Water pump device characteristic curve correction method based on clustering analysis and trust region algorithm

Abstract

The invention discloses a water pump device characteristic curve correction method based on clustering analysis and a trust region algorithm, and the method comprises the steps: carrying out the clustering processing of the steady state working condition data of a water pump device through a K-means clustering method, combining a large number of similar working conditions through iteration, and screening out the typical working condition points of the water pump device; on the basis of the typical working condition points of the water pump device, an original characteristic curve of the water pump device is obtained through a moving least square method or cubic spline or polynomial fitting; and parameters are introduced to transform an original characteristic curve of the water pump device to realize three-dimensional affine transformation, and characteristic curve correction of the water pump device is realized based on a Dogbox algorithm in a trust region algorithm. The method has the advantages that a model test or a field test is not needed, correction is carried out according to basic operation monitoring data of a common large axial flow / mixed flow pump station, universality is high, certain precision can still be kept in a working condition interval lacking data, and scientific basis and technical support can be provided for actual dispatching of the pump station.

Description

Technical Field

[0001] The invention relates to the technical field of optimization scheduling of large water pumping stations, and in particular to a method for correcting a characteristic curve of a water pump device based on cluster analysis and a trust region algorithm. Background Art

[0002] As an important control structure in the inter-basin water diversion project, the pump station will transform the traditional dispatching method dominated by manual experience into scientific, efficient and intelligent decision-making in the process of promoting the construction of digital twin intelligence, so as to achieve the goal of reducing costs and increasing efficiency. The flow-head-blade angle characteristic curve of the pump device is a curve that describes the relationship and change law between the three performance parameters of the pump flow, head and blade placement angle. It reflects the operating status of the pump, and its accuracy directly affects the dispatching decision-making judgment.

[0003] At present, the flow-head-blade angle characteristic curves of the water pump device used in the dispatching work are mostly obtained through model tests at the factory. During the operation process, affected by factors such as installation conditions, inflow conditions, equipment degradation, and scale effect, there is a difference between the actual operating state and the ideal operating condition, and the characteristic curve is offset, resulting in the pump station optimization dispatching model based on this being unable to calculate an operating plan that conforms to the actual situation. In order to ensure the accuracy of the pump station dispatching decision, the flow-head-blade angle characteristic curve of the water pump device needs to be corrected. However, considering the safety of project operation, the actual operating conditions of the pump station are generally concentrated in a small design operating range, and it is difficult to obtain the measured data under the remaining stable conditions through on-site unit tests.

[0004] After reviewing relevant literature, it is found that there are many studies on the prediction and fitting of the characteristic curve of water pump devices, but there is a lack of research on the offset and correction of the characteristic curve in engineering practice. The existing methods and technologies cannot provide technical support for correcting the flow-head-blade angle characteristic curve of water pump devices. Summary of the invention

[0005] The object of the present invention is to provide a method for correcting a characteristic curve of a water pump device based on cluster analysis and a trust region algorithm, thereby solving the above-mentioned problems existing in the prior art.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] A method for correcting characteristic curve of water pump device based on cluster analysis and trust region algorithm.

[0008] S1. Cluster analysis to select typical operating points:

[0009] The K-means clustering method is used to cluster the steady-state operating data of the water pump device, and the typical operating points of the water pump device are screened out by iteratively merging a large number of similar operating conditions.

[0010] S2. Extraction of original characteristic curve:

[0011] Based on the typical operating points of the water pump device, the original characteristic curve of the water pump device is obtained by using the moving least square method or cubic spline or polynomial fitting;

[0012] S3, Characteristic curve correction based on typical operating points:

[0013] Parameters are introduced to transform the original characteristic curve of the water pump device to realize three-dimensional affine transformation, and the characteristic curve of the water pump device is corrected based on the Dogbox algorithm in the trust region algorithm.

[0014] Preferably, the K-means clustering method in step S1 specifically includes the following contents:

[0015] S11. Select K discrete initial cluster centers through K-means++ algorithm;

[0016] S12, assign the data set to the cluster corresponding to the cluster center through the classic K-means++ algorithm;

[0017] S13, calculating the mean of each cluster and using the mean as the new cluster center;

[0018] S14, repeating steps S12 to S13 until the cluster center remains unchanged or changes less than the change threshold;

[0019] S15. Use the elbow method to determine whether it is the optimal K value. If not, go to step S11. If so, divide the steady-state operating data of the water pump device into K cluster centers to screen out typical operating points of the water pump device.

[0020] Preferably, step S2 specifically comprises:

[0021] The flow-head-blade angle characteristic curve of the water pump device is fitted by the method of binary quadratic polynomial, and the original characteristic curve equation of the water pump device is obtained, which is in the following form:

[0022] θ=p00+p10*Q+p01*H+p20*Q 2 +p11*Q*H+p02*H 2

[0023] Among them, θ is the blade installation angle of the water pump device; Q is the flow rate of the water pump device; H is the head of the water pump device; p00, p10, p01, p20, p11, and p02 are the coefficients of each term in the fitted polynomial equation.

[0024] Preferably, step S3 specifically includes the following contents:

[0025] S31, introduce parameters a, b, c, r to transform the original characteristic curve equation obtained in step S2 to achieve three-dimensional affine transformation. The equation after linear transformation is as follows:

[0026]

[0027] Among them, a, b, c, r are new parameters introduced by the modified equation; Q' is the flow value of the water pump device after correction; H' is the head value of the water pump device after correction; θ' is the blade placement angle value of the water pump device after correction;

[0028] Assume that the normal operating range of the water pump device is Q∈[Q min ,Q max ],H∈[H min ,H max ],θ∈[θ min ,θ max ]; the search domains are specified as a∈[-0.3*|Q min -Q max |,0.3*|Q min -Q max |],b∈[-0.3*|H min -H max |,0.3*|H min -H max |],c∈[-0.3*|θ min -θ max |,0.3*|θ min -θ max |]; Q min and Q max are the minimum flow rate and maximum flow rate of the water pump device respectively; H min and H max are the minimum head and maximum head of the water pump device respectively; θ min and θ max They are the minimum and maximum angles for placing the blades of the water pump device respectively.

[0029] S32. Based on the Dogbox algorithm in the trust region algorithm, the fitting parameters are determined by minimizing the residual sum of squares, that is, the characteristic curve is translated and rotated in the three-dimensional space concept, so that the characteristic curve approaches the selected typical operating points, and the characteristic curve correction of the water pump device is completed.

[0030] Preferably, before step S1, the step further includes:

[0031] S0. Data collection and preprocessing: According to the actual engineering scheduling and operation conditions at different time scales, the water conditions and working conditions of the water pump device are statistically summarized, the data is processed using data cleaning methods, and the steady-state operating data of the water pump device is screened out using the adaptive Gaussian filtering method and the confidence method.

[0032] Preferably, the water condition and working condition related data include the water level below the station, the water level above the station, the flow monitoring value of each unit, and the blade angle monitoring value of each unit of the water pump device;

[0033] Data processing includes removing null values, outliers and abnormal values.

[0034] Preferably, in step S0, the noise is filtered using an adaptive Gaussian filter method, and the calculation formula is:

[0035] S(t,σ)=f(t)*g(t,σ)

[0036] Among them, S(t,σ) is the smoothing result of the signal or data after Gaussian filtering; f(t) is the original signal or data sequence, which is a function that changes with time t; t is the time variable; g(t,σ) is the Gaussian kernel function; * is the convolution operator; σ is the standard deviation of the Gaussian function;

[0037]

[0038] Where m is the sampling time point; f(m) is the historical data signal; f k+1 (m) is the output value after smoothing at the k+1th iteration; f k (m+j) is the value of the signal or data at position m+j in the kth iteration; w k (m+j,σ) is the weight of each point in the window; j is the offset in the convolution window, which represents the time offset relative to m; M is the width parameter of the convolution window.

[0039] Preferably, the confidence method is specifically:

[0040] After the data is denoised by the adaptive Gaussian filtering method, the error probability distribution function of the steady-state data of the direct monitoring indicators is obtained, so that the change threshold of the indicators can be determined under the premise of a certain confidence interval as the basis for judging the steady state; when all indicators are in a steady state, the operating point is considered to be a steady-state operating point.

[0041] Preferably, after step S3, the method further includes:

[0042] S4. Extraction of corrected characteristic curve: Use origin software to draw the corrected flow-head-blade angle characteristic curve.

[0043] Preferably, step S4 specifically comprises:

[0044] The corrected characteristic curve equation obtained in step S3 is discretized by using origin software, and a corrected integer angle contour map is drawn to obtain a corrected flow-head-blade angle characteristic curve map.

[0045] The beneficial effects of the present invention are as follows: 1. The present invention selects steady-state operating data from long-sequence monitoring data by combining an adaptive Gaussian filtering method with a confidence method. The adaptive Gaussian filtering method can retain the original signal change information to the greatest extent while filtering out noise. 2. The present invention performs clustering processing on long-sequence historical data based on the K-means clustering method, greatly reducing the amount of data, improving the calculation speed, and making the data points cover as many working conditions as possible, avoiding the curve overfitting phenomenon that may occur in the linear regression process due to the excessive density of point sets in some common working condition intervals during the operation of the same pump station. 3. The present invention introduces new parameters to transform the original characteristic curve equation for affine transformation, so that the curve fits the data points as much as possible without destroying the original curve space structure, avoiding the abnormal deformation of the characteristic curve that does not conform to the hydraulic characteristics of the water pump caused by direct fitting, thereby improving the curve correction quality. 4. The present invention performs parameter optimization based on the Dogbox algorithm in the trust region algorithm, and has good robustness, stability and convergence characteristics when solving high-dimensional nonlinear parameter optimization problems that may occur due to the introduction of new parameters. 5. The present invention has strong versatility. For axial flow / mixed flow pump stations with long-sequence historical monitoring data, this method can be used to generate corrected water pump device characteristic curves. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a flow chart of a method in an embodiment of the present invention;

[0047] Figure 2 3 is a flow-lift-blade angle characteristic curve diagram of the water pump after correction in the embodiment of the present invention (vertical line). DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation methods described herein are only used to explain the present invention and are not used to limit the present invention.

[0049] like Figure 1 As shown, in this embodiment, a method for correcting the characteristic curve of a water pump device based on cluster analysis and trust region algorithm is provided, which includes the following five parts:

[0050] 1. Data Collection and Preprocessing

[0051] According to the actual engineering scheduling and operation conditions at different time scales, the water conditions, working conditions and other related data of the water pump device are statistically summarized, the data are processed using data cleaning methods, and the steady-state operating data of the water pump device are screened out using adaptive Gaussian filtering and confidence methods.

[0052] Specifically: According to the actual project scheduling and operation conditions at different time scales, the water conditions, work conditions and other related data are statistically summarized. The data mainly includes the water level below the pump station, the water level above the pump station, the flow monitoring values ​​of each unit of the pump station, the blade angle monitoring values ​​of each unit of the pump station and other water / work conditions data. After data cleaning to remove null values, outliers and abnormal values, the adaptive Gaussian filtering method is used to filter out noise while retaining the original signal change information to the greatest extent; the filtering process of the adaptive Gaussian filtering method is,

[0053] S(t,σ)=f(t)*g(t,σ)

[0054] Among them, S(t,σ) is the smoothed result of the signal or data after Gaussian filtering; f(t) is the original signal or data sequence, which is a function that changes with time t; t is the time variable; g(t,σ) is the Gaussian kernel function; * is the convolution operator; σ is the standard deviation of the Gaussian function.

[0055]

[0056] Where m is the sampling time point; f(m) is the historical data signal; f k+1 (m) is the output value after smoothing at the k+1th iteration; f k (m+j) is the value of the signal or data at position m+j in the kth iteration; w k (m+j,σ) is the weight of each point in the window; j is the offset in the convolution window, which represents the time offset relative to m; M is the width parameter of the convolution window.

[0057] The confidence method is as follows: after the data is denoised, the error probability distribution function of the steady-state data of direct monitoring indicators such as flow Q and water level H is obtained, so as to determine the change threshold of the indicator under the premise of a certain confidence interval as the basis for judging the steady state. When all indicators are in a steady state, the operating point can be considered as the steady-state operating point of the water pump device.

[0058] 2. Cluster analysis to select typical operating points

[0059] The K-means clustering method is used to cluster the steady-state operating data of the water pump device. By iteratively merging a large number of similar operating conditions, the typical operating points of the water pump device are screened out to avoid the excessive density of operating points in some intervals affecting the overall correction effect.

[0060] Specifically: because the data points of the unit operating conditions of the same pump station are concentrated and have high similarity during operation, the point density is too large in some common operating condition intervals, and too small in uncommon operating condition intervals, which may cause overfitting during the correction process; therefore, the K-means clustering method is considered to cluster the long sequence of historical data, and a large number of similar operating conditions are merged through iteration, so as to facilitate the subsequent correction of the pump device characteristic curve faster and more accurately. The K-means clustering method specifically includes the following steps:

[0061] 2.1. Select K discrete initial cluster centers through K-means++ algorithm.

[0062] 2.2. Assign the data set to the cluster corresponding to the cluster center through the classic K-means++ algorithm.

[0063] 2.3. Calculate the mean of each cluster and use this mean as the new cluster center.

[0064] 2.4. Repeat steps 2.2 and 2.3 until the cluster center remains unchanged or changes very little (such as the change is less than the set threshold).

[0065] 2.5. Use the elbow method to determine whether it is the optimal K value. If not, go to step 2.1. If so, divide the typical operating points into K cluster centers to form the unit's correction training set (i.e., typical operating points).

[0066] 3. Extraction of original characteristic curve

[0067] Based on the typical operating points of the water pump device, the original characteristic curve of the water pump device is obtained by using the moving least square method or cubic spline or polynomial fitting.

[0068] Specifically, the characteristic curve of the water pump device can be numerically processed by using a variety of methods such as moving least squares method, cubic spline and polynomial. The present invention intends to use a two-variable quadratic polynomial method to fit the flow-head-blade angle characteristic curve of the water pump device, which is in the following form:

[0069] θ=p00+p10*Q+p01*H+p20*Q 2 +p11*Q*H+p02*H 2

[0070] Where θ is the blade placement angle of the water pump device (unit: °); Q is the flow rate of the water pump device (unit: m 3 / s); H is the head of the water pump device (unit: m). p00, p10, p01, p20, p11, and p02 are the coefficients of each term in the fitted polynomial equation.

[0071] 4. Characteristic curve correction based on typical operating points

[0072] Parameters are introduced to transform the original characteristic curve of the water pump device to realize three-dimensional affine transformation, and the characteristic curve of the water pump device is corrected based on the Dogbox algorithm in the trust region algorithm.

[0073] Specifically: introduce appropriate parameters to transform the initial characteristic curve equation to achieve three-dimensional affine transformation. Introduce parameters a, b, c, r to transform the characteristic curve equation obtained in the third part. The equation after linear transformation is as follows:

[0074]

[0075] Among them, a, b, c, and r are new parameters introduced by the modified equation; Q′ is the flow value of the water pump device after correction; H′ is the head value of the water pump device after correction; θ′ is the blade placement angle value of the water pump device after correction.

[0076] Assume that the normal operating range of the pump group is Q∈[Q min ,Q max ],H∈[H min ,H max ],θ∈[θ min ,θ max ]; then the search domains are specified as: a∈[-0.3*|Q min -Q max |,0.3*|Q min -Q max |],b∈[-0.3*|H min -H max |,0.3*|H min -H max |],c∈[-0.3*|θ min -θ max |,0.3*|θ min -θ max |]; Q min and Q max are the minimum flow rate and maximum flow rate of the water pump device respectively; H min and H max are the minimum head and maximum head of the water pump device respectively; θ min and θ max They are the minimum and maximum angles for placing the blades of the water pump device respectively.

[0077] In order to solve the high-dimensional nonlinear optimization problem caused by the three-dimensional affine transformation process, based on the Dogbox algorithm in the trust region algorithm, the fitting parameters are determined by minimizing the residual sum of squares, that is, the characteristic curve is translated and rotated in the three-dimensional space concept, so that the characteristic curve is close to the selected typical operating points and the characteristic curve correction is completed.

[0078] 5. Extraction of characteristic curve after correction

[0079] Use origin software to draw the corrected flow-head-blade angle characteristic curve.

[0080] Specifically: using origin software, the corrected characteristic curve equation obtained in the fourth part is discretized, and the corrected integer angle contour map is drawn, that is, the corrected flow-head-blade angle characteristic curve map is obtained.

[0081] Embodiment 2

[0082] In order to better illustrate the execution process of the method of the present invention, this embodiment takes the 3# water pump unit of a pump station as an example, and performs the blade angle characteristic curve correction and result display of the water pump device based on this method, which specifically includes the following contents:

[0083] 1. Data Collection and Preprocessing

[0084] The water conditions, working conditions and other related data of the 3# pump unit of the pump station during the scheduling period from 2022.10.21 to 2023.4.6 are collected and summarized, and the steady-state working condition data are selected from the long sequence monitoring data by using the adaptive Gaussian filtering method and the confidence method. The data requirements are shown in Table 1.

[0085] Table 1 Data requirements table

[0086]

[0087] 2. Cluster analysis to select typical operating points

[0088] The selected monitoring data of the 3# pump unit of the pump station is clustered by the K-means clustering method for long-sequence historical data, and a large number of similar working conditions are merged through iteration to facilitate the subsequent correction of the characteristic curve of the pump unit. The specific steps are as follows:

[0089] (1) Select K discrete initial cluster centers using the K-means++ algorithm.

[0090] (2) Assign the data set to the cluster corresponding to the cluster center through the classic K-means++ algorithm.

[0091] (3) Calculate the mean of each cluster and use this mean as the new cluster center.

[0092] (4) Go to (2) and (3) until the cluster center remains unchanged or changes very little.

[0093] (5) Use the elbow method to determine whether it is the optimal K value. If not, go to (1).

[0094] Finally, the optimal K value for unit clustering was determined to be 200, that is, the previous 10,276 operating points were divided into 200 cluster centers through clustering to form the correction training set of the unit.

[0095] 3. Extraction of original characteristic curve

[0096] The characteristic curve of the water pump device is numerically processed, and the flow-head-blade angle characteristic curve of the water pump device is fitted by using the method of binary quadratic polynomial, which is in the following form:

[0097] θ=-20.694+0.267*Q-0.625*H+0.0025*Q 2 +0.013*Q*H+0.127*H 2

[0098] Where θ is the blade placement angle of the water pump device (°), Q is the flow rate of the water pump device (m 3 / s), H is the head of the water pump device (m). Its correlation coefficient R 2 ≥0.98 has met the relevant accuracy requirements.

[0099] 4. Characteristic curve correction based on typical operating points

[0100] (1) Introducing appropriate parameters and search domain: Introducing parameters a, b, c, and r to transform the characteristic curve equation obtained in the third part, the equation after linear transformation is as follows:

[0101]

[0102] Assume that the normal operating range of the pump group is Q∈[25m 3 / s,40m 3 / s], H∈[3.8m,6.5m], θ∈[-7°,-1°]; then the search domains are specified as: a∈[-7.5,7.5], b∈[-2,2], c∈[-2,2].

[0103] (2) Based on the Dogbox algorithm in the trust region algorithm, the fitting parameters are determined by minimizing the residual sum of squares, that is, the characteristic curve is translated and rotated in the three-dimensional space concept, so that the characteristic curve approaches the selected typical operating point and completes the characteristic curve correction.

[0104] 5. Extraction of characteristic curve after correction

[0105] Using the origin software, the corrected curve equation obtained in the fourth part is discretized, and the corrected integer angle contour map is drawn, that is, the corrected flow-head-blade angle characteristic curve is obtained, as shown in Figure 2 The correction effect is shown in Table 2.

[0106] Table 2. Angle calculation accuracy index after correction

[0107]

[0108] By adopting the above technical solution disclosed in the present invention, the following beneficial effects are obtained:

[0109] The present invention provides a method for correcting the characteristic curve of a water pump device based on cluster analysis and a trust region algorithm. The present invention selects steady-state operating data from long-sequence monitoring data by combining an adaptive Gaussian filtering method with a confidence method. The adaptive Gaussian filtering method can retain the original signal change information to the greatest extent while filtering out noise. The present invention performs clustering processing on long-sequence historical data based on the K-means clustering method, greatly reduces the amount of data, improves the calculation speed, and makes the data points cover as many working conditions as possible, avoiding the curve overfitting phenomenon that may occur in the linear regression process due to the excessive density of point sets in some common working condition intervals during the operation of the same pump station. The present invention introduces new parameters to transform the original characteristic curve equation for affine transformation, and makes the curve fit the data points as much as possible without destroying the original curve space structure, avoiding the abnormal deformation of the characteristic curve that does not conform to the hydraulic characteristics of the water pump caused by direct fitting, thereby improving the curve correction quality. The present invention performs parameter optimization based on the Dogbox algorithm in the trust region algorithm, and has good robustness, stability and convergence characteristics when solving the high-dimensional nonlinear parameter optimization problem that may occur due to the introduction of new parameters. The present invention has strong versatility, and can be used to process and generate corrected water pump device characteristic curves for axial flow / mixed flow pump stations with long-sequence historical monitoring data.

[0110] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be considered as the scope of protection of the present invention.

Claims

1. A method for correcting a characteristic curve of a water pump device based on cluster analysis and trust region algorithm, characterized in that: The following steps are included: S1. Cluster analysis to select typical operating points: The K-means clustering method is used to cluster the steady-state operating data of the water pump device, and the typical operating points of the water pump device are screened out by iteratively merging a large number of similar operating conditions. S2. Extraction of original characteristic curve: Based on the typical operating points of the water pump device, the original characteristic curve of the water pump device is obtained by using the moving least square method or cubic spline or polynomial fitting; S3, Characteristic curve correction based on typical operating points: Parameters are introduced to transform the original characteristic curve of the water pump device to realize three-dimensional affine transformation, and the characteristic curve of the water pump device is corrected based on the Dogbox algorithm in the trust region algorithm.

2. The method for correcting the characteristic curve of a water pump device based on cluster analysis and trust region algorithm according to claim 1, characterized in that: The K-means clustering method in step S1 specifically includes the following contents: S11. Select K discrete initial cluster centers through K-means++ algorithm; S12, assign the data set to the cluster corresponding to the cluster center through the classic K-means++ algorithm; S13, calculating the mean of each cluster and using the mean as the new cluster center; S14, repeating steps S12 to S13 until the cluster center remains unchanged or changes less than the change threshold; S15. Use the elbow method to determine whether it is the optimal K value. If not, go to step S11. If so, divide the steady-state operating data of the water pump device into K cluster centers to screen out typical operating points of the water pump device.

3. The method for correcting the characteristic curve of a water pump device based on cluster analysis and trust region algorithm according to claim 2, characterized in that: Step S2 specifically includes: The flow-head-blade angle characteristic curve of the water pump device is fitted by the method of binary quadratic polynomial, and the original characteristic curve equation of the water pump device is obtained, which is in the following form: θ=p00+p10*Q+p01*H+p20*Q 2 +p11*Q*H+p02*H 2 Among them, θ is the blade installation angle of the water pump device; Q is the flow rate of the water pump device; H is the head of the water pump device; p00, p10, p01, p20, p11, and p02 are the coefficients of each term in the fitted polynomial equation.

4. The method for correcting the characteristic curve of a water pump device based on cluster analysis and trust region algorithm according to claim 3 is characterized in that: Step S3 specifically includes the following contents: S31, introduce parameters a, b, c, r to transform the original characteristic curve equation obtained in step S2 to achieve three-dimensional affine transformation. The equation after linear transformation is as follows: Among them, a, b, c, r are new parameters introduced by the modified equation; Q' is the flow value of the water pump device after correction; H' is the head value of the water pump device after correction; θ' is the blade placement angle value of the water pump device after correction; Assume that the normal operating range of the water pump device is Q∈[Q min ,Q max ],H∈[H min ,H max ],θ∈[θ min ,θ max ]; the search domains are specified as a∈[-0.3*|Q min -Q max |,0.3*|Q min -Q max |],b∈[-0.3*|H min -H max |,0.3*|H min -H max |],c∈[-0.3*|θ min -θ max |,0.3*|θ min -θ max |]; Q min and Q max are the minimum flow rate and maximum flow rate of the water pump device respectively; H min and H max are the minimum head and maximum head of the water pump device respectively; θ min and θ max They are the minimum and maximum angles for placing the blades of the water pump device respectively. S32. Based on the Dogbox algorithm in the trust region algorithm, the fitting parameters are determined by minimizing the residual sum of squares, that is, the characteristic curve is translated and rotated in the three-dimensional space concept, so that the characteristic curve approaches the selected typical operating points, and the characteristic curve correction of the water pump device is completed.

5. The method for correcting the characteristic curve of a water pump device based on cluster analysis and trust region algorithm according to claim 1, characterized in that: Before step S1, the method further includes: S0. Data collection and preprocessing: According to the actual engineering scheduling and operation conditions at different time scales, the water conditions and working conditions of the water pump device are statistically summarized, the data is processed using data cleaning methods, and the steady-state operating data of the water pump device is screened out using the adaptive Gaussian filtering method and the confidence method.

6. The method for correcting the characteristic curve of a water pump device based on cluster analysis and trust region algorithm according to claim 5, characterized in that: Water and work condition related data include the water level below the pump station, the water level above the pump station, the flow monitoring value of each unit, and the blade angle monitoring value of each unit; Data processing includes removing null values, outliers and abnormal values.

7. The method for correcting the characteristic curve of a water pump device based on cluster analysis and trust region algorithm according to claim 5, characterized in that: In step S0, the noise is filtered using the adaptive Gaussian filtering method, and the calculation formula is: S(t,σ)=f(t)*g(t,σ) Among them, S(t,σ) is the smoothing result of the signal or data after Gaussian filtering; f(t) is the original signal or data sequence, which is a function that changes with time t; t is the time variable; g(t,σ) is the Gaussian kernel function; * is the convolution operator; σ is the standard deviation of the Gaussian function; Where m is the sampling time point; f(m) is the historical data signal; f k+1 (m) is the output value after smoothing at the k+1th iteration; f k (m+j) is the value of the signal or data at position m+j in the kth iteration; w k (m+j,σ) is the weight of each point in the window; j is the offset in the convolution window, which represents the time offset relative to m; M is the width parameter of the convolution window.

8. The method for correcting the characteristic curve of a water pump device based on cluster analysis and trust region algorithm according to claim 7, characterized in that: The specific confidence method is: After the data is denoised by the adaptive Gaussian filtering method, the error probability distribution function of the steady-state data of the direct monitoring indicators is obtained, so that the change threshold of the indicators can be determined under the premise of a certain confidence interval as the basis for judging the steady state; when all indicators are in a steady state, the operating point is considered to be a steady-state operating point.

9. The method for correcting the characteristic curve of a water pump device based on cluster analysis and trust region algorithm according to claim 1, characterized in that: Step S3 also includes: S4. Extraction of corrected characteristic curve: Use origin software to draw the corrected flow-head-blade angle characteristic curve.

10. The method for correcting the characteristic curve of a water pump device based on cluster analysis and trust region algorithm according to claim 9, characterized in that: Step S4 specifically includes: The corrected characteristic curve equation obtained in step S3 is discretized by using origin software, and a corrected integer angle contour map is drawn to obtain a corrected flow-head-blade angle characteristic curve map.

Images