warning signal; if the real-time remaining life of the membrane component to be predicted is less than the corresponding threshold and the health index of the membrane component to be predicted is greater than or equal to the corresponding threshold, send a second-level warning signal; if the real-time remaining life of the membrane component to be predicted is greater than or equal to the corresponding threshold and the health index of the membrane component to be predicted is less than the corresponding threshold, send a third-level warning signal; if the real-time remaining life of the membrane component to be predicted is greater than or equal to the corresponding threshold and the health index of the membrane component to be predicted is greater than or equal to the corresponding threshold, do not send a warning signal; When the system identifies a first-level warning signal, send an instruction to notify the administrator to replace the membrane component; when the system identifies a second-level warning signal, send an instruction to notify the administrator to perform routine maintenance on the membrane component; when the system identifies a third-level warning signal, send an instruction to notify the administrator to perform routine maintenance or repair on the membrane component.

6. An artificial intelligence-based membrane component data management system, which is implemented by applying the artificial intelligence-based membrane component data management method described in any one of claims 1-5, and is characterized in that, The system includes a data preprocessing module, a health analysis module, a life prediction model construction module, and a real-time warning module; The data preprocessing module is used to obtain the historical operation data of the membrane component, preprocess the historical operation data of the membrane component, achieve spatio-temporal alignment of the operation parameters with different acquisition cycles, and perform spatial interpolation; The health analysis module analyzes the data characteristics of the preprocessed time series of each operation parameter of the membrane component, forms a multi-dimensional feature vector, analyzes the reconstruction error according to the multi-dimensional feature vector, and builds a health model based on the reconstruction error; The life prediction model construction module constructs a life analysis data pair based on the historical operation data of the whole process of the membrane component from the start of use to the failure state, and builds a corresponding life prediction model based on the life analysis data pair; The real-time warning module obtains the life prediction model parameters of the historical membrane components related to the membrane component to be predicted, constructs a real-time life prediction model of the membrane component to be predicted, and performs warning analysis according to the real-time remaining life and health index of the membrane component to be predicted.

7. The membrane component data management system based on artificial intelligence according to claim 6, wherein: The data preprocessing module includes a data acquisition unit and a spatio-temporal alignment and expansion unit; The data acquisition unit is used to obtain the historical operation data of the membrane component, generate the time series of each operation parameter; binarize the membrane surface topography image, fully cover the foreground area based on the preset window set of the system, and count the number of windows of each size that complete the full coverage of the foreground area; form data pairs based on the size of each window and the corresponding number of windows, perform linear fitting on the data pairs, and record the slope of the fitting line as the image feature value; The spatio-temporal alignment and expansion unit is used to extract the time series of any two operation parameters, analyze the shortest path of each operation parameter value in the two time series, perform interpolation based on the shortest path, and perform spatial interpolation using the Kriging method; traverse all types of operation parameters to form the expanded time series of each operation parameter.

8. The membrane component data management system based on artificial intelligence according to claim 6, characterized in that: The health analysis module includes a feature analysis unit and a health model construction unit; The feature analysis unit is used to obtain the expanded time series of the transmembrane pressure difference, and analyze the coefficient of variation of the transmembrane pressure difference; analyze the theoretical value of the membrane flux according to the expanded time series of the transmembrane pressure difference and Darcy's law, and analyze the residual characteristics by comparing the theoretical value and the actual value of the membrane flux; perform Fourier transform on the expanded time series of the membrane flux to analyze the spectral energy and main frequency components; obtain the expanded time series of the produced water quality, flow rate and inlet water temperature, and analyze the mean and standard deviation of each expanded time series; The health model construction unit constructs a multi-dimensional feature vector based on the coefficient of variation of the transmembrane pressure difference, the mean and standard deviation of each expanded time series, the residual characteristics, the spectral energy and the main frequency components; Compress the multi-dimensional feature vector to form a latent space representation, and then reconstruct the latent space representation into the original data; optimize the parameters of the encoder and decoder through backpropagation to minimize the reconstruction error; Construct a health model based on the reconstruction error.

9. The membrane component data management system based on artificial intelligence according to claim 6, characterized in that: The life prediction model construction module includes a life prediction model construction unit and a parameter optimization unit; The life prediction model construction unit obtains the historical operation data of the whole process of the membrane component from the start of use to the failure state, constructs life analysis data pairs based on the historical operation data, and constructs a life prediction model according to the life analysis data pairs; The parameter optimization unit is used to discretize the life prediction model into a difference equation; Use the gradient descent method to update the degradation rate coefficient and the health index influence coefficient until the total error reaches the minimum value or the maximum number of iterations is reached; Bind the latest degradation rate coefficient and health index influence coefficient to the corresponding membrane component.

10. A membrane component data management system based on artificial intelligence according to claim 6, characterized in that: The real-time warning module includes a real-time model construction unit and a warning unit; The real-time model construction unit is used to extract the degradation rate coefficient and the health index influence coefficient corresponding to a preset number of historical membrane components, calculate the average value of the degradation rate coefficient and the average value of the health index influence coefficient of the historical membrane components, and use them as the initial values of the degradation rate coefficient and the health index influence coefficient of the membrane component to be predicted; the production coordinates of the preset number of historical membrane components extracted are the same as those of the membrane component to be predicted; construct a real-time life prediction model based on the initial values of the degradation rate coefficient and the health index influence coefficient of the membrane component to be predicted; and perform real-time remaining life prediction; The warning unit classifies warning signals according to the real-time remaining life and health indicators, and gives system prompts according to the warning signal levels.

Images