Pump station group collaborative scheduling method and system based on digital twinning

Abstract

The application discloses a pump station group cooperative scheduling method and system based on digital twinning, and belongs to the technical field of optimization control based on digital twinning. The method comprises the following steps: collecting multi-source heterogeneous data of the pump station group, constructing a basic framework of a digital twinning model of each pump station group, and synchronously updating; deeply fusing health degree, energy efficiency ratio, load capacity and response speed data, and calculating a pump station scheduling response willingness factor; using an improved algorithm, embedding a dynamic scheduling capability weight vector into a balance fitness function and a response weighted search mechanism for solving; using the digital twinning model of the pump station group, pre-animating and verifying a cooperative scheduling optimization scheme of each pump station, confirming hydraulic balance and operation safety, generating target rotating speed and start-stop instructions, and delivering the instructions to each pump station for execution. The application quantifies the pump station scheduling response willingness by constructing a digital twinning model, thereby guiding cooperative optimization, and realizing willingness distribution driven scheduling.

Description

Technical Field

[0001] This invention relates to the field of optimization control technology based on digital twins, and in particular to a collaborative scheduling method and system for pump station groups based on digital twins. Background Technology

[0002] With the continuous advancement of national water conservancy modernization, pumping stations, as core hub facilities in water conservancy projects, are widely used in multiple key areas such as flood control and drainage, farmland irrigation, urban water supply, and inter-basin water transfer. Their operational efficiency and scheduling rationality are directly related to the optimal allocation of water resources, water security, and energy conservation levels, and are of great significance to regional economic and social development and ecological environment stability. In recent years, with the frequent occurrence of extreme climate events and the continuous increase in the intensity of watershed development, the traditional pumping station group scheduling mode is no longer suitable for complex and ever-changing hydrological conditions and diversified scheduling needs. Promoting the transformation of pumping station group scheduling towards digitalization, intelligence, and collaboration has become an inevitable trend in the industry. At present, the application of digital twin technology in pumping station individual operation monitoring and fault early warning has become increasingly mature. Some studies have attempted to apply digital twin technology to pumping station group scheduling, and have initially realized the visualization monitoring and scheduling optimization of pumping station group operation status.

[0003] However, given the current state of technological development, there are still substantial defects, including the following: 1. Fixed parameters, which makes it impossible to accurately depict the real-time differentiated operating status of pump stations; 2. The use of centralized scheduling ignores the real-time status differences and asynchronous coordination intentions between pump stations, leading to conflicts between global optimization and local feasibility; 3. Inability to effectively utilize real-time status information to guide the search direction; 4. Lack of safety verification before the execution of scheduling schemes, which leads to the triggering of equipment protection mechanisms and hydraulic shocks after the schemes are issued. Summary of the Invention

[0004] This application provides a collaborative scheduling method and system for pump station groups based on digital twins. The core of the method is to dynamically quantify the pump station scheduling response intention by constructing a digital twin model, and to guide the improvement of the firefly algorithm for collaborative optimization, so as to solve the above four existing technical problems.

[0005] Therefore, this application provides the following solution: This application discloses a collaborative scheduling method for pump station groups based on digital twins, comprising: S1, collecting multi-source heterogeneous data of the pump station group, constructing the basic framework of the pump station group digital twin model and updating it synchronously; S2, based on the pump station group digital twin model, using a nonlinear mapping algorithm to fuse health, energy efficiency ratio, load capacity and response speed data, calculating the scheduling response willingness factor of each pump station, and obtaining a dynamic scheduling capability weight vector; S3, using an improved firefly algorithm, embedding the dynamic scheduling capability weight vector into a balance fitness function and a response weighted search mechanism for solving, and obtaining a collaborative scheduling optimization scheme for each pump station; S4, using the pump station group digital twin model, pre-simulating and verifying the collaborative scheduling optimization scheme for each pump station, confirming hydraulic balance and operational safety, generating target speed and start / stop commands and issuing them to each pump station for execution.

[0006] Furthermore, its S1 includes: S101. Collect multi-source heterogeneous data from the pump station group, including hydraulic data, equipment status data, energy consumption data, and environmental data, as the raw data stream; S102. Clean and preprocess the raw data stream to obtain a structured real-time data vector; S103. Construct a hydraulic characteristic twin model of the pump station using a multinomial fitting method, construct an equipment health twin model using a proportional risk model based on Weibull distribution, construct an energy consumption twin model using a real-time energy efficiency ratio mapping model, and complete the construction of a digital twin model of the pump station group. S104. Based on the structured real-time data vector, the recursive least squares method is used to correct the hydraulic characteristic curve parameters online and calculate the health index, energy efficiency ratio, load capacity and response speed to obtain the real-time state vector.

[0007] Furthermore, its S2 includes: S201. Receive the real-time status vector and extract the health index, energy efficiency ratio, load capacity and response speed components from it to obtain a four-dimensional independent vector. S202. Calculate the willingness contribution value of the health index, energy efficiency ratio, load capacity, and response speed component dimensions in the four-dimensional independent vector; the formula for calculating the willingness contribution value of the health index is: In the formula F H (HI i ) contributes to health intention; HI i The equipment health index is represented by k1, the kurtosis coefficient by θ1, and the health threshold by e. The energy efficiency ratio calculation formula is as follows: In the formula F E (EER i Energy Efficiency Contribution (EER) is the contribution value of the willingness to improve energy efficiency. i Real-time energy efficiency ratio; EER 优For optimal energy efficiency reference value, 80% of the pump station's rated efficiency is taken; EER min The minimum acceptable energy efficiency ratio is 50% of the pump station's rated efficiency. This is the normalized value of the energy efficiency compliance rate; the formula for calculating the intended contribution value of load capacity is: In the formula F L (LC i ) represents the load willingness contribution value; -k2 is the attenuation coefficient; e is the natural constant; LC i The remaining load capacity; the formula set for calculating the willingness contribution value for response speed is as follows: In the formula F R (RS i ) represents the contribution value of the willingness to respond; RS i For response speed index; RS max The maximum effective response speed exponential threshold is set to a typical value of 1.5.

[0008] Furthermore, its S2 also includes: S203. The health willingness contribution value, energy efficiency willingness contribution value, load willingness contribution value and response willingness contribution value are weighted and combined to form the dispatch response willingness factor, with the weight allocation being 0.3→0.25→0.25→0.2 respectively; S204. Normalize the dynamic scheduling capability weight vector of each pump station to generate a dynamic scheduling capability weight vector for improving the response weighted initialization of the firefly algorithm.

[0009] Furthermore, its S3 includes: S301. Construct a mathematical model with the goal of minimizing total energy consumption and embedding weight constraint penalties using optimization modeling methods, and initialize the control parameters of the improved firefly algorithm. S302. After initialization, obtain the dynamic scheduling capability weight vector, and use the weight-guided scaling factor to perform non-uniform amplification on the uniform random number to obtain the initial population that preferentially gathers in the high flow area of ​​the high-willing pump station. S303. Using the dynamic scheduling capability weight vector, construct a balanced fitness function carrying the intention reverse penalty coefficient, evaluate the fitness of each individual in the initial population, and obtain a comprehensive fitness value that reflects energy consumption, task completion degree and constraint penalty.

[0010] Furthermore, its S3 also includes: S304. Using the intention-coupled attraction method, the brightness of the comprehensive fitness value, which reflects energy consumption, task completion degree and constraint penalty, is compared and the position is updated to obtain the scheduling flow allocation solution for the new generation of individuals in the population. S305. Using the elite retention strategy, the fitness of the scheduling flow allocation solution is selected for optimal retention and convergence determination. Finally, the global optimal solution is extracted as the collaborative scheduling optimization scheme for each pumping station.

[0011] Furthermore, its S4 includes: S401. Perform integrity verification, supply and demand balance deviation calculation, and secondary confirmation of health hard constraints on the collaborative scheduling optimization scheme to obtain an executable scheduling scheme that meets the feasibility prediction. S402. Using the digital twin model of the pump station group, hydraulic transient simulation and dynamic speed calculation are performed on the executable scheduling scheme to evaluate the stability of the transient process and obtain the target speed and stability verification results of each pump station. S403. By using power verification, vibration and temperature trend prediction, and start-stop cycle constraint checks, the target speed and stability verification results of each pump station are filtered by safety boundaries to obtain the final executable scheduling scheme that passes all safety constraints.

[0012] Furthermore, its S4 also includes: S404. Convert the target flow rate in the final executable scheduling scheme into target speed and valve opening, and generate control command data packets for each pump station. S405. Send control command data packets to each pump station to execute operations, and build a log database to store the operation evidence.

[0013] Furthermore, a collaborative scheduling system for pump station groups based on digital twins includes: The first module is used to collect multi-source heterogeneous data from the pump station group, build the basic framework for constructing a digital twin model of the pump station group, and update it synchronously. The second module is used to integrate health status, energy efficiency ratio, load capacity and response speed data to calculate the scheduling response willingness factor of each pumping station. The third module is used to embed the dynamic scheduling capability weight vector into the balanced fitness function and response weighted search mechanism for solving; The fourth module is used to pre-test and verify the coordinated scheduling optimization scheme of each pumping station. After confirming hydraulic balance and operational safety, it generates target speed and start / stop commands and sends them to each pumping station for execution.

[0014] Compared with the prior art, this application achieves at least one of the following beneficial effects: 1. A method combining multi-source heterogeneous data acquisition with online correction using recursive least squares is employed to process hydraulic data, equipment status, energy consumption, and environmental data of a pumping station group, and to construct and synchronously update digital twin models. This yields real-time state vectors for each pumping station, solving the problem that existing technical parameters are fixed, which makes it impossible to accurately depict the real-time differentiated operating status of pumping stations.

[0015] 2. By using a nonlinear mapping function fusion and weighted normalization method, the four-dimensional state data of health, energy efficiency ratio, load capacity and response speed are nonlinearly transformed and fused to generate a scheduling response willingness factor, realizing the paradigm shift from task-driven to capability-driven scheduling. This solves the problem that the existing technology uses centralized scheduling to ignore the real-time state differences and asynchronous coordination willingness between pump stations, resulting in a conflict between global optimality and local feasibility.

[0016] 3. An improved firefly algorithm with embedded weight vectors is used to solve the optimization problem guided by dynamic weight vectors through a balanced fitness function with response-weighted initialization, intention-backward penalty, and intention-coupled attraction mechanism. The optimal scheme for coordinated scheduling flow allocation of each pumping station is obtained, which solves the problem that the optimization algorithm used in the existing technology cannot effectively use real-time status information to guide the search direction.

[0017] 4. By combining digital twin dynamic simulation pre-play with multi-dimensional safety boundary constraint filtering, the scheduling scheme is subjected to hydraulic transient simulation, speed calculation and item-by-item verification of power, vibration, temperature and start-stop constraints. Executable control commands are generated and issued through industrial communication protocols, realizing a closed-loop connection from optimization scheme to physical execution. This solves the problem that the existing technology lacks safety verification before the scheduling scheme is executed, which leads to the triggering of equipment protection mechanisms and hydraulic shock after the scheme is issued. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments or examples of the present invention, the drawings used in the embodiments or examples will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained according to these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a collaborative scheduling method for pump station groups based on digital twins according to the present invention. Figure 2 This is an architecture diagram of a collaborative scheduling system for pump station groups based on digital twins, according to the present invention. Detailed Implementation

[0020] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0021] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0022] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0023] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0024] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0025] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0026] The following are examples provided: See appendix Figure 1A collaborative scheduling method for pump station groups based on digital twins includes: S1, collecting multi-source heterogeneous data of the pump station group, constructing the basic framework of the pump station group digital twin model and updating it synchronously; S2, based on the pump station group digital twin model, using a nonlinear mapping algorithm to fuse health, energy efficiency ratio, load capacity and response speed data, calculating the scheduling response willingness factor of each pump station, and obtaining a dynamic scheduling capability weight vector; S3, using an improved firefly algorithm, embedding the dynamic scheduling capability weight vector into a balance fitness function and a response weighted search mechanism for solving, obtaining the collaborative scheduling optimization scheme of each pump station; S4, using the pump station group digital twin model, pre-running and verifying the collaborative scheduling optimization scheme of each pump station, confirming hydraulic balance and operational safety, generating target speed and start / stop commands and issuing them to each pump station for execution.

[0027] The core technical points of this invention include: 1- Taking the scheduling response willingness factor driven by the real-time status of pump stations as the theoretical core, for the parameters of the four dimensions of health index, real-time energy efficiency ratio, remaining load capacity and response speed index output by the digital twin model of each pump station, mapping functions (including sigmoid function, truncated linear function, exponential function and piecewise linear function) that conform to physical nonlinear characteristics are constructed respectively. After weighted fusion, a dynamic scheduling capability weight vector is generated. The weight reflects the strength of each pump station's willingness to undertake scheduling tasks in real time, providing state guidance signals for the optimization algorithm; 2- Improving key parameters based on the firefly algorithm. Improvement 1: Embedding DRW weights as inverse dynamic coefficients in the constraint penalty term of the fitness function to construct a balanced fitness function, which automatically tends to allocate traffic to high-willing pump stations; Improvement 2: Introducing a response weighted search mechanism, using DRW weights to guide population initialization and correct the calculation of attraction between individuals, so that the search process focuses on the high-willing region; The improved method reduces the search space and improves the convergence speed, and reduces system complexity without the need to build a complex hierarchical architecture.

[0028] In specific implementation, S1 includes: S101. Collect multi-source heterogeneous data from the pump station group, including hydraulic data, equipment status data, energy consumption data, and environmental data, as the raw data stream; S102. Clean and preprocess the raw data stream to obtain a structured real-time data vector; S103. Construct a hydraulic characteristic twin model of the pump station using a multinomial fitting method, construct an equipment health twin model using a proportional risk model based on Weibull distribution, construct an energy consumption twin model using a real-time energy efficiency ratio mapping model, and complete the construction of a digital twin model of the pump station group. S104. Based on the structured real-time data vector, the recursive least squares method is used to correct the hydraulic characteristic curve parameters online and calculate the health index, energy efficiency ratio, load capacity and response speed to obtain the real-time state vector.

[0029] The process of constructing three twin models in step S103 includes the construction of a hydraulic characteristic twin model, an equipment health twin model, and an energy consumption twin model; The construction of the hydraulic characteristic twin model includes: Step 1) Use the polynomial fitting method to establish the mathematical relationships between flow rate and head, flow rate and efficiency, and flow rate and power to form a hydraulic characteristic twin model; Step 2), the formula for fitting the characteristic curve of flow rate-head based on Step 1) is: In the formula H 1,i (Q) represents the theoretical head of the i-th pump station at flow rate Q (used to describe the pump station's ability to lift fluid to a certain height); Q is the instantaneous flow rate at the pump station outlet; a H0,i The zero-order fitting coefficient (intercept term, determined by both the pump impeller outer diameter and rated speed); a H1,i These are the first-order fitting coefficients (linear terms, describing the linear slope of the head as a function of flow rate); a H2,i These are the second-order fitting coefficients (nonlinear terms that describe the degree of nonlinear curvature of the head as the flow rate changes). Step 3), the formula for fitting the flow-efficiency characteristic curve based on Step 1) is: In the formula η 2,i (Q) is the theoretical operating efficiency of the i-th pump station at flow rate Q (used to describe the conversion efficiency of the pump station in converting electrical energy into fluid potential energy). , , and These are the zero-order, first-order, second-order, and third-order fitting coefficients, respectively (the zero-order coefficient is used as the efficiency baseline constant, the first-order coefficient is used to describe the slope of the linear change in efficiency with flow rate, the second-order coefficient is used as the curvature term of the efficiency curve, and the third-order coefficient is used as a higher-order correction term to more accurately fit the asymmetric characteristics of rapid efficiency decline in the high flow rate region). Step 4), based on steps 2) and 3), fits the flow-power characteristic curve. The formula for fitting the characteristic curve is: In the formula P 3,i (Q) represents the shaft power of the i-th pump station at flow rate Q (used to determine the energy consumption of the pump station operation); ρ represents the density of the fluid being transported; g represents the acceleration due to gravity (assuming a standard value of 9.81 m / s²). 2 ); H 1,i (Q) is the theoretical head of the i-th pump station at flow rate Q (obtained in step 2); η 2,i (Q) represents the theoretical operating efficiency of the i-th pump station at flow rate Q (obtained in step 3); 10 -3 Unit conversion factor (converting watts (W) to kilowatts (kW)); Step 5) Based on steps 1) to 4), fit the three mathematical relationships to the values ​​(i.e., H).1,i (Q), η 2,i (Q) and P 3,i (Q)) are combined to form a set of fundamental hydraulic characteristic functions, which serve as a twin model of hydraulic characteristics.

[0030] The construction of the device health twin model includes: Step 1) Use a proportional hazards model based on the Weibull distribution to describe the evolution of the health index with key monitoring quantities (including cumulative vibration degradation, temperature effect degradation, and runtime degradation); Step 2), based on Step 1), the calculation formula for vibration cumulative degradation is as follows: In the formula, △HI 振动,i V represents the amount of equipment health degradation caused by vibration. 实测,i V represents the current measured vibration amplitude of the i-th pumping station; 阈值,i The vibration amplitude alarm threshold (the rated value at the time of equipment leaving the factory); c v,i β is the vibration degradation weighting coefficient (reflecting the sensitivity of the health impact of vibration on this type of equipment, determined by fitting historical maintenance data); 振 The vibration degradation shape parameter (used to control the nonlinearity of the degradation curve; when >1, it indicates superlinear accelerated degradation). Step 3), based on Step 1), the formula for calculating the degradation due to temperature effects is: In the formula, △HI 温度,i Temperature-induced health degradation; T 实测,i T represents the currently measured temperature of the bearing (or winding) at the i-th pump station; 参,i The normal reference operating temperature (i.e., the operating procedure for the equipment); T max,i The maximum permissible operating temperature (i.e., the insulation class of the equipment); β 温 This is the shape parameter for temperature degradation (used to control the nonlinearity of temperature-accelerated degradation). Step 4), based on Step 1), the calculation formula for runtime degradation is: In the formula, △HI 时长,i T represents the amount of health degradation caused by accumulated runtime. 时长,i T represents the cumulative operating time of the i-th pumping station; 总,i c is the total lifespan of the equipment during its design phase; r,i β is the runtime degradation weighting coefficient (obtained by fitting based on equipment life statistics); 振 For runtime degradation shape parameter (β) 振 When β > 1, it indicates accelerated wear in the later stages. 振 When the value is less than 1, it indicates rapid early degradation. Step 5), based on steps 1) to 4), the general formula for constructing the device health twin model is: HI i (t)=HI 0,i -△HI 振动,i -△HI 温度,i -△HI 时长,i HI i (t) represents the comprehensive health index of the i-th pump station at time t (as the final parameter of the equipment health twin model); △HI 振动,i △HI represents the amount of equipment health degradation caused by vibration. 温度,i Temperature-induced health degradation; △HI 时长,i This refers to the amount of health degradation caused by accumulated runtime.

[0031] The energy consumption twin model is constructed by defining it as an energy efficiency ratio (EER) mapping model based on real-time operating conditions. The real-time energy efficiency ratio of the model is defined by the following formula: , where EER i (t) represents the real-time energy efficiency ratio (directly quantifying the efficiency of converting electrical energy into hydraulic energy under the current operating conditions); ρ represents the fluid density; and g represents the acceleration due to gravity (assuming a standard value of 9.81 m / s²). 2 ); Q i (t) represents the instantaneous flow rate of the i-th pump station at time t (used to measure the pump station's output capacity); H i (t) represents the actual head of the i-th pump station at time t (reflecting the pump station's ability to overcome pipeline resistance and raise the water level); P i (t) represents the input active power of the i-th pump station at time t (the electrical power absorbed by the pump station from the grid); ρgQ i (t)H i (t) represents the hydraulic power output of the pumping station (i.e., the theoretical effective power).

[0032] It uses the recursive least squares method to correct the hydraulic characteristic curve parameters online. The problem is defined as follows: in actual operation, due to impeller wear, changes in pipeline characteristics, and other factors, the flow-head curve may deviate from the basic model. Therefore, the fitting coefficients need to be dynamically updated. The definition formulas for the observation vector and the parameter vector to be estimated are as follows: In the formula, φ(t) is the observation vector; θ(t) is the parameter vector to be estimated; and 1 is a constant term. and These correspond to the first and second terms of the observation vector, respectively; , and These correspond to the first and second terms of the parameter vector to be estimated, respectively; T is the vector transpose; when new measured points are obtained, RLS update and correction are performed, and the update and correction formula set is as follows: In the formula, K(t) is the gain vector (used to determine the strength and direction of the correction of the parameter estimate by the current observation data); P(t-1) is the error covariance matrix of the previous time step; φ(t) is the observation vector (obtained by defining the formula set); λ is the forgetting factor (values ​​0.95-0.99); φ T (t)·P(t-1)·φ(t) is the result of the quadratic operation (reflecting the information increment that the current observation point can provide); This is the parameter estimation vector updated at the current time (used to update the model); The parameter estimation vector for the previous time step (obtained by defining a set of formulas); H i (t) represents the measured head at the current moment; The predicted head is based on the parameters of the previous moment and the current flow rate; P(t) is the prediction error (i.e., the difference between the actual measured value and the model prediction value); P(t) is the updated error covariance matrix at the current time step (used to calculate the gain vector in the next recursive step to achieve closed-loop update); P(t-1) is the error covariance matrix at the previous time step; I is a 3×3 identity matrix (used to ensure the correctness of the dimensions). It is the outer product of the gain vector and the observation vector (used to describe the information removal effect of the current observation on the covariance matrix).

[0033] The specific implementation method for calculating its health index, energy efficiency ratio, load capacity and response speed is as follows: The real-time vibration, temperature and running time in the structured real-time data vector are input into the digital twin model of the pump station group (which includes the equipment health twin model, hydraulic characteristic twin model and energy consumption twin model) for separate processing. 1- Use the equipment health twin model to calculate the health index at the current moment; 2- Use the hydraulic characteristic twin model to calculate the current energy efficiency ratio; 3- Use the energy consumption twin model to calculate the remaining load capacity and response speed index, and combine the health index, energy efficiency ratio, load capacity and response speed to form a real-time state vector.

[0034] In specific implementation, S2 includes: S201. Receive the real-time status vector and extract the health index, energy efficiency ratio, load capacity and response speed components from it to obtain a four-dimensional independent vector. S202. Calculate the willingness contribution value of the health index, energy efficiency ratio, load capacity, and response speed component dimensions in the four-dimensional independent vector; the formula for calculating the willingness contribution value of the health index is: In the formula F H (HI i ) contributes to health intention; HI iHere, k1 is the equipment health index; θ1 is the kurtosis coefficient (typically 8); e is the health threshold (typically 0.6); and e is the natural constant. The energy efficiency ratio calculation formula is as follows: In the formula F E (EER i Energy Efficiency Contribution (EER) is the contribution value of the willingness to improve energy efficiency. i Real-time energy efficiency ratio; EER 优 For optimal energy efficiency reference value, 80% of the pump station's rated efficiency is taken; EER min The minimum acceptable energy efficiency ratio is 50% of the pump station's rated efficiency. This is the normalized value of the energy efficiency compliance rate; the formula for calculating the intended contribution value of load capacity is: In the formula F L (LC i ) represents the load willingness contribution value; -k2 is the attenuation coefficient (typically 3); e is the natural constant; LC i The remaining load capacity; the formula set for calculating the willingness contribution value for response speed is as follows: In the formula F R (RS i ) represents the contribution value of the willingness to respond; RS i For response speed index; RS max The maximum effective response speed exponential threshold is set to a typical value of 1.5.

[0035] In specific implementation, S2 also includes: S203. The health willingness contribution value, energy efficiency willingness contribution value, load willingness contribution value and response willingness contribution value are weighted and combined to form the dispatch response willingness factor, with the weight allocation being 0.3→0.25→0.25→0.2 respectively; S204. Normalize the dynamic scheduling capability weight vector of each pump station to generate a dynamic scheduling capability weight vector for improving the response weighted initialization of the firefly algorithm.

[0036] The above-mentioned normalization process is applied to the dynamic scheduling capacity weight vector of each pumping station, and the formula for calculating the dynamic scheduling capacity weight is as follows: In the formula w i (t) represents the dynamic scheduling capacity weight of the i-th pump station at time t; DRW i (t) represents the scheduling response willingness factor of the i-th pump station at time t (reflecting the willingness of the current pump station to undertake more scheduling tasks at the current time); ε is a very small positive number (taken as 10). -6To prevent the denominator from being zero when the scheduling response willingness factor of all pump stations is zero); N is the total number of pump stations participating in scheduling in the pump station group; i is the dynamic scheduling capability weight index; j is the index used to traverse all pump stations in the pump station group.

[0037] In specific implementation, S3 includes: S301. Construct a mathematical model with the goal of minimizing total energy consumption and embedding weight constraint penalties using optimization modeling methods, and initialize the control parameters of the improved firefly algorithm. S302. After initialization, obtain the dynamic scheduling capability weight vector, and use the weight-guided scaling factor to perform non-uniform amplification on the uniform random number to obtain the initial population that preferentially gathers in the high flow area of ​​the high-willing pump station. S303. Using the dynamic scheduling capability weight vector, construct a balanced fitness function carrying the intention reverse penalty coefficient, evaluate the fitness of each individual in the initial population, and obtain a comprehensive fitness value that reflects energy consumption, task completion degree and constraint penalty.

[0038] Its mathematical model, which aims to minimize total energy consumption and incorporates weighted constraint penalties, includes: Step 1) Define each firefly as a scheduling scheme (i.e., the flow distribution vector of each pump station). Step 2) With minimizing the total energy consumption of the pump station group as the main optimization objective, and considering the completion degree of scheduling tasks and constraint penalties, construct the objective function. The formula for constructing the objective function is as follows: In the formula, F(X) is the fitness function value (a quantitative index for comprehensively evaluating the quality of a scheduling scheme, and the optimization objective is to minimize this value); E 总 (X) represents the total energy consumption of the pump station group (as optimization objective item 1); T 未完成 (X) represents the amount of unfinished scheduled tasks (as optimization objective 2); P 惩 (X,W) represents the comprehensive constraint penalty term (as optimization objective term 3); E 参 For reference energy consumption benchmark values ​​(the average energy consumption value for the same period in history); Q 供 The total target water supply (used to dimensionlessly represent the task deficit); β1, β2, and β3 are the weighting coefficients for target item 1, target item 2, and target item 3, respectively; for target item 1, the calculation formula is: E in the formula 总 (X) represents the total energy consumption of the pump station group; P i (Q i Let Q be the flow rate of the i-th pump station. i The operating power is Δt, which is the scheduling duration (one scheduling cycle is taken as 1 hour); N is the total number of pump stations in the pump station group; i is the pump station group index; for target item 2, its calculation formula is: In the formula, T 未完成(X) represents the amount of unfinished scheduled tasks; Q 供 The total target water supply is N; the total number of pump stations in the pump station group is N; the index of the pump station group is i; Q is Q. i Let be the target flow rate of the i-th pumping station; for target item 3, its calculation formula is: In the formula P 惩 (X,W) represents the comprehensive constraint penalty term; N is the total number of pump stations in the pump station group; i is the pump station group index; λ is the global penalty factor (taking a typical range value of 10). 2 -10 3 ); w i (t) represents the dynamic scheduling capacity weight of the i-th pump station at time t; ε is a very small positive number; Q 越,i This is the upper limit of the safe flow rate for the i-th pump station (a penalty is incurred when the flow rate exceeds the safe limit, and the value is 0 if the flow rate does not exceed the limit). The negative penalty coefficient for intention (when w) i When (t) is large, a smaller coefficient results in a lighter penalty for constraint out-of-bounds errors; when w i When (t) is small, the coefficient is large and the punishment is severe. Step 3) Set upper and lower bound constraints for flow rate. The flow constraint formula set is as follows: In the formula Q 下界,i and Q 上界,i These are the lower and upper bounds of the flow (used to limit the safety margin); Q max,i The maximum flow rate allowed by the factory design of the pumping station; P 额定,i ρ is the rated power of the pumping station; γ is the fluid density; H is the power safety factor (taken as 0.9 to prevent motor overload); i This represents the actual head under the current operating conditions.

[0039] The method of using a weighted scaling factor to perform non-uniform amplification of uniform random numbers is as follows: For the k-th individual (k=1,2,...,M) in the population, its i-th dimension component Q i,k The initialization formula is: In the formula Q represents the initial flow rate of the k-th individual in the initial population at the i-th pump station dimension; i is the pump station group index; k is the individual index; 下界,i and Q 上界,i These are the lower and upper bounds of the flow rate, respectively (obtained from the above set of constraint formulas); r i,k σ(w) represents a random number that follows a uniform distribution in [0,1] (generated independently for each dimension and for each individual). i ) is a weighted scaling factor (used to non-uniformly amplify uniform random numbers); the formula for calculating the weighted scaling factor is: σ(w) i)=0.5+0.5·w i (t)·N, where 0.5 is the basic offset constant (the purpose is to ensure that when w i When (t)=0, the scaling factor still has a value of 0.5, maintaining a basically random distribution); w i (t) represents the dynamic scheduling capacity weight of the i-th pump station at time t; N is the total number of pump stations in the pump station group; 0.5·w i (t)·N is the weight amplification term (used to linearly map weight information to the increment of the scaling factor); its execution logic is as follows: 1- Receive the dynamic scheduling capability weight vector, and independently generate random numbers that follow a uniform distribution [0,1] for each individual k and each dimension i; 2- Calculate the scaling factor using the above scaling factor; 3- Calculate the initial flow value using the above initialization formula; 4- When the health protection hard constraint is triggered (i.e., HI) i If the value is less than 0.5, then the initial flow value is forced to 0 to obtain the initial population.

[0040] Its construction of the equilibrium fitness function carrying the intention-based reverse penalty coefficient includes: Step 1) Define the penalty term for individuals in the initial population as the comprehensive constraint penalty term (i.e., the above objective term 3); Step 2) Based on Step 1), call the hydraulic characteristic twin model in the digital twin model of the pump station group to calculate the response energy consumption of each pump station; use the formula for calculating the amount of unfinished scheduling tasks (i.e., the above objective item 2) to calculate the task completion degree; use the formula for calculating the comprehensive constraint penalty to calculate the constraint penalty; use the constructed mathematical model with the goal of minimizing total energy consumption and embedding weighted constraint penalties to calculate the fitness function value of the three items of response energy consumption, task completion degree and constraint penalty of each pump station, as the comprehensive fitness value.

[0041] In practical implementation, S3 also includes: S304. Using the intention-coupled attraction method, the brightness of the comprehensive fitness value, which reflects energy consumption, task completion degree and constraint penalty, is compared and the position is updated to obtain the scheduling flow allocation solution for the new generation of individuals in the population. S305. Using the elite retention strategy, the fitness of the scheduling flow allocation solution is selected for optimal retention and convergence determination. Finally, the global optimal solution is extracted as the collaborative scheduling optimization scheme for each pumping station.

[0042] The implementation process of brightness comparison and position update includes: Step 1) Define individual brightness (for fireflies, brightness is inversely proportional to fitness value), and the formula for defining brightness is: In the formula I k Let F(X) be the brightness value of the k-th firefly individual (used to describe the superiority of the individual); kThe fitness function value of the k-th individual (i.e., the fitness value of a single individual); 1 is a constant term (used to ensure that the denominator is non-zero and avoid division by zero error); Step 2), based on step 1), for two random individuals j and k, when I j >I k When j is brighter than k, then the attractive force β of j on k is... j,k The formula is defined as follows: In the formula β j,k βj represents the attraction strength of individual j to individual k (used to determine the step size coefficient for individual k to move towards the brighter individual j); β0 is the maximum attraction constant (typically 1); γ is the light absorption coefficient (typically 1.0). Let be the square of the Euclidean distance between individual j and individual k; This is the distance attenuation factor (an attenuation coefficient used to map spatial distance to [0,1]). Let be the intention coupling factor; for the Euclidean distance between individual j and individual k, the formula for calculating the Euclidean distance between individuals is: In the formula r jk Q is the normalized Euclidean distance between individual j and individual k (used to quantify the overall difference in flow allocation between the two scheduling schemes); i,j Let Q be the target flow rate value for individual j in the i-th pump station dimension; i,k Let be the target flow rate value for individual k in the i-th pump station dimension; the formula for the intention coupling factor is: In the formula ε is the coupling factor between the intentions of individual j and individual k; ε is a very small positive number (a small constant to prevent division by zero); w i ·Q i,j Will-weighted traffic; w i (t) Step 3) Based on step 2), update the position. The update formula set is as follows: In the formula Let be the position vector of the k-th individual in the (t+1)-th generation (after update); Let be the position vector of the k-th individual in the t-th generation (before the update); Let β be the position vector of individual j that is brighter than k in generation t (as the target scheduling scheme of the attraction source); j,k The strength of attraction between individual j and individual k (obtained in step 2); This is the position difference vector (used to describe the distance from the current solution to the optimal solution); α (t) α is the randomization step size factor for generation t (used to control the magnitude of the random perturbation term);(t) ·ε k The random perturbation component is used to avoid prematurely falling into local optima; the scheduling flow allocation solution for the next generation of individuals is finally obtained through the execution steps.

[0043] It performs fitness-based retention and convergence determination on the scheduling traffic allocation solution, and extracts the globally optimal solution, including: Step 1) Sort all individuals in the current population in ascending order of fitness value to obtain an ordered list of individuals. Take the top K individuals after sorting as elite individuals (i.e., 10% of the population size) and directly copy the marked K elite individuals into the next generation of the population. Step 2) Set the maximum number of iterations to 50, and stop iterating when the maximum number of iterations is reached (to avoid failing to converge and getting stuck in an infinite loop). Step 3) Calculate the relative rate of change of the optimal fitness over the last L generations, using the following formula: In the formula, δ is the convergence threshold (typically 10). -4 ); is the fitness value of the best individual in the population in generation t (i.e., the current generation) (reflecting the quality of the optimal scheduling scheme found by the current algorithm). The fitness value of the best individual in the tL-th generation population (as a benchmark). This represents the absolute change in optimal fitness over L generations. Step 4) After the iterations from Step 1) to Step 3) stop, select the individual with the smallest fitness value in the population as the optimal scheduling scheme (i.e., the collaborative scheduling optimization scheme of each pump station).

[0044] In specific implementation, S4 includes: S401. Perform integrity verification, supply and demand balance deviation calculation, and secondary confirmation of health hard constraints on the collaborative scheduling optimization scheme to obtain an executable scheduling scheme that meets the feasibility prediction. S402. Using the digital twin model of the pump station group, hydraulic transient simulation and dynamic speed calculation are performed on the executable scheduling scheme to evaluate the stability of the transient process and obtain the target speed and stability verification results of each pump station. S403. By using power verification, vibration and temperature trend prediction, and start-stop cycle constraint checks, the target speed and stability verification results of each pump station are filtered by safety boundaries to obtain the final executable scheduling scheme that passes all safety constraints.

[0045] In its specific implementation, S4 also includes: S404. Convert the target flow rate in the final executable scheduling scheme into target speed and valve opening, and generate control command data packets for each pump station. S405. Send control command data packets to each pump station to execute operations, and build a log database to store the operation evidence.

[0046] Its integrity verification, supply and demand balance deviation calculation, and secondary confirmation of health hard constraints include: Step 1) Check that the dimensions of the collaborative scheduling optimization scheme meet the following conditions: 1- Ensure that it is consistent with the scale N of the pump station group; 2- Each component meets the numerical reasonableness (non-negative, non-infinite and non-NaN); If the conditions are not met, the verification fails and the scheme recalculation request is triggered to roll back to S3. Step 2) Obtain the total water supply of the coordinated scheduling optimization plan, compare it with the target demand of the current scheduling task, and calculate the supply-demand deviation rate. The formula for calculating the deviation rate is: In the formula δ Q This refers to the supply-demand deviation rate. Q represents the total water supply for the coordinated scheduling optimization scheme. i实际 The target demand for the current scheduling task; 100% represents the conversion of supply and demand deviation into deviation rate; the preset value is when δ... Q When the tolerance threshold is 5%, the solution is determined to not meet the basic requirements of the task and an early warning is issued. Step 3) Read the current health index of each pump station i from the digital twin model of the S1 pump station group. When the health index HI i <Typical value 0.5 and Q i >ε (ε=10) -3 m 3 / s), then force Q to i Set the value to 0 and record that the pump station has exited the current scheduling task. The corresponding shortfall will be redistributed by the system in subsequent fine-tuning.

[0047] Its hydraulic transient simulation and rotational speed dynamic calculation, and evaluation of transient process stability include: Step 1) Take the actual operating status of each pumping station at the current moment as the starting point of the simulation, and set the time step size to 1-5 seconds; Step 2) Calculate the target flow rate Q of each pumping station in the collaborative scheduling optimization scheme. i This is converted into a speed regulation command to simulate the dynamic process of each pumping station transitioning from its current operating condition to the target operating condition. The approximate formula for the relationship between speed and flow rate is as follows: In the formula Q i (t) represents the instantaneous flow rate of the i-th pump station at simulation time t; n i (t) represents the instantaneous rotational speed of the i-th pump station at simulation time t; Q i and n i Let $i$ be the target flow rate of the $i$-th pumping station and $j$ be the target rotational speed required for the $i$-th pumping station to achieve the target flow rate, respectively. Step 3), based on step 2), where n iThe solution is obtained by simultaneously solving the flow-head mathematical relationship curve and the pipeline characteristic curve in the hydraulic characteristic twin model of the pump station group digital twin model. The solution formula is as follows: In the formula n i The target rotational speed required for the i-th pump station to achieve the target flow rate; H 静,i S is the net head of the i-th pump station; i Let n be the resistance coefficient of the pipeline corresponding to the i-th pump station; 额定,i Let be the rated speed of the i-th pump station; For dynamic head loss; Step 4) Based on step 3), update the rotational speed using the following formula: In the formula, is The rotational speed at the next simulation moment (the updated rotational speed value, used to advance the simulation clock); n i (t) represents the rotational speed at the current simulation moment; Δt 下 T represents the simulation time step (taken as 1-5); 斜,i Let n be the ramp time constant for the speed regulation of the i-th pump station; (n i -n i (t) represents the deviation between the current speed and the target speed (i.e., the remaining amount of speed adjustment needed). Step 5) Assess the stability of the transient process. The formula is: In the formula S 稳 ΔP is a comprehensive index of the system's hydraulic stability. max ΔP represents the maximum peak value of the total pressure fluctuation in the pipeline network during the simulation period. 压 The maximum allowable pressure fluctuation limit for the pipeline network; max i σ Q,i The value representing the largest flow overshoot among all pumping stations (using the case of the most severe pumping station to represent the overall flow overshoot level of the system); σ Q,上限 The maximum allowable flow overshoot limit; |dL / dt| max The maximum absolute value of the water level change rate during the simulation period (taking the moment of the most intense water level fluctuation during the transition process); (dL / dt) 上限 The upper limit of the allowable rate of water level change (determined by the stability of the channel / pond slope); , and These are pressure fluctuation index, flow overshoot index, and water level change rate index.

[0048] The safety boundary filtering of the target speed and stability verification results for each pumping station includes: Step 1) Predict the operating power P of each pump station at the target flow rate using the flow-power mathematical relationship curve in the hydraulic characteristic twin model of the pump station group digital twin model. This prediction is made when P ≦ Y·P 额定,i When the rated power of the pumping station is reached, the safety check is passed. If the check is not met, the upper limit of the pumping station's flow rate is lowered and the local optimization is retried. Step 2) When the pump station's flow rate Q is in this collaborative scheduling optimization scheme i ≈0 (less than the start / stop threshold Q) stop =0.20 × the rated flow of the i-th pump station), and if it is currently in operation, then a shutdown operation is planned; if the maximum number of start-stop operations is exceeded, then the pump station will be kept at the lowest speed (not completely shut down) to avoid equipment damage, thus completing the safety boundary filtering and obtaining the final executable scheduling scheme.

[0049] The conversion between the target rotational speed and the valve opening includes: Step 1) For pump stations that need to be operated, directly extract the verified target rotational speed value from the hydraulic transient simulation results; Step 2), the formula for converting the valve opening degree based on Step 1) is: In the formula V i Q represents the target valve opening. i Let ΔP be the target flow rate of the i-th pumping station; 压,i The pressure difference across the valve; ρ is the fluid density; It is an inverse function of the valve flow coefficient.

[0050] See appendix Figure 2 The present invention also provides a collaborative scheduling system for pump station groups based on digital twins, comprising: The first module is used to collect multi-source heterogeneous data from the pump station group, build the basic framework for constructing a digital twin model of the pump station group, and update it synchronously. The second module is used to integrate health status, energy efficiency ratio, load capacity and response speed data to calculate the scheduling response willingness factor of each pumping station. The third module is used to embed the dynamic scheduling capability weight vector into the balanced fitness function and response weighted search mechanism for solving; The fourth module is used to pre-test and verify the coordinated scheduling optimization scheme of each pumping station. After confirming hydraulic balance and operational safety, it generates target speed and start / stop commands and sends them to each pumping station for execution.

[0051] Those skilled in the art will recognize that the implementation steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

Claims

1. A collaborative scheduling method for pump station groups based on digital twins, characterized in that, include: S1. Collect multi-source heterogeneous data of the pump station group, construct the basic framework of the digital twin model of the pump station group and update it synchronously; S2. Based on the digital twin model of the pump station group, a nonlinear mapping algorithm is used to fuse health, energy efficiency ratio, load capacity and response speed data to calculate the scheduling response willingness factor of each pump station and obtain the dynamic scheduling capability weight vector. S3. Using the improved firefly algorithm, the dynamic scheduling capability weight vector is embedded into the balanced fitness function and response weighted search mechanism for solution, so as to obtain the collaborative scheduling optimization scheme of each pumping station. S4. Using the digital twin model of the pump station group, the coordinated scheduling optimization scheme of each pump station is pre-tested and verified. After confirming the hydraulic balance and operational safety, the target speed and start / stop commands are generated and sent to each pump station for execution.

2. The method for collaborative scheduling of pump station groups based on digital twins according to claim 1, characterized in that, Its S1 includes: S101. Collect multi-source heterogeneous data from the pump station group, including hydraulic data, equipment status data, energy consumption data, and environmental data, as the raw data stream; S102. Clean and preprocess the raw data stream to obtain a structured real-time data vector; S103. Construct a hydraulic characteristic twin model of the pump station using a multinomial fitting method, construct an equipment health twin model using a proportional risk model based on Weibull distribution, construct an energy consumption twin model using a real-time energy efficiency ratio mapping model, and complete the construction of a digital twin model of the pump station group. S104. Based on the structured real-time data vector, the recursive least squares method is used to correct the hydraulic characteristic curve parameters online and calculate the health index, energy efficiency ratio, load capacity and response speed to obtain the real-time state vector.

3. The method for collaborative scheduling of pump station groups based on digital twins according to claim 2, characterized in that, Its S2 includes: S201. Receive the real-time status vector and extract the health index, energy efficiency ratio, load capacity and response speed components from it to obtain a four-dimensional independent vector. S202. Calculate the willingness contribution value of the health index, energy efficiency ratio, load capacity, and response speed components in the four-dimensional independent vector; the formula for calculating the willingness contribution value of the health index is: In the formula F H (HI i ) contributes to health intention; HI i The equipment health index is represented by k1, the kurtosis coefficient by θ1, and the health threshold by e. The energy efficiency ratio calculation formula is as follows: In the formula F E (EER i Energy Efficiency Contribution (EER) is the contribution value of the willingness to improve energy efficiency. i Real-time energy efficiency ratio; EER 优 For optimal energy efficiency reference value, 80% of the pump station's rated efficiency is taken; EER min The minimum acceptable energy efficiency ratio is 50% of the pump station's rated efficiency. This is the normalized value of the energy efficiency compliance rate; the formula for calculating the intended contribution value of load capacity is: In the formula F L (LC i ) represents the load intention contribution value; -k2 is the attenuation coefficient; e is the natural constant; LC i The remaining load capacity; the formula set for calculating the willingness contribution value for response speed is as follows: In the formula F R (RS i ) represents the contribution value of the willingness to respond; RS i For response speed index; RS max The maximum effective response speed exponential threshold is set to a typical value of 1.

5.

4. The method for collaborative scheduling of pump station groups based on digital twins according to claim 3, characterized in that, Its S2 also includes: S203. The health willingness contribution value, energy efficiency willingness contribution value, load willingness contribution value and response willingness contribution value are weighted and combined to form the dispatch response willingness factor, with the weight allocation being 0.3→0.25→0.25→0.2 respectively; S204. Normalize the dynamic scheduling capability weight vector of each pump station to generate a dynamic scheduling capability weight vector for improving the response weighted initialization of the firefly algorithm.

5. A collaborative scheduling method for pump station groups based on digital twins according to claim 4, characterized in that, Its S3 includes: S301. Construct a mathematical model with the goal of minimizing total energy consumption and embedding weight constraint penalties using optimization modeling methods, and initialize the control parameters of the improved firefly algorithm. S302. After initialization, obtain the dynamic scheduling capability weight vector, and use the weight-guided scaling factor to perform non-uniform amplification on the uniform random number to obtain the initial population that preferentially gathers in the high flow area of ​​the high-willing pump station. S303. Using the dynamic scheduling capability weight vector, construct a balanced fitness function carrying the intention reverse penalty coefficient, evaluate the fitness of each individual in the initial population, and obtain a comprehensive fitness value that reflects energy consumption, task completion degree and constraint penalty.

6. The method for collaborative scheduling of pump station groups based on digital twins according to claim 5, characterized in that, Its S3 also includes: S304. Using the intention-coupled attraction method, the brightness of the comprehensive fitness value, which reflects energy consumption, task completion degree and constraint penalty, is compared and the position is updated to obtain the scheduling flow allocation solution for the new generation of individuals in the population. S305. Using an elite retention strategy, the fitness of the scheduling flow allocation solution is selected for optimal retention and convergence determination. Finally, the global optimal solution is extracted as the collaborative scheduling optimization scheme for each pumping station.

7. A collaborative scheduling method for pump station groups based on digital twins according to claim 6, characterized in that, Its S4 includes: S401. Perform integrity verification, supply and demand balance deviation calculation, and secondary confirmation of health hard constraints on the collaborative scheduling optimization scheme to obtain an executable scheduling scheme that meets the feasibility prediction. S402. Using the digital twin model of the pump station group, hydraulic transient simulation and dynamic speed calculation are performed on the executable scheduling scheme to evaluate the stability of the transient process and obtain the target speed and stability verification results of each pump station. S403. By using power verification, vibration and temperature trend prediction, and start-stop cycle constraint checks, the target speed and stability verification results of each pump station are filtered by safety boundaries to obtain the final executable scheduling scheme that passes all safety constraints.

8. A collaborative scheduling method for pump station groups based on digital twins according to claim 7, characterized in that, Its S4 also includes: S404. Convert the target flow rate in the final executable scheduling scheme into target speed and valve opening, and generate control command data packets for each pump station. S405. Send the control command data packet to each pump station to execute the operation, and build a log database to store the operation evidence.

9. A collaborative scheduling system for pump station groups based on digital twins, characterized in that, include: The first module is used to collect multi-source heterogeneous data from the pump station group, build the basic framework for constructing a digital twin model of the pump station group, and update it synchronously. The second module is used to integrate health status, energy efficiency ratio, load capacity and response speed data to calculate the scheduling response willingness factor of each pumping station. The third module is used to embed the dynamic scheduling capability weight vector into the balanced fitness function and response weighted search mechanism for solving; The fourth module is used to pre-test and verify the coordinated scheduling optimization scheme of each pumping station. After confirming hydraulic balance and operational safety, it generates target speed and start / stop commands and sends them to each pumping station for execution.

Images