Water supply network data quality control method coupling multiscale decomposition and independent noise reduction

Abstract

The application discloses a water supply pipe network data quality control method coupling multi-scale decomposition and independent noise reduction, comprising the following steps: (1) decomposing a monitoring data sequence and determining a noisy IMF component; (2) performing independent noise reduction on the noisy IMF component based on Gaussian filtering; and (3) reconstructing the monitoring data sequence. The method can improve the noise reduction effect of the monitoring data sequence of the water supply pipe network, has robustness in reducing monitoring data under different leakage intensity leakage scenes, is an important way for real-time monitoring data quality control of the water supply pipe network, provides technical support for realizing real-time hydraulic model construction of the pipe network and data analysis driven intelligent management and control of the pipe network, and has important significance for improving the comprehensive performance of the intelligent system of the water supply pipe network.

Description

Technical Field

[0001] This invention belongs to the field of urban water supply network technology, specifically relating to a data quality control method for water supply networks that couples multi-scale decomposition and independent noise reduction. Background Technology

[0002] In recent years, with the continuous development of IoT and sensor network technologies, real-time monitoring systems for water supply networks, constructed from sensors distributed throughout the network, can record monitoring data online, such as node pressure and pipeline flow. However, monitoring data inevitably contains some noise, which gradually increases as equipment ages. Due to advancements in data transmission and storage technologies, the volume of data has exploded, highlighting the growing importance of data quality. Controlling data quality is fundamental to achieving intelligent management and control of water supply networks through data analysis and model-driven approaches.

[0003] As a whole system, the actual monitoring data of a water supply network, especially pressure monitoring data, often changes synchronously with fluctuations in water consumption. Therefore, actual pressure data is often linearly correlated, while noise is usually considered an independent and uncorrelated stationary time series. This significant difference between noise and actual monitoring data can provide new ideas for noise reduction in water supply network monitoring data.

[0004] Independent denoising increases the complexity of the data structure and alters the relative positions of the original data's sample points. Subsequent multi-scale decomposition techniques can further impact the decomposition results, leading to instability. A combined approach, integrating independent denoising into the multi-scale decomposition process, effectively reduces data distortion and is more suitable for pipeline network data quality management. Therefore, before massive monitoring data sequences are applied to intelligent management of water supply networks, real-time monitoring data quality control is urgently needed at both theoretical and technical levels.

[0005] In view of this, this invention proposes a multi-scale decomposition technology based on fusion independent denoising to meet the quality control requirements of water supply network monitoring data. It combines ICEEMDAN with Gaussian filtering, determines the boundary point through a modified CMSE criterion, and uses correlation coefficient, signal-to-noise ratio, root mean square error, and permutation entropy as evaluation indicators to characterize the denoising performance for data reconstruction. Finally, the denoised monitoring data sequence is obtained, which effectively improves the denoising effect of water supply network monitoring data sequence. Summary of the Invention

[0006] This invention aims to provide a data quality control method for water supply networks that couples multi-scale decomposition with independent noise reduction. To achieve the above objective, this invention adopts the following steps: (1) Decompose the monitoring data sequence and determine the noisy IMF component; (2) Independent noise reduction of noisy IMF components based on Gaussian filtering; (3) Reconstruct the monitoring data sequence.

[0007] Furthermore, step 1 specifically includes: 1.1 Obtaining different IMF components based on ICEEMDAN time series decomposition: To monitor time series Add to i Gaussian white noise with a mean of 0 At the same time, the amplitude of noise is added. ,form i A sequence to be decomposed ,in i =1,2,3…,n: ; Regarding the above i Perform EMD decomposition on the sequence to obtain i Local means, the original signal Subtract this i The first IMF component is obtained by averaging the local means: ; In the formula: It is a local mean function; The second IMF component is: ; Repeat the above steps until the final remainder is a monotonic function or a constant. 1.2 Determining whether the IMF component contains noise using the modified CMSE criterion: After the data sequence is decomposed by ICEEMDAN, the following results are obtained: m One IMF component sorted by frequency from highest to lowest and one remainder. res The boundary between noisy IMF components and noisy IMF components is determined based on the CMSE criterion, and the calculation method is as follows:

[0008] ; In the formula x The original signal, N For signal length, k The number of decomposition layers; Since the noise is high-frequency, its energy gradually decreases as ICEEMDAN is decomposed. The global minimum point of the IMF component energy is used as the dividing point. j See formula: ; This method modifies the above formula. When a local minimum exists before the global minimum is calculated using the CMSE criterion, the boundary point is... j The formula should be the position of the number corresponding to the first local minimum plus 1. ; The IMF components before the dividing point are noisy components and require further independent noise reduction; the IMF components after the dividing point are noise-free components, that is, data sequences composed of effective information.

[0009] Furthermore, step 2 specifically includes: Each noisy IMF component obtained in step 1.2 above is subjected to Gaussian filtering for independent noise reduction. The one-dimensional expression for Gaussian filtering is as follows: ; Where σ is the standard deviation of the function, with a range of 0.2-0.65 and values ​​taken at intervals of 0.05. Each component will produce 10 sets of filtering results under different parameters.

[0010] Furthermore, step 3 specifically includes: 3.1 Assumptions for k Reconstruct each IMF component, and select a set of data from the independent noise reduction results of each component, totaling... Group data, Data and noise-free IMF components are reconstructed to obtain Group noise reduction monitoring data; 3.2 The correlation coefficient CORR, signal-to-noise ratio SNR, root mean square error RMSE, and permutation entropy were selected as evaluation indicators for noise reduction performance. 3.3 Calculate the correlation coefficient between the denoised data and the original data. The closer the value is to 1, the more similar the denoised data is to the original data, that is, the better the denoising effect. The formula is as follows: ; in, The original signal, The signal after noise reduction; 3.4 Calculate the signal-to-noise ratio (SNR) of the denoised data. A higher SNR indicates better denoising performance of the algorithm. The calculation formula is as follows: ; In the formula, N For data length, The original signal, The signal after noise reduction; 3.5 Calculate the root mean square error between the denoised data and the original data. The smaller the value, the better the denoising effect. The calculation formula is as follows: ; 3.6 Sorting Entropy Calculate the permutation entropy of the denoised data. The larger the value, the more noise is removed and the better the denoising effect. The specific steps for calculating the permutation entropy are as follows: 3.6.1 Given a length of N Time series data By arranging these data, different sequences are obtained, for a length of... N Time series can generate N There are 2 different arrangements, and each arrangement corresponds to a new sequence.

[0011] 3.6.2 For each permutation, the entropy value is calculated using the entropy formula in information theory. The Shannon entropy is calculated by calculating the frequency distribution of each permutation and its entropy value, and then averaging the entropy values ​​of all permutations to obtain the permutation entropy as the entropy value of the entire time series data. The calculation formula is as follows: ; in, This represents the total number of permutations. Indicates the first i The probability of a permutation; 3.7 Characterization based on the above evaluation indicators The noise reduction performance of the reconstructed data sequences under different parameters was evaluated, and the best-performing set of parameters and its corresponding reconstructed data were determined, ultimately yielding the noise-reduced monitoring data sequence.

[0012] The present invention achieves the above technical solution and has the following technical effects or advantages: Due to advancements in data transmission and storage technologies, the volume of data in water supply networks has exploded. Data noise has led to a decline in data quality, significantly hindering the implementation of data-driven intelligent management and control methods for water supply networks. Multi-scale decomposition denoising can reduce data complexity, but it also results in the removal of some true data. Independent denoising, on the other hand, increases the complexity of the data structure and reduces data stability. The water supply network data quality control method based on multi-scale decomposition incorporating independent denoising integrates independent denoising into the multi-scale decomposition process for data processing. This effectively reduces data distortion, improves the denoising effect of water supply network monitoring data sequences, and demonstrates robustness in monitoring data quality control under leakage scenarios with varying leakage intensities. This method is of great significance for improving the overall performance of intelligent water supply network systems. Attached Figure Description

[0013] Figure 1 A flowchart illustrating a data quality control method for water supply networks that couples multi-scale decomposition with independent noise reduction; Figure 2 A schematic diagram of the water supply network in City J; Figure 3 Waveform diagram of IMF components; Figure 4 This is a continuous mean square error curve for the IMF. Figure 5 The root mean square error curve of the IMF component noise reduction result; Figure 6(a) shows the correlation coefficients of the denoising results of Gaussian filtering, coupled multi-scale decomposition, and independent denoising methods. ); Figure 6(b) shows the signal-to-noise ratio of the denoising results obtained by Gaussian filtering, coupled multi-scale decomposition, and independent denoising methods. ); Figure 6(c) shows the root mean square error of the denoising results of Gaussian filtering, coupled multi-scale decomposition, and independent denoising methods. ); Figure 7(a) shows the correlation coefficient of the denoising results of the coupled multi-scale decomposition and independent denoising methods. ); Figure 7(b) shows the signal-to-noise ratio of the denoising results of the coupled multi-scale decomposition and independent denoising methods. ); Figure 7(c) shows the root mean square error of the denoising results of the coupled multi-scale decomposition and independent denoising methods. ); Figure 8(a) shows the correlation coefficient of the denoising results of the coupled multi-scale decomposition and independent denoising methods (leakage condition and normal condition). Figure 8(b) shows the signal-to-noise ratio of the denoising results of the coupled multi-scale decomposition and independent denoising methods (leakage condition and normal condition). Figure 8(c) shows the root mean square error of the denoising results of the coupled multi-scale decomposition and independent denoising methods (leakage condition and normal condition). Detailed Implementation

[0014] This invention aims to propose a data quality control method for water supply networks that couples multi-scale decomposition and independent noise reduction (hereinafter referred to as "this invention"). The implementation method of this invention will be further described in detail below with reference to the accompanying drawings and examples. Figure 1 The flowchart of a data quality control method for water supply networks that couples multi-scale decomposition and independent noise reduction is presented. The specific steps are as follows: Construct a hydraulic model of the water supply network and obtain the data sequence monitored by the water supply network monitoring system; like Figure 2The image shows a real water supply network for a medium-sized city, J. J has three water plants supplying water, with 491 water demand points and 640 pipe sections. It has 20 pressure sensors installed. n s =20).

[0015] Establish a node water demand prediction model to predict future water demand in real time. b Water demand at each time period; t The predicted water demand at each water demand node is used as input to the hydraulic model of the water supply network. The pressure values ​​at the monitoring points (pressure sensors) are simulated using EPANET software (DDA) to obtain... t Time-based pressure prediction vector The unit is meters (m), and the formula is as follows: ; in, for t Time of the first Pressure prediction data from a pressure sensor, This refers to the number of pressure sensors. for The column vector of a row; Table 1 t Predicted water demand at time =1 Node number 1 2 … 240 … 490 491 Water demand (CMH) 6.83 7.31 … 42.52 … 23.07 36.71 ; Will As model input, simulate Real-time pressure monitoring data The formula is as follows: ; in, This is the pressure prediction data for the first time step, presented as a 20-row column vector. In this invention, data from 240 time steps are simulated to ultimately obtain... This data set comprises data from 20 monitoring points at 240 time points. The simulated pressure value matrix for these 20 monitoring points at 240 time points is presented here:

[0016] ; The simulated pressure value is affected by pressure measurement disturbances and water demand disturbances. The following is an example of adding disturbance values: Pressure measurement perturbations are added to simulated pressure values ​​to mimic real data. The error of the pressure measurement perturbation follows a standard normal distribution. The noise intensity is given below. Example of pressure vector at time: ; Add a water demand perturbation to the simulated pressure value. The error of the water demand perturbation follows a standard normal distribution N(0, 1). σ ), ; In the formula, σ To add the standard deviation of the perturbation, p For the perturbation ratio, μ This represents the predicted water demand at the water demand node. The water demand disturbance is given here. Noise intensity Example of adding a pressure vector to account for water demand disturbance: .

[0017] This invention provides a data quality control method for water supply networks that couples multi-scale decomposition and independent noise reduction, comprising the following steps: (1) Decompose the monitoring data sequence and determine the noisy IMF component.

[0018] Step (1) specifically includes: (1.1) Obtaining different IMF components based on ICEEMDAN time series decomposition: To monitor time series Add to i Gaussian white noise with a mean of 0 At the same time, the amplitude of noise is added. ,form i A sequence to be decomposed ,in i =1,2,3…,n: ; Regarding the above i Perform EMD decomposition on the sequence to obtain i Local means, the original signal Subtract this i The first IMF component is obtained by averaging the local means: ; In the formula: It is a local mean function; The second IMF component is: ; Repeat the above steps until the final remainder is a monotonic function or a constant. Figure 3 To determine the decomposition results when adding pressure perturbations to the simulated pressure values, the original data was decomposed into five IMF components (IMF1 to IMF5) with frequencies ranging from high to low, and a remainder term. res .

[0019] (1.2) Determine whether the IMF component contains noise using the modified CMSE criterion: After the data sequence is decomposed by ICEEMDAN, the following results are obtained: m One IMF component sorted by frequency from highest to lowest and one remainder. res The boundary between noisy IMF components and noisy IMF components is determined based on the CMSE criterion, and the calculation method is as follows:

[0020] ; In the formula, x The original signal, N For signal length, k The number of decomposition layers; Since the noise is high-frequency, its energy gradually decreases as ICEEMDAN is decomposed. The global minimum point of the IMF component energy is used as the dividing point. j See formula: ; This method modifies the above formula. When a local minimum exists before the global minimum is calculated using the CMSE criterion, the boundary point is... j The formula should be the position of the number corresponding to the first local minimum plus 1. ; The IMF components before the dividing point are noisy components and require further independent noise reduction; the IMF components after the dividing point are noise-free components, that is, data sequences composed of effective information.

[0021] First, calculate the correlation coefficient between each IMF component and the original data, using the following formula: ; in, Indicates the first K Each IMF component at time... t The value, h(t) Indicates the original data at time [time]. t The value, N This indicates the length of the data sequence. The correlation coefficient between the IMF components and the original data is as follows: Figure 4 As shown, when the IMF serial number is 3, the IMF continuous mean square error curve shows its first extreme point. According to the correlation coefficient criterion, the IMF at which the first correlation coefficient reaches its minimum value is taken as the dividing point. Therefore, IMF1, IMF2, and IMF3 are noisy data, while IMF4 and IMF5 are noiseless data. (2) Independent noise reduction of noisy IMF components based on Gaussian filtering.

[0022] Step (2) specifically includes: The one-dimensional expression for Gaussian filtering is: ; Where σ is the standard deviation of the function, the standard deviation ranges from 0.2 to 0.65, and the values ​​are taken at intervals of 0.05. Gaussian filtering is performed on the noisy IMF components, and 10 independent noise reduction results are generated for each component under different parameters.

[0023] Denoising is performed on each noisy IMF component using Gaussian filtering, with the parameters of the Gaussian filter... σ Taking a value in the range of 0.2-0.65 (interval 0.05) as an example, each noisy IMF component (IMF1, IMF2 and IMF3) generates 10 corresponding denoised components, resulting in a total of 1000 (10^3 = 1000) sets of denoised data sequences.

[0024] (3) Reconstruct the monitoring data sequence.

[0025] Step (3) specifically includes: (3.1) Assume that for k Reconstruct each IMF component, and select a set of data from the independent noise reduction results of each component, totaling... Group data, Data and noise-free IMF components are reconstructed to obtain Group noise reduction monitoring data; (3.2) The correlation coefficient CORR, signal-to-noise ratio SNR, root mean square error RMSE and permutation entropy are selected as the evaluation indicators for noise reduction performance. (3.3) Calculate the correlation coefficient between the denoised data and the original data. The closer the value is to 1, the more similar the denoised data is to the original data, that is, the better the denoising effect. The formula is as follows: ; in, The original signal, The signal after noise reduction; (3.4) Calculate the signal-to-noise ratio (SNR) of the denoised data. The higher the SNR, the better the denoising performance of the algorithm. The calculation formula is as follows: ; In the formula, N For data length, The original signal, The signal after noise reduction; (3.5) Calculate the root mean square error between the denoised data and the original data. The smaller the value, the better the denoising effect. The calculation formula is as follows: .

[0026] The 1000 denoised data sequences are reconstructed with the remaining IMF4 and IMF5 data sets to obtain a 1000-set denoised dataset. The CORR, SNR, and RMSE of the denoised data and the original data are then calculated, such as... Figure 5 As shown, taking the calculation of RMSE as an example, the data with 233 groups has the smallest RMSE value, and the corresponding standard deviations are 0.4, 0.45 and 0.25, respectively. That is, the data with 233 groups has the best noise reduction effect.

[0027] Based on the above method, the data sequences of 20 monitoring points were traversed, and their noise reduction performance evaluation indicators were calculated to obtain the set of parameters with the best noise reduction performance and their corresponding reconstructed data. The larger the CORR, the more similar the noise-reduced data is to the data without added noise, that is, more noise is removed. SNR represents the ratio of real data to noise in the data. The larger the SNR, the less noise in the data. RMSE represents the square root of the average degree of difference between the noise-reduced data and the original data without added noise. The smaller the RMSE, the better the noise reduction effect. Figures 6(a) to 6(c) show the noise intensity when... The noise reduction results were compared between using Gaussian filtering alone and using the technique of this invention. The average CORR of the Gaussian filtering method was 0.995475, the average SNR was 43.5782, and the average RMSE was 0.197428. The average CORR of the ICEEDMAN method combined with Gaussian filtering was 0.996323, the average SNR was 44.9861, and the average RMSE was 0.187623. All three evaluation metrics indicate that the noise reduction effect of the technique of this invention is better.

[0028] (3.6) Sorting entropy Calculate the permutation entropy of the denoised data. The larger the value, the more noise is removed and the better the denoising effect. The steps to calculate the permutation entropy are as follows: (3.6.1) Given a length of N Time series data By arranging these data, different sequences are obtained, for a length of... N Time series can generate N There are several different arrangements, each of which corresponds to a new sequence; (3.6.2) For each permutation, the entropy value is calculated using the entropy formula in information theory. The Shannon entropy is calculated by calculating the frequency distribution of each permutation and then calculating its entropy value; then the entropy values ​​of all permutations are averaged to obtain the permutation entropy as the entropy value of the entire time series data. The calculation formula is as follows: ; in, This represents the total number of permutations. Indicates the first i The probability of a permutation; Calculate the permutation entropy of the two sets of filtered residuals. Permutation entropy reflects the degree of disorder and randomness of the data. The larger the permutation entropy, the more disordered the data, indicating that more noise has been removed. The noise intensity is given here. The sorting entropy of the two sets of filter residuals under the condition of...

[0029] As shown in Table 2, the average permutation entropy of the Gaussian filter residuals of the 20 monitoring points is 0.9491, and the average permutation entropy of the Gaussian filter residuals combined with ICEEMDAN is 0.9577. That is, the present invention achieves more noise removal and has better noise reduction performance.

[0030] Table 2. Noise Residual Arrangement Entropy Based on Two Algorithms Monitoring point number Gaussian filter permutation entropy ICEEMDAN combined with Gaussian filter to arrange entropy 1 0.9479 0.9543 2 0.9518 0.9527 3 0.9523 0.9649 4 0.9469 0.9531 5 0.9507 0.9587 6 0.9454 0.9642 7 0.9520 0.9629 8 0.9473 0.9538 9 0.9455 0.9496 10 0.9522 0.9586 11 0.9490 0.9529 12 0.9488 0.9642 13 0.9481 0.9491 14 0.9530 0.9638 15 0.9491 0.9545 16 0.9507 0.9634 17 0.9467 0.9537 18 0.9497 0.9621 19 0.9481 0.9639 20 0.9485 0.9538 (3.7) Based on the above evaluation indicators, the characteristics are... The noise reduction performance of the reconstructed data sequences under different parameters was evaluated, and the best-performing set of parameters and its corresponding reconstructed data were determined, ultimately yielding the noise-reduced monitoring data sequence.

[0031] Experimental results: Figures 7(a) to 7(c) show the water demand disturbance as noise intensity The noise reduction results of this invention are as follows: The original data showed an average CORR of 0.9948, an average SNR of 42.8184, and an average RMSE of 0.1987. After noise reduction, the average CORR increased by 0.0006, the average SNR increased by 0.4013, and the average RMSE decreased by 0.0089. The data for all three evaluation indicators demonstrate that this invention maintains good noise reduction performance under water demand disturbance scenarios.

[0032] Table 3 shows the noise intensity The water demand disturbances are respectively At that time, the residual permutation entropy of the present invention. As the water demand perturbation increases, the average permutation entropy continuously decreases, meaning that the filtered-out true components continuously increase. This indicates that the water demand perturbation ratio affects the noise reduction performance of the present invention.

[0033] Table 3. Noise Residual Arrangement Entropy Based on Multi-Scale Decomposition Techniques for Integrated Independent Noise Reduction ( ) Monitoring point number =5% =10% 1 0.9368 0.9304 2 0.9451 0.9332 3 0.9443 0.9299 4 0.9389 0.9317 5 0.9467 0.9306 6 0.9461 0.9338 7 0.9354 0.9308 8 0.9361 0.9291 9 0.9419 0.9346 10 0.9396 0.9321 11 0.9515 0.9322 12 0.9468 0.9307 13 0.9426 0.9323 14 0.9472 0.9328 15 0.9489 0.9298 16 0.9518 0.9336 17 0.9342 0.9325 18 0.9347 0.9313 19 0.9475 0.9302 20 0.9363 0.9296 Figures 8(a) to 8(c) show the CORR, SNR, and RMSE indices of the present invention after noise reduction under leakage and normal operating conditions. The leakage condition (water demand node 350 has an intensity of...) is used as an example. (Leakage) accompanied by noise intensity ( Taking [example] as an example, after denoising, the CORR, SNR, and RMSE of the data are generally close under normal and leakage conditions. For some nodes, the three indicators under normal conditions are slightly higher than those under leakage conditions, while for other nodes, the three indicators under leakage conditions are slightly higher than those under normal conditions. This indicates that the present invention is suitable for quality control of monitoring data in leakage scenarios, and that the present invention has a certain robustness in denoising data related to leakage control in water supply networks.

Claims

1. A data quality control method for water supply networks that couples multi-scale decomposition and independent noise reduction, characterized in that, Includes the following steps: (1) Decompose the monitoring data sequence and determine the noisy IMF component; (2) Independent noise reduction of noisy IMF components based on Gaussian filtering; (3) Reconstruct the monitoring data sequence; Specifically, step (1) is as follows: (1.1) Obtaining different IMF components based on ICEEMDAN time series decomposition: To monitor time series Add to i Gaussian white noise with a mean of 0 At the same time, the amplitude of noise is added. ,form i A sequence to be decomposed ,in i =1,2,3…,n: ； Regarding the above i Perform EMD decomposition on the sequence to obtain i Local means, the original signal Subtract this i The first IMF component is obtained by taking the first-order remainder of the average of the local means: ； In the formula: It is a local mean function; The second IMF component is: ； Repeat the above steps until the final remainder is a monotonic function or a constant. (1.2) Determine whether the IMF component contains noise using the modified CMSE criterion: After the data sequence is decomposed by ICEEMDAN, the following results are obtained: m One IMF component sorted by frequency from highest to lowest and one remainder. res The boundary between noisy IMF components and noisy IMF components is determined based on the CMSE criterion, and the calculation method is as follows: ； In the formula, x The original signal, N For signal length, k The number of decomposition layers; Since the noise is high-frequency, its energy gradually decreases as ICEEMDAN is decomposed. The global minimum point of the IMF component energy is used as the dividing point. j See formula: ； The above formula has been modified so that when a local minimum exists before the global minimum is calculated using the CMSE criterion, the boundary point is... j The formula should be the position of the number corresponding to the first local minimum plus 1. ； The IMF components before the dividing point are noisy components and require further independent noise reduction; the IMF components after the dividing point are noise-free components, that is, data sequences composed of effective information. Specifically, step (3) is as follows: (3.1) Assume that for k Reconstruct each IMF component, and select a set of data from the independent noise reduction results of each component, totaling... Group data, Data and noise-free IMF components are reconstructed to obtain Group noise reduction monitoring data; (3.2) The correlation coefficient CORR, signal-to-noise ratio SNR, root mean square error RMSE and permutation entropy are selected as the evaluation indicators for noise reduction performance. (3.3) Calculate the correlation coefficient between the denoised data and the original data. The closer the value is to 1, the more similar the denoised data is to the original data, that is, the better the denoising effect. The formula is as follows: ； in, The original signal, The signal after noise reduction; (3.4) Calculate the signal-to-noise ratio (SNR) of the denoised data. The higher the SNR, the better the denoising performance of the algorithm. The calculation formula is as follows: ； In the formula, N For data length, The original signal, The signal after noise reduction; (3.5) Calculate the root mean square error between the denoised data and the original data. The smaller the value, the better the denoising effect. The calculation formula is as follows: ； (3.6) Sorting entropy Calculate the permutation entropy of the denoised data. The larger the value, the more noise is removed and the better the denoising effect. The steps to calculate the permutation entropy are as follows: (3.6.1) Given a length of N Time series data By arranging these data, different sequences are obtained, for a length of... N Time series can generate N There are several different arrangements, each of which corresponds to a new sequence; (3.6.2) For each permutation, the entropy value is calculated using the entropy formula in information theory. The Shannon entropy is calculated by calculating the frequency distribution of each permutation and then calculating its entropy value; then the entropy values ​​of all permutations are averaged to obtain the permutation entropy as the entropy value of the entire time series data. The calculation formula is as follows: ； in, This represents the total number of permutations. Indicates the first i The probability of a permutation; (3.7) Based on the above evaluation indicators, the characteristics are... The noise reduction performance of the reconstructed data sequences under different parameters was evaluated, and the best-performing set of parameters and its corresponding reconstructed data were determined, ultimately yielding the noise-reduced monitoring data sequence.

2. The water supply network data quality control method with coupled multi-scale decomposition and independent noise reduction as described in claim 1, characterized in that, Step (2) is as follows: The one-dimensional expression for Gaussian filtering is: ； Where σ is the standard deviation of the function, the standard deviation ranges from 0.2 to 0.65, and the values ​​are taken at intervals of 0.

05. Gaussian filtering is performed on the noisy IMF components, and 10 independent noise reduction results are generated for each component under different parameters.

Images