An intelligent scheduling method, system, device and medium for coordinated operation of a variable-speed multi-pump parallel system

Abstract

The present invention relates to the technical field of optimized control of centrifugal pumps, and specifically, to an intelligent scheduling method, system, device and medium for coordinated operation of a variable-speed multi-pump parallel system; the method first calls numerical methods to simulate and calculate the performance characteristics of parallel centrifugal pumps and the pipeline loss characteristics of the variable-speed multi-pump parallel system to obtain digital characteristic curves; secondly, with the minimum total shaft power as the objective function, an intelligent regulation mathematical model of the variable-speed multi-pump parallel system is established under the given demand targets and various constraint conditions of the variable-speed multi-pump parallel system; then, the particle swarm optimization algorithm is improved according to the golden sine algorithm to solve and obtain operation data; and a surrogate model is established by calling a feedforward artificial neural network to regress and predict the operation data; finally, according to the prediction results of the surrogate model, an optimal regulation scheme for the variable-speed multi-pump parallel system is obtained, which improves the energy efficiency and stability of the variable-speed multi-pump parallel system while achieving more accurate prediction of the scheduling results.

Description

Technical Field

[0001] The present invention relates to the technical field of optimized control of centrifugal pumps, and more specifically, to an intelligent scheduling method, system, device and medium for coordinated operation of a variable-speed multi-pump parallel system. Background Art

[0002] Currently, there are mainly two methods for regulating a variable-speed multi-pump parallel system in the fluid transportation link:

[0003] 1. Manual adjustment based on historical data and manual experience: Manual adjustment generally includes speed adjustment method and valve adjustment method. During the operation of the equipment, according to the demand for the pumped fluid, the rotation speed or the valve opening is adjusted to change the flow distribution of the multi-pump parallel system to respond to different task requirements;

[0004] 2. Automatic control scheduling using intelligent algorithms: The automatic control scheduling method using intelligent algorithms is to use a computer to collect and learn historical data to provide a control decision-making scheme for the operation of the multi-pump parallel system.

[0005] Regarding the existing technical means and control methods, the specific description is as follows:

[0006] (1) Speed adjustment method: The speed adjustment method can be divided into variable-speed adjustment method and throttling adjustment method. Among them, the variable-speed adjustment method realizes the adjustment of the pump operating conditions by changing the rotation speed of the pump prime mover. This method benefits from the progress of frequency conversion technology, making the application of variable-speed adjustment in pump condition adjustment more extensive. During variable-speed adjustment, the operator determines the rotation speed of the centrifugal pump and the demand changes of the water supply pressure and flow rate, so that the pump can maintain a relatively high operating efficiency within a certain range. The throttling adjustment method is relatively simple. The operator manually adjusts the opening of the centrifugal pump discharge valve to change the flow rate. As the valve opening becomes smaller, the flow rate and flow velocity decrease accordingly. However, this adjustment method will cause relatively large throttling losses and poor economy. If the valve is closed too much and the working flow rate is very small, it will also cause safety hazards such as pump body heating and cavitation damage;

[0007] (2) Valve adjustment method: The valve adjustment method mainly manually adjusts the opening of the outlet valve to change the output pressure of the pump. This method is simple and economical. By changing the opening of the outlet valve, the output pressure of the pump can be effectively adjusted. This method is suitable for occasions where the water pressure needs to be quickly adjusted. However, due to over-reliance on personal experience during manual operation and the lack of quantitative execution means, this method will cause unstable system operation and insignificant efficiency improvement when applied;

[0008] (3)Automatic control scheduling using intelligent algorithms: In traditional algorithm control, the data processing of computers and intelligent algorithms are mainly used to optimize and adjust the operating parameters of the control system. By optimizing the operating parameters based on historical data, a control scheme for improving system efficiency and stability is obtained. Currently, this method usually includes steps such as data collection, data processing, establishing optimization objectives and optimization models, and solving the model through intelligent algorithms to find the optimal operating parameters and achieve the best performance of the system. However, it is difficult to adapt to complex systems with multiple pumps in parallel relying solely on the historical data of system operation. The existing intelligent algorithms have poor adaptability in the pumping transportation link, and the accuracy of the provided optimization and regulation decisions is relatively low.

[0009] Manual adjustment methods rely on the experience judgment of operation and maintenance personnel. During the execution process, there is a lack of quantitative analysis methods and they are greatly affected by changes in the external environment. It is difficult to make accurate judgments according to the characteristics of different types of equipment units when referring to historical data in the process of making distribution decisions through manual adjustment, resulting in insufficient energy conservation of the equipment and low efficiency improvement.

[0010] The existing scheduling control technologies relying on intelligent algorithms often establish constraint conditions through the collection and analysis of historical data, and adopt different optimization strategies to achieve the adjustment of low-energy consumption and high-efficiency working configurations. However, the currently common intelligent optimization methods lack algorithm designs for fluid machinery such as centrifugal pumps. They only construct function models based on historical data while ignoring the research on the parallel operation characteristics of centrifugal pumps. The applicability of the optimization methods is poor, and the accuracy of result prediction and analysis is relatively low.

[0011] In addition, in the optimization algorithms in existing engineering applications, the genetic algorithm has strong global search ability, good parallelism, and strong adaptability, but it has the deficiencies of high computational complexity, complex programming implementation, and poor late-stage search ability; the artificial bee colony algorithm has the advantages of strong global optimization ability, wide application range, and strong robustness, but it has the problems of being easily trapped in local optima and being only applicable to the global optimization of continuous functions. It makes little contribution to the scheduling optimization of the overall system in engineering practice, and the response time of the optimization strategy is too long. Summary of the Invention

[0012] In view of the problems existing in the existing operation optimization methods for multi-pump parallel systems, such as difficult quantitative decision-making, inaccurate performance prediction, poor adaptability of existing optimization algorithms, and slow response speed, the present invention proposes an intelligent scheduling method, system, device, and medium for coordinated operation of a variable-speed multi-pump parallel system; this method first calls numerical methods to simulate and calculate the performance characteristics of parallel centrifugal pumps and the pipeline loss characteristics of a variable-speed multi-pump parallel system to obtain digital characteristic curves; secondly, with the minimum total shaft power as the objective function, under the given demand targets of the variable-speed multi-pump parallel system, namely the demand flow Q a and the demand head Ha and under various constraint conditions, establish an intelligent regulation mathematical model for a variable-speed multi-pump parallel system; then improve the particle swarm optimization algorithm according to the golden sine algorithm to solve and obtain operation data; and call a feedforward artificial neural network to establish a surrogate model to regress and predict the operation data; finally, according to the prediction results of the surrogate model, obtain the optimal regulation scheme for the variable-speed multi-pump parallel system, while achieving a more accurate prediction of the scheduling results, improving the energy efficiency and stability of the variable-speed multi-pump parallel system.

[0013] The specific implementation content of the present invention is as follows:

[0014] An intelligent scheduling method for the coordinated operation of a variable-speed multi-pump parallel system. First, call a numerical method to simulate and calculate the variable-speed performance characteristics of parallel centrifugal pumps and the pipeline loss characteristics of the variable-speed multi-pump parallel system to obtain digital characteristic curves; secondly, according to the digital characteristic curves, with the minimum total shaft power as the objective function, under the given demand targets of the variable-speed multi-pump parallel system, namely the demand flow Q a and the demand head H a , and under various constraint conditions, establish an intelligent regulation mathematical model for the variable-speed multi-pump parallel system; then improve the particle swarm optimization algorithm according to the golden sine algorithm to solve the intelligent regulation mathematical model of the variable-speed multi-pump parallel system to obtain operation data; and call a feedforward artificial neural network to establish a surrogate model to regress and predict the operation data; finally, according to the prediction results of the surrogate model, obtain the optimal regulation scheme for the variable-speed multi-pump parallel system.

[0015] To better implement the present invention, further, the intelligent scheduling method for the coordinated operation of the variable-speed multi-pump parallel system specifically includes the following steps:

[0016] Step S1: Call a numerical method to simulate and calculate the variable-speed performance characteristics of parallel centrifugal pumps and the pipeline loss characteristics of the variable-speed multi-pump parallel system to obtain digital characteristic curves;

[0017] Step S2: According to the digital characteristic curves, with the minimum total shaft power as the objective function, under the given demand targets of the variable-speed multi-pump parallel system, namely the demand flow Q a and the demand head H a , and under various constraint conditions, establish an intelligent regulation mathematical model for the variable-speed multi-pump parallel system;

[0018] Step S3: According to the golden sine algorithm, adaptive strategy, and global optimal-guided dynamic reverse learning to improve the particle swarm optimization algorithm, solve the intelligent regulation mathematical model of the variable-speed multi-pump parallel system to obtain operation data;

[0019] Step S4: Call the feedforward artificial neural network to establish a surrogate model, regressively predict the operating data, and obtain the prediction result;

[0020] Step S5: According to the prediction result of the surrogate model, combine with the scheduling decision variables to obtain the optimal coordinated operation scheduling scheme of the variable-speed multi-pump parallel system.

[0021] To better implement the present invention, further, the specific steps of step S1 are as follows:

[0022] Step S11: According to the factory characteristic curve of a single centrifugal pump, or by establishing a three-dimensional model of a single centrifugal pump, calculate the characteristic curve of the single centrifugal pump by using the numerical simulation method based on performance prediction. Fit the characteristic curve of the single centrifugal pump by the polynomial fitting method to obtain the mathematical expressions of the single-pump flow Q -head H 、single-pump flow Q -shaft power P and single-pump flow Q -efficiency η characteristic curves;

[0023] Step S12: According to the similarity theory of centrifugal pumps, calculate the characteristic curve of a single centrifugal pump under variable speed conditions, and establish the mathematical expressions of the single-pump flow Q -head H 、single-pump flow Q -shaft power P and single-pump flow Q -efficiency η characteristic curves under variable speed conditions of a single centrifugal pump;

[0024] Step S13: According to the characteristics of flow superposition and constant head of parallel centrifugal pumps, calculate the characteristic curve of the variable-speed multi-pump parallel system, and establish the mathematical expressions of the total flow Q t -total shaft power P t and total flow Q t -parallel pump head H t characteristic curves of the variable-speed multi-pump parallel system;

[0025] Step S14: Construct the pipeline loss characteristic curve according to the net head of the pipeline, the pipe resistance coefficient of the branch pipe before the parallel node, the pipe resistance coefficient of the main pipeline of the variable-speed multi-pump parallel system, the flow rate of the branch pipeline, and the total flow rate of the parallel centrifugal pumps.

[0026] To better implement the present invention, further, the specific steps of step S2 are as follows:

[0027] Step S21: Construct an objective function based on the characteristic curve of a single centrifugal pump, the characteristic curve under variable speed conditions of a single centrifugal pump, the characteristic curve of a variable speed multi-pump parallel system, the pipeline loss characteristic curve, the total shaft power of the variable speed multi-pump parallel system, the shaft power of each centrifugal pump, the state factor of the centrifugal pump, the centrifugal pump speed ratio, the number of operating parallel centrifugal pumps, and the single-pump flow rate.

[0028] Step S22: Construct constraint conditions; the construction of the constraint conditions includes: constructing a head constraint for the variable speed multi-pump parallel system according to the pipeline loss characteristic curve, the head characteristic curve of the parallel pumps, and the required head, constructing a start-stop number constraint according to the maximum number of operating units, constructing a flow rate constraint for the variable speed multi-pump parallel system according to the water supply demand flow rate, constructing a speed ratio constraint according to the rated speed of the centrifugal pump, and constructing a centrifugal pump operating condition constraint according to the rated flow rate of the centrifugal pump and the set constraint interval.

[0029] Step S23: Establish an intelligent scheduling mathematical model for the variable speed multi-pump parallel system based on the objective function and the constraint conditions.

[0030] To better implement the present invention, further, step S3 specifically includes the following steps:

[0031] Step S31: Construct a position update strategy according to the golden sine algorithm; the position update strategy is used to update the position of the particle.

[0032] Step S32: Construct an adaptive strategy according to the set weight factor and learning factor; the adaptive strategy is used to adaptively update the weight factor and the learning factor.

[0033] Step S33: Construct a globally optimal-guided dynamic reverse learning strategy according to the reverse solution of the value of the u th individual in the current population on the v dimension; the globally optimal-guided dynamic reverse learning strategy is used to complete dynamic reverse learning according to the globally optimal-guided dynamic reverse learning strategy after the particle swarm algorithm iteration is completed, and screen out the top N individuals with the best fitness, and merge the reverse population with the initial population to form a new population.

[0034] Step S34: Improve the particle swarm optimization algorithm according to the position update strategy, the adaptive strategy, and the globally optimal-guided dynamic reverse learning strategy, and solve the intelligent scheduling mathematical model of the variable speed parallel pump group according to the improved particle swarm optimization algorithm to obtain operation data.

[0035] To better implement the present invention, further, step S4 specifically includes the following steps:

[0036] Step S41: Call a feedforward neural network with a single hidden layer to construct a surrogate model.

[0037] Step S42: Adjust the regression coefficients of the proxy model, use the operation data as the input layer, and train through a single hidden layer to obtain the prediction result.

[0038] To better implement the present invention, further, the objective function established in step S21 is:

[0039]

[0040] Wherein, P t is the total shaft power of the variable-speed multi-pump parallel system; P i is the shaft power of the i th centrifugal pump; W i is the state factor of the i th centrifugal pump; b 1, b 2, b 3, b 4 are the fitting coefficients of the polynomial fitting of the single-pump flow-power curve of the centrifugal pump; k is the speed ratio of the centrifugal pump; m o is the number of operating centrifugal pumps in the variable-speed multi-pump parallel system; Q is the flow rate of a single pump.

[0041] Based on the above-mentioned intelligent scheduling method for the coordinated operation of the variable-speed multi-pump parallel system, to better implement the present invention, further, an intelligent scheduling system for the coordinated operation of the variable-speed multi-pump parallel system is proposed, which is used to execute the above-mentioned intelligent scheduling method for the coordinated operation of the variable-speed multi-pump parallel system; it includes a simulation unit, a construction unit, a solution unit, a prediction unit, and an output unit;

[0042] The simulation unit is used to call numerical methods to simulate and calculate the variable-speed performance characteristics of the parallel centrifugal pumps and the pipeline loss characteristics of the variable-speed multi-pump parallel system to obtain the digital characteristic curves;

[0043] The construction unit is used to, according to the digital characteristic curves, with the minimum total shaft power as the objective function, under the given demand targets of the variable-speed multi-pump parallel system, namely the demand flow Q a and the demand head H a , and various constraint conditions, establish an intelligent regulation mathematical model for the variable-speed multi-pump parallel system;

[0044] The solution unit is used to solve the intelligent regulation mathematical model of the variable-speed multi-pump parallel system according to the improved particle swarm optimization algorithm based on the golden sine algorithm to obtain the operation data;

[0045] The prediction unit is used to call a feedforward artificial neural network to establish a surrogate model and perform regression prediction on the operation data;

[0046] The output unit is used to obtain the optimal regulation scheme of the variable-speed multi-pump parallel system according to the prediction result of the surrogate model.

[0047] Based on the above-mentioned intelligent scheduling method for the coordinated operation of a variable-speed multi-pump parallel system, in order to better implement the present invention, further, an electronic device is proposed, including a memory and a processor; a computer program is stored on the memory; when the computer program is executed on the processor, the above-mentioned intelligent scheduling method for the coordinated operation of a variable-speed multi-pump parallel system is implemented.

[0048] Based on the above-mentioned intelligent scheduling method for the coordinated operation of a variable-speed multi-pump parallel system, in order to better implement the present invention, further, a computer-readable storage medium is proposed, and a computer instruction is stored on the computer-readable storage medium; when the computer instruction is executed on the above-mentioned electronic device, the above-mentioned intelligent scheduling method for the coordinated operation of a variable-speed multi-pump parallel system is implemented.

[0049] The present invention has the following beneficial effects:

[0050] (1) The method for simulating and fitting the performance characteristic curve of a variable-speed centrifugal pump proposed by the present invention accurately constructs the performance characteristic curve of a variable-speed multi-pump parallel system. Compared with the method of constructing a function model only based on historical data in the prior art, the applicability of the optimization method is improved, and the accuracy of result prediction and analysis is improved;

[0051] (2) The present invention analyzes the constraint conditions for the operation of a variable-speed multi-pump parallel system in a specific high-efficiency range, and establishes an optimization mathematical model for the coordinated operation of a variable-speed multi-pump parallel system, solving the problem of mismatch between the operating conditions and requirements of a variable-speed multi-pump parallel system, namely the required flow rate Q a and the required head H a ;

[0052] (3) The present invention takes the minimum total shaft power as the objective function and, under various constraint conditions, establishes an accurate optimization model for finding the best scheduling scheme for a variable-speed multi-pump parallel system under different flow rate requirements;

[0053] (4) The improved particle swarm optimization algorithm GPSO based on golden sine proposed by the present invention is used to calculate the objective function of a variable-speed multi-pump parallel system under multiple constraint conditions, solving the problems of long calculation time, low calculation accuracy, and slow convergence speed of traditional algorithms in the pump station optimization scheduling problem;

[0054] (5) The present invention integrates the parameters obtained by algorithm solving to form a database, constructs an artificial neural network for predicting the allocation result, and quickly generates an optimal operation scheduling scheme for the variable-speed multi-pump parallel system according to real-time input data, solving the problem of a large amount of electric energy waste caused by the traditional variable-speed multi-pump parallel system regulation depending on the experience of the staff. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a flowchart of the coordinated operation optimization method for the variable-speed multi-pump parallel system provided by the present invention.

[0056] Figure 2 It is a flowchart for solving the mathematical model.

[0057] Figure 3 It is a schematic structural diagram of a single-hidden-layer feedforward artificial neural network.

[0058] Figure 4 It is a schematic diagram of the principle of the surrogate model.

[0059] Figure 5 It is a bar chart of the search speed of the partially improved particle swarm optimization algorithm GPSO and other optimization algorithms.

[0060] Figure 6 It is an iterative convergence curve of the partially improved particle swarm optimization algorithm GPSO and other optimization algorithms.

[0061] Figure 7 It is a box plot of the statistical indicators of the partially improved particle swarm optimization algorithm GPSO and other optimization algorithms.

[0062] Figure 8 It is a schematic structural diagram of the single-hidden-layer feedforward artificial neural network in Embodiment 3.

[0063] Figure 9 It is a regression prediction curve graph of the surrogate model.

[0064] Figure 10 It is a curve graph of the highest water supply demand and the required net head in a region in Embodiment 3. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0065] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. It should be understood that the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments, and therefore should not be regarded as a limitation of the protection scope. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0066] In the description of the present invention, it should be noted that unless otherwise clearly specified and limited, the terms "set", "connected", and "connected to" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can also be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0067] Embodiment 1:

[0068] This embodiment proposes to first call a numerical method to simulate and calculate the variable-speed performance characteristics of a parallel centrifugal pump and the pipeline loss characteristics of a variable-speed multi-pump parallel system to obtain digital characteristic curves; secondly, according to the digital characteristic curves, with the minimum total shaft power as the objective function, under the given demand target of the variable-speed multi-pump parallel system, that is, the required flow Q a and the required head H a , and under various constraint conditions, establish an intelligent regulation mathematical model for the variable-speed multi-pump parallel system; then improve the particle swarm optimization algorithm according to the golden sine algorithm, solve the intelligent regulation mathematical model of the variable-speed multi-pump parallel system to obtain operation data; and call a feedforward artificial neural network to establish a surrogate model to regress and predict the operation data; finally, according to the prediction results of the surrogate model, obtain the optimal regulation scheme for the variable-speed multi-pump parallel system; specifically including the following steps:

[0069] Step S1: Call a numerical method to simulate and calculate the variable-speed performance characteristics of a parallel centrifugal pump and the pipeline loss characteristics of a variable-speed multi-pump parallel system to obtain digital characteristic curves.

[0070] The specific steps of step S1 are as follows:

[0071] Step S11: According to the factory characteristic curve of a single centrifugal pump or by establishing a three-dimensional model of a single centrifugal pump, use a numerical simulation method based on performance prediction to calculate the characteristic curve of a single centrifugal pump. Fit the characteristic curve of a single centrifugal pump by the polynomial fitting method to obtain the single-pump flow Q - head H , single-pump flow Q - shaft power P and the mathematical expressions of the single-pump flow Q - efficiency η characteristic curves;

[0072] Step S12: According to the similarity theory of centrifugal pumps, calculate the characteristic curve of a single centrifugal pump under variable-speed conditions, and establish the single-pump flow Q - head of a single centrifugal pump under variable-speed conditions H , single-pump flow Q - shaft powerP and the single pump flow rate Q - Efficiency η The mathematical expression of the characteristic curve;

[0073] Step S13: According to the characteristics of flow rate superposition and constant head of parallel centrifugal pumps, calculate the characteristic curve of the variable-speed multi-pump parallel system, and establish the total flow rate of the variable-speed multi-pump parallel system Q t - Total shaft power P t and the total flow rate Q t - Head of parallel pumps H t The mathematical expression of the characteristic curve;

[0074] Step S14: According to the net head of the pipeline, the pipe resistance coefficient of the branch pipe before the parallel node, the pipe resistance coefficient of the main pipeline of the variable-speed multi-pump parallel system, the flow rate of the branch pipeline, and the total flow rate of the parallel centrifugal pumps, construct the pipeline loss characteristic curve.

[0075] Step S2: Based on the digital characteristic curve, with the minimum total shaft power as the objective function, under the given demand targets of the variable-speed multi-pump parallel system, namely the required flow rate Q a and the required head H a , and under various constraint conditions, establish an intelligent regulation mathematical model for the variable-speed multi-pump system.

[0076] The specific steps of Step S2 include the following steps:

[0077] Step S21: According to the characteristic curve of a single centrifugal pump, the characteristic curve of a single centrifugal pump under variable-speed conditions, the characteristic curve of the variable-speed multi-pump parallel system, the pipeline loss characteristic curve, the total shaft power of the variable-speed multi-pump parallel system, the shaft power of each centrifugal pump, the state factor of the centrifugal pump, the centrifugal pump speed ratio, the number of operating parallel centrifugal pumps, and the single pump flow rate, construct the objective function;

[0078] Step S22: Construct the constraint conditions; the construction of the constraint conditions includes: constructing the head constraint of the variable-speed multi-pump parallel system according to the pipeline loss characteristic curve, the head characteristic curve of the parallel pumps, and the required head, constructing the start-stop number constraint according to the maximum number of operating units, constructing the flow rate constraint of the variable-speed multi-pump parallel system according to the water supply demand flow rate, constructing the speed ratio constraint according to the rated speed of the centrifugal pump, and constructing the operating condition constraint of the centrifugal pump according to the rated flow rate of the centrifugal pump and the set constraint interval;

[0079] Step S23: According to the objective function and the constraint conditions, establish an intelligent scheduling mathematical model for the variable-speed multi-pump parallel system.

[0080] Step S3: Improve the particle swarm optimization algorithm according to the golden sine algorithm, adaptive strategy, and global-optimal-guided dynamic reverse learning to solve the intelligent regulation mathematical model of the variable-speed multi-pump parallel system and obtain the operation data.

[0081] The specific steps of step S3 are as follows:

[0082] Step S31: Construct a position update strategy according to the golden sine algorithm; the position update strategy is used to update the position of the particle;

[0083] Step S32: Construct an adaptive strategy according to the set weight factor and learning factor; the adaptive strategy is used to adaptively update the weight factor and learning factor;

[0084] Step S33: According to the reverse solution of the value of the u th individual in the current population on the v th dimension, construct a global-optimal-guided dynamic reverse learning strategy; the global-optimal-guided dynamic reverse learning strategy is used to complete dynamic reverse learning according to the global-optimal-guided dynamic reverse learning strategy after the particle swarm algorithm iteration is completed, and screen out the top N individuals with the best fitness, and merge the reverse population with the initial population to form a new population;

[0085] Step S34: Improve the particle swarm optimization algorithm according to the position update strategy, position update strategy, and global-optimal-guided dynamic reverse learning strategy, and solve the intelligent scheduling mathematical model of the variable-speed parallel pump group according to the improved particle swarm optimization algorithm to obtain the operation data.

[0086] Step S4: Call a feedforward artificial neural network to establish a surrogate model, regress and predict the operation data, and obtain the prediction result.

[0087] The specific steps of step S4 are as follows:

[0088] Step S41: Call a feedforward neural network with a single hidden layer to construct a surrogate model;

[0089] Step S42: Adjust the regression coefficient of the surrogate model, use the operation data as the input layer, and train through a single hidden layer to obtain the prediction result.

[0090] The objective function established in step S21 is:

[0091]

[0092] where, P t is the total shaft power of the variable-speed multi-pump parallel system; P i is the shaft power of the i th centrifugal pump; Wi is the state factor of the i th centrifugal pump; b 1, b 2, b 3, b 4 are the fitting coefficients of the polynomial fitting of the single - pump flow - power curve of the centrifugal pump; k is the speed ratio of the centrifugal pump; m o is the number of operating centrifugal pumps in the variable - speed multi - pump parallel system; Q is the flow rate of a single pump.

[0093] Step S5: According to the prediction results of the surrogate model, combined with the scheduling decision variables, obtain the optimal coordinated operation scheduling scheme for the variable - speed multi - pump parallel system.

[0094] Working principle: In this embodiment, first, a numerical method is used to simulate and calculate the variable - speed performance characteristics of the parallel centrifugal pumps and the pipeline loss characteristics of the variable - speed multi - pump parallel system to obtain the digital characteristic curves; secondly, according to the digital characteristic curves, with the minimum total shaft power as the objective function, under the given demand targets of the variable - speed multi - pump parallel system, that is, the demand flow rate Q a and the demand head H a , and under various constraint conditions, establish an intelligent regulation mathematical model for the variable - speed multi - pump parallel system; then improve the particle swarm optimization algorithm according to the golden sine algorithm to solve and obtain the operating data; and call the feed - forward artificial neural network to establish a surrogate model to regress and predict the operating data; finally, according to the prediction results of the surrogate model, obtain the optimal regulation scheme for the variable - speed multi - pump parallel system, while achieving more accurate prediction of the scheduling results, improving the energy efficiency and stability of the variable - speed parallel pump group.

[0095] Embodiment 2:

[0096] Based on the above - mentioned Embodiment 1, as Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 , Figure 8 , Figure 9 , Figure 10 shown, a specific embodiment is used for detailed description.

[0097] Step S1: Use a numerical method to numerically simulate and calculate the variable - speed performance characteristics of the centrifugal pump and the pipeline loss characteristics of the variable - speed multi - pump parallel system to predict the variable - speed performance of the centrifugal pump and determine the single - pump flow rate Q - head H of the single pump, single - pump flow rateQ - Shaft power P and the flow rate of a single pump Q - Efficiency η The mathematical expressions of the characteristic curves.

[0098] The performance of a centrifugal pump is usually measured by a series of core parameters, including rotational speed n , the flow rate of a single pump Q , head H , shaft power P and efficiency η . These parameters provide detailed information about the operating characteristics of the centrifugal pump. Further, the characteristic curves of a centrifugal pump are an intuitive way to describe its performance, and they reflect the mutual relationship between the flow rate, head, power, and efficiency of the centrifugal pump at a specific rotational speed.

[0099] The operating state of a centrifugal pump system will change due to factors such as changes in demand and the regulation of a variable-speed multi-pump parallel system. Therefore, high-precision digital modeling of the characteristic curves of a centrifugal pump is a prerequisite for establishing a mathematical model for intelligent regulation of a variable-speed multi-pump parallel system.

[0100] In this embodiment, the characteristic curves of the centrifugal pump Q - H are fitted with a quadratic polynomial as shown in Equation (1.1): (1.1)

[0101] where a 1, a 2, a 3 are fitting coefficients, and their values are determined through comparative analysis of the test results and the fitting curve.

[0102] Q - P The curve fitting is performed using a cubic polynomial, and its expression is as shown in Equation (1.2):

[0103] (1.2)

[0104] where b 1, b 2, b 3, b 4 are fitting coefficients.

[0105] Q - η The curve is obtained by fitting using the cubic polynomial fitting method, and its expression is as shown in Equation (1.3):

[0106] (1.3)

[0107] where c 1,[[]]c 2, c 3, c 4 are fitting coefficients.

[0108] The method for constructing the digital characteristic curve of a centrifugal pump with variable speed is as follows:

[0109] When the speed of the centrifugal pump changes, its characteristic curve will change accordingly. Equation (1.4) reveals the specific proportional relationship existing between its performance parameters when the centrifugal pump is geometrically similar and has similar flow conditions:

[0110] , , , (1.4)

[0111] Wherein, n 0 is the rated speed, are respectively the characteristic parameters of the centrifugal pump at the rated speed. η 1 and η 2 are respectively the efficiencies of the centrifugal pump corresponding to two speeds.

[0112] Define the speed ratio k = n / n 0. Substitute the above equations into Equations (1.1) to (1.3) respectively, then the digital expression of the characteristic curve of the centrifugal pump under variable speed is shown in Equations (1.5) - (1.7):

[0113] (1.5)

[0114] (1.6)

[0115] (1.7)

[0116] The method for constructing the digital characteristic curve of parallel centrifugal pumps with variable speed is as follows:

[0117] When m o identical centrifugal pumps of the same model are operating in parallel, the total flow rate Q t and the total shaft power P t are respectively Q times the flow rate P and the shaft power m o of a single centrifugal pump, that is:

[0118] (1.8)

[0119] (1.9)

[0120] Substitute equations (1.8), (1.9) and the speed ratio k into equations (1.5) and (1.6), and the digital characteristic curves of variable-speed multi-pump parallel operation can be obtained: Q t - H t and Q t - P t Digital characteristic curves:

[0121] (1.10)

[0122] (1.11)

[0123] where H t is the head of the parallel centrifugal pumps, and the head of each pump is H t at this time; P t is the total shaft power of the parallel centrifugal pumps; Q t is the total flow rate of the parallel centrifugal pumps; m o is the number of operating centrifugal pumps in the variable-speed multi-pump parallel system; k is the speed ratio.

[0124] The method for constructing the calculation of pipeline loss characteristics is as follows:

[0125] When constructing the mathematical model of the variable-speed multi-pump parallel system, not only the variable-speed characteristics of the parallel pumps need to be considered, but also the pipeline loss of the variable-speed multi-pump parallel system needs to be accurately predicted and modeled. Assuming that the flow of water in the pipeline is in an ideal state, that is, the flow is relatively stable and incompressible, the relationship between the pipeline flow rate and the head loss can be expressed as:

[0126] (1.12)

[0127] Among them, h loss is the pipeline head loss, and λ is the friction coefficient of the inner wall of the pipeline; L is the length of the long straight pipeline, with the unit of m; D is the inner diameter of the long straight pipe, with the unit of m; Q g is the volume flow rate in the pipe, with the unit of m³ / h; g is the acceleration due to gravity.

[0128] When there is a certain elevation difference between the inlet and outlet of a variable-speed multi-pump parallel system, that is, when the net head of the pipeline is not zero, it is also necessary to overcome the work done by the net head. Since each pump absorbs water from the same position and the pipeline layout is the same, ignoring the loss differences of each branch pipe before the parallel node, the pipeline loss characteristic curve model of the variable-speed multi-pump parallel system can be written in the following form:

[0129] (1.13)

[0130] In the formula, H x is the pipe resistance head of the variable-speed multi-pump parallel system, with the unit of m; H st is the net head of the pipeline, with the unit of m; S i is the pipe resistance coefficient of each branch pipe before the parallel node, with the unit of h 2 / m 5 ; S p is the pipe resistance coefficient of the main pipeline of the variable-speed multi-pump parallel system, with the unit of h 2 / m 5 ;

[0131] Preferably, in this embodiment, by constructing a digital model of the variable-speed characteristics of the multi-pump parallel system through numerical simulation, establishing an objective function, setting constraint conditions, etc., a mathematical model for the intelligent scheduling of the variable-speed multi-pump parallel system is constructed.

[0132] Step S2: Based on the digital characteristic curve obtained by the above solution, determine the objective function with the minimum total shaft power and the constraint conditions for operation in a specific high-efficiency interval, and establish a mathematical model for the intelligent regulation of the variable-speed multi-pump parallel system. The method for establishing the objective function is as follows:

[0133] The different choices of the objective function will have an important impact on the overall solution difficulty and solution results of the optimization problem. Taking the minimum total shaft power of the multi-pump parallel system as the objective function, therefore:

[0134] (2.1)

[0135] In the formula, P t represents the total shaft power of the variable-speed multi-pump parallel system; P i represents the i th centrifugal pump's shaft power; W i is the i th centrifugal pump's state factor, indicating the iOperating state of a centrifugal pump (1 - operating state, 0 - non-operating state); k is the speed ratio of the centrifugal pump; m o is the number of operating centrifugal pumps in a variable-speed multi-pump parallel system; Q is the flow rate of a single pump.

[0136] Preferably, in this embodiment, the constraint conditions may include the head constraint condition of the variable-speed multi-pump parallel system, the starting and stopping number constraint condition, the flow rate constraint condition of the variable-speed multi-pump parallel system, the speed ratio constraint condition, the pump group operating condition constraint condition, etc. The method of setting the constraint conditions is as follows:

[0137] Setting of the head constraint condition of the variable-speed multi-pump parallel system:

[0138] Since the variable-speed multi-pump parallel system must overcome the net head of the pipeline and loss work during operation. At the same time, considering the stability of the operation of the variable-speed multi-pump parallel system, its head constraint is expressed as:

[0139] (2.2)

[0140] In the formula, m o is the number of operating centrifugal pumps in the variable-speed multi-pump parallel system; is the head of each operating centrifugal pump; H a is the required head.

[0141] Setting of the starting and stopping number constraint condition:

[0142] (2.3)

[0143] In the formula, m max is the maximum number of pumps designed to operate in the pump station.

[0144] Setting of the flow rate constraint condition of the variable-speed multi-pump parallel system:

[0145] (2.4)

[0146] Among them, Q a is the required flow rate.

[0147] Setting of the speed ratio constraint condition:

[0148] Due to the limitations of support, lubrication and heat dissipation, the operating speed of the centrifugal pump and the motor generally does not exceed the rated speed. When its operating speed is less than a certain value, the motor efficiency will decrease significantly. The speed ratio constraint formula of the centrifugal pump can be taken as shown in formula (2.5):

[0149] (2.5)

[0150] Setting of operating condition constraints for parallel centrifugal pumps:

[0151] When the centrifugal pump operates near the design operating point, the flow lines inside the pump are relatively stable and have high efficiency. At this time, the operating reliability of the pump is relatively high; when it operates seriously off the design condition, the efficiency will decrease significantly, and there are often situations such as vortices, cavitation and erosion in the pump, which will reduce the service life of the centrifugal pump. At this time, the operating reliability of the pump is relatively low. In this embodiment, the centrifugal pump adopts a kQ r , 1.1 kQ r operating flow rate range of operating condition constraints to ensure the stability of operation. The operating condition constraints are shown in formula (2.6):

[0152] (2.6)

[0153] In the formula, Q r is the design flow rate at the rated speed.

[0154] Then the operable total flow rate range of the variable-speed multi-pump parallel system at any speed is:

[0155] (2.7)

[0156] Step S3: Combining the characteristics of the particle swarm optimization algorithm and the golden sine algorithm, an improved particle swarm algorithm GPSO based on the golden sine is proposed. By introducing an adaptive strategy and dynamic reverse learning with global optimal guidance, the mathematical model is solved and calculated.

[0157] In this embodiment, the mathematical model is solved by constructing an improved particle swarm algorithm GPSO based on the golden sine. By introducing an adaptive strategy and dynamic reverse learning with global optimal guidance, the performance of global search and local search is balanced.

[0158] In this embodiment, the improved particle swarm algorithm, that is, the improved particle swarm algorithm GPSO based on the golden sine, which integrates the adaptive acceleration and dynamic reverse learning strategy with global optimal guidance.

[0159] As Figure 2 shown, the specific implementation steps of the improved particle swarm algorithm GPSO are as follows:

[0160] Step 1: Given the population size N, the maximum number of iterations T, randomly initialize the positions of the population;

[0161] Step 2: Calculate the fitness value of each individual according to the minimum total shaft power of the objective function;

[0162] Step 3: Record the global optimal and individual optimal positions and determine whether the termination criterion is satisfied. When the termination criterion is not satisfied, execute Step 4; when the termination criterion is satisfied, execute Step 8;

[0163] Step 4: Perform adaptive factor update, changing the weight factor and learning factor;

[0164] Step 5: Update the velocity and position of the particles;

[0165] Step 6: Perform global-optimal-guided reverse dynamic learning;

[0166] Step 7: Solve using the golden sine optimization algorithm and perform boundary condition processing. After the boundary condition processing is completed, return to execute Step 2;

[0167] Step 8: Output the solution result.

[0168] Among them, the termination criterion includes the maximum number of iterations and the required accuracy; when the maximum number of iterations or the required accuracy is reached, it will be judged that the termination criterion is satisfied, the loop will stop, and the solution result will be output.

[0169] In this embodiment, the position update strategy based on the golden sine algorithm is specifically as follows:

[0170] Based on the particle swarm algorithm, let the particles traverse all points on the unit circle, so that the individuals tend to the optimal value. Since both the particle swarm optimization algorithm and the golden sine algorithm take particles as individuals, the position update formula of the golden sine algorithm is:

[0171] (3.1)

[0172] In the formula, x u, l is the position of individual l at the u -th iteration, P o is its optimal position, R 1 and R 2 are random numbers, which respectively determine the moving distance and position update direction of the individual in the next iteration, and .

[0173] θ 1 and θ 2 are the introduced golden ratio numbers, as shown in formula (3.2):

[0174] ; (3.2)

[0175] Among them, τ is the golden ratio, .

[0176] In this embodiment, the adaptive strategy is as follows:

[0177] The adaptive strategy is to adjust each parameter of the algorithm, mainly by adjusting the weight factor in formula (3.3) w， so as to balance the global search ability and local exploration ability of the algorithm, and adjust the learning factor in formula (3.4) c , to affect the local exploration ability of the algorithm. The following are the adjustment formulas for the weight factor and the learning factor:

[0178] (3.3)

[0179] In the formula, w max = 1.2, w min = 0.4 are respectively the maximum and minimum values of the selected weight factor, d u,v is the distance of the v latitude, d max,v is the v maximum distance of the latitude;

[0180] (3.4)

[0181] In the formula, c 1,u,v is the individual learning factor, c 1,max = 2.5, c 1,min = 0.5 are respectively the maximum and minimum values of the selected individual learning factor;

[0182] (3.5)

[0183] In the formula, c 2,u,v is the overall learning factor, c 2,max = 3.0, c 2,min = 0.6 are respectively the maximum and minimum values of the selected overall learning factor.

[0184] In this embodiment, the dynamic reverse learning strategy guided by the global optimum is as follows:

[0185] The reverse learning strategy is: evaluate the fitness values of the current solution and the reverse solution at the same time, and select the better solution as the next generation individual. Let x u,v be the u individual in the current population at thev If it is the value on a certain dimension, its reverse solution is:

[0186] (3.6)

[0187] In the formula, ub is the upper bound of the population value range, lb is the lower bound of the population value range, and RAND(0,1) represents generating a random number within (0,1).

[0188] To improve the global performance of the particle swarm algorithm, after each iteration is completed, the global optimal position guides the population to complete dynamic reverse learning by formula (3.6), and screens out the first N individuals with better fitness, and merges the reverse population with the initial population to form a new population. The reverse learning strategy can avoid blind search from causing too long time while expanding the search space, and improves the population distribution while ensuring the convergence of the algorithm.

[0189] Step S4: Use the feedforward artificial neural network FFNN to perform regression prediction on each parameter and establish a surrogate model.

[0190] This embodiment is based on a single-hidden-layer feedforward artificial neural network to construct a neural network surrogate model: When considering practical problems, the electricity price, flow demand, head, etc. often change in equal time periods. If the parameters are substituted into the optimization algorithm to solve each time, it is often difficult to obtain the required optimization results in time, that is, the response time is long. Therefore, this embodiment proposes to use a feedforward artificial neural network to establish a surrogate model, perform regression prediction on the data, and perform fitting processing to solve the problem of long response time.

[0191] Preferably, in this embodiment, the single-hidden-layer feedforward artificial neural network adopted is specifically as follows:

[0192] (4.1)

[0193] In formula (4.1), x q is the input variable; y p,q is the output signal of the p th neuron in the q th layer; w is the weight factor; b p,q is the bias of the p th neuron in the q th layer; Neu is the total number of neurons in this layer; Neu p-1 is the total number of neurons in the p -1th layer; lis the number of iterations; is the p weight factor of the e th neuron in the l -1 layer when the th iteration is performed; f p,q () is the activation function; e is the neuron counting variable.

[0194] 70% of the data in the sample database is used to train the model, and 30% is used to evaluate the performance of the model. To quantify the fitting effect of the surrogate model, the regression coefficient is used to judge the fitting performance. Adjust the regression coefficient can represent the regression performance of the model prediction. Its formula is shown in Equation (4.2). Adjusting the regression coefficient can ignore the influence of the sample quantity on the result. The closer the value is to 1, the better the performance of the model in predicting data. Generally, it is considered that

[0195] (4.2)

[0196] In the formula, S n is the total number of samples, n p is the number of features, y z is the sample target value, is the sample target mean value, is the model prediction value, and z is the number of features.

[0197] This kind of network structure enables the network to learn more complex feature representations, so it has better performance when dealing with complex problems. By introducing the prefix structure, the network can process input data more flexibly and improve the generalization ability of the model.

[0198] Step S5: Obtain the optimal regulation scheme of the variable-speed multi-pump parallel system through the result prediction of the surrogate model.

[0199] In this embodiment, based on the prediction result of the surrogate model, the optimal regulation scheme of the multi-pump parallel system is obtained by combining the scheduling decision variables as follows:

[0200] As Figure 3 shown, the single-hidden-layer feedforward neural network mainly includes a hidden layer, an activation function, and an output layer. As Figure 4 shown, the various data obtained by solving the mathematical model by the improved particle swarm optimization algorithm GPSO are used as the input layer. After training by the hidden layer, the final output layer takes the predicted parameter values and combines the scheduling decision variables to obtain the optimal scheduling scheme; among them, the scheduling decision variable is Qa , which is the target flow rate of the variable-speed multi-pump parallel system.

[0201] The model quickly generates an optimal operation scheduling plan for the pumping station according to real-time input data. These plans include decisions on starting and stopping pumps, speed adjustment, etc., aiming to minimize energy consumption and operating costs while meeting water supply demands.

[0202] Other parts of this embodiment are the same as those of the above-mentioned Embodiment 1, so they will not be elaborated here.

[0203] Embodiment 3:

[0204] Based on any one of the above-mentioned Embodiments 1 - 2, Figure 5 、 Figure 6 、 Figure 7 、 Figure 8 、 Figure 9 、 Figure 10 As shown, taking the parallel operation of four centrifugal pumps of the same model as an example for detailed description.

[0205] Establish a mathematical model of the variable-speed multi-pump parallel system; in this embodiment, the parallel operation of four centrifugal pumps of the same model is used as the test scheme, and the scheme before optimization always keeps the same speed operation of each centrifugal pump.

[0206] By adopting this implementation method, through the way of establishing a mathematical model that accurately describes the operation characteristics of the variable-speed multi-pump parallel system and improving the mathematical model, the working conditions of the variable-speed multi-pump parallel system can be better described.

[0207] To sum up, the fitness function is:

[0208] (4.3)

[0209] In the formula, fit is the fitness value.

[0210] Performance test of the improved particle swarm optimization algorithm GPSO; specifically, to verify the performance of the improved particle swarm optimization algorithm GPSO of the present invention, comparative tests are carried out with the basic particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, the genetic algorithm GA, and the sparrow search algorithm SSA on 23 test functions.

[0211] Set relevant parameters for the relevant optimization algorithms: c 1 = 2, c 2 = 2. For the adaptive particle swarm optimization algorithm APSO, set the relevant parameters: c 1 = 2, c 2 = 2, w max = 0.9,w min = 0.4, and the relevant parameters for the chaotic particle swarm optimization algorithm CPSO are c 1 = 2, c 2 = 2, φ = c 1 + c 2, the contraction factor χ is , and the relevant parameters for GA are P c = 0.8, P m = 0.05. The algorithm has the same settings in all test functions, where the population size N is 50, the maximum number of iterations is 1000, and the number of runs is 30.

[0212] When testing the performance of the algorithm, four indicators, namely the optimal value, the worst value, the median, and the standard deviation, are commonly used. However, relying solely on the common statistical indicators is not sufficient to convince the performance test of the algorithm. In the optimization algorithm, the search speed is an important performance indicator, which measures the ability of the algorithm to find the optimal solution within a given time. Improving the search speed usually means that the algorithm can converge to the optimal solution faster, thus reducing the computing time and resource consumption. Therefore, this paper also uses CRI to examine the search speed of the algorithm. As Figure 5 shown, where (a) is the bar chart of the search speed of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA on the test function f 6, where (b) is the bar chart of the search speed of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA on the test function f 13 on, where (c) is the bar chart of the search speed of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA on the test function f 15 on, where (d) is the bar chart of the search speed of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA on the test function f 18 on.

[0213] From Table 1-1 and Figure 5It can be seen that the improved particle swarm optimization algorithm GPSO performs excellently in most of the indicators. Whether it is a low-dimensional problem or a high-dimensional problem, the improved particle swarm optimization algorithm GPSO shows significant advantages. Compared with other optimization algorithms, the improved particle swarm optimization algorithm GPSO can often quickly solve and converge. From this, it can be seen that the improved particle swarm optimization algorithm GPSO has an obvious advantage in the search speed. Even when the number of iterations is relatively high, it still maintains a relatively fast search speed. Having only an obvious advantage in the search speed is not enough. This test also conducts a comprehensive performance comparison between the improved particle swarm optimization algorithm GPSO and other optimizations. The speed of approaching the optimal solution and the speed of a sequence approaching its limit value are also indicators of the performance of the optimization algorithm.

[0214] As Figure 6 shown, where (a) is the iterative convergence curve of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA tested on the test function f 5, where (b) is the iterative convergence curve of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA tested on the test function f 8, where (c) is the iterative convergence curve of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA tested on the test function f 14 and (d) is the iterative convergence curve of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA tested on the test function f 19 . The improved particle swarm optimization algorithm GPSO is significantly superior to other optimization algorithms in terms of convergence speed and is also superior to other optimization algorithms in terms of fitness.

[0215] As Figure 7 shown, where (a) is the box plot of the statistical indicators of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA tested on the test function f7, where (b) is the box plot of the statistical indicators of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA tested on the test function f 8, and (c) is the box plot of the statistical indicators of the particle swarm optimization algorithm PSO, the adaptive particle swarm optimization algorithm APSO, the chaotic particle swarm optimization algorithm CPSO, the improved particle swarm optimization algorithm GPSO, and the genetic algorithm GA tested on the test functionf 20 Box plot of statistical indicators tested above, where (d) is the particle swarm optimization algorithm PSO, adaptive particle swarm optimization algorithm APSO, chaotic particle swarm optimization algorithm CPSO, improved particle swarm optimization algorithm GPSO, and genetic algorithm GA in the test function f 22 Box plot of statistical indicators tested above. At the same time, the optimal value and standard deviation are selected from the statistical indicators, such as Figure 7 As can be seen from Table 1-1, Table 1-2 and Table 1-3, among the statistical indicators, the improved particle swarm optimization algorithm GPSO is significantly better than other optimization algorithms in most cases. At the same time, in order to better examine the test results of the solution accuracy, as shown in Table 1-4, the Friedman test with a significance level of 0.05 is used to compare various optimization algorithms participating in the test, and it is found that the improved particle swarm optimization algorithm GPSO has an obvious advantage in dealing with this problem, followed by the adaptive particle swarm optimization algorithm APSO. The solution effects of the particle swarm optimization algorithm PSO and the chaotic particle swarm optimization algorithm CPSO are similar, and the genetic algorithm GA has the worst effect.

[0216] Table 1-1 Comparison table of search speeds between the improved particle swarm optimization algorithm GPSO and other optimization algorithms

[0217]

[0218] Table 1-2 Comparison table of optimal values between the improved particle swarm optimization algorithm GPSO and other optimization algorithms

[0219]

[0220] Table 1-3 Comparison table of standard deviations between the improved particle swarm optimization algorithm GPSO and other optimization algorithms

[0221]

[0222] Table 1-4 Friedman ranking result table of the improved particle swarm optimization algorithm GPSO and other optimization algorithms

[0223]

[0224] Establish a surrogate model;

[0225] In the test of this example, as Figure 8 shown, due to the small number of fitting data samples, only a single-hidden-layer cascaded feedforward neural network is used for fitting and prediction, and it is compared with the actual optimization data in the later stage.

[0226] The process of simulation training uses the gradient descent method as the training algorithm. The training objective is that the root mean square error between the predicted value and the actual value is less than 0.001, and the number of training steps does not exceed 1000 steps. During the training process, 70% of the data in the sample database is used to train the model, and 30% is used to evaluate the performance of the model. To quantify the fitting effect of the surrogate model, the regression coefficient is used to judge the fitting performance.

[0227] Adjust the regression coefficient It can represent the regression performance of the model prediction. Fit the 100 sets of data solved when the net head is 10m by the above method, and the results are as Figure 9 shown, where (a) is the Q a - P t prediction curve of the surrogate model; (b) is the Q a - k prediction curve. Q a - P t and Q a - k The adjusted regression coefficients of are 0.9768 and 0.9149 respectively. Although the sample data is small, the adjusted regression coefficients are both greater than 0.9. In actual engineering cases, the algorithm can be used to solve in the initial stage, and a database can be formed after a period of time, and then a surrogate model can be established to quickly obtain the optimal solution.

[0228] Specifically, the above scheme is applied to an actual case:

[0229] The scale of a water supply pump station in a certain area of City A is 80000m 3 / d. The pump station uses 4 double-suction centrifugal pumps of the same model. The pump parameters are shown in Table 2-1. The highest water supply demand and time-of-use electricity price in a certain area of the city are as Figure 10 shown. The current dispatching method of the pump station is to increase the number of operating pumps only when the existing opened pumps are not enough to meet the water supply demand, and all four pumps operate at a constant speed. The staff of the pump station will adjust the number of pumps and valve openings of the pump station according to the water supply demand. The electricity consumption cost of the pump station motor is priced according to the national standard for time-of-use. Among them, the basic electricity price is 0.617 yuan, and the electricity price from 11:00 to 2:00 and from 17:00 to 21:00 is 0.952 yuan. The prices at other times are all the basic prices. The current operation results of the pump station are shown in Table 2-2, and the operation optimization results using the optimization algorithm in this paper are shown in Table 2-3.

[0230] As can be seen from the results in Table 2-2 and Table 2-3, the current daily operating cost of the pumping station is 18,614.34 yuan. By using the optimization algorithm of this embodiment for solution, the optimized daily operating cost of the pumping station can be obtained as 13,742.09 yuan. The power saving rate of the improved optimization scheme is:

[0231]

[0232] Compared with the power saving rate of about 10% of the existing improved optimization algorithm, the energy consumption loss of this optimization algorithm is lower. And compared with the current scheme, the improved scheme reduces the number of pump starts and stops, which improves the service life of the equipment to a certain extent.

[0233] Using the single-hidden-layer prefix neural network in the artificial neural network to perform regression prediction fitting on the data that has been optimized and solved, and verifying according to the actual optimization data, it is found that the prediction effect is good. And applying the improved scheme to an actual engineering example, it is found that compared with the current operation scheme of this pumping station, the power saving rate of the improved optimization scheme is 26.17%, and the number of pump starts and stops is reduced.

[0234] Table 2-1 Parameter Table of Variable-Speed Double-Suction Pump Group

[0235]

[0236] Table 2-2 Current Daily Operation Results Table of the Pumping Station

[0237]

[0238] Table 2-3 Optimized Results of the Pumping Station's Daily Operation

[0239]

[0240] Other parts of this embodiment are the same as any one of the above-mentioned Embodiment 1 - Embodiment 2, so they will not be elaborated here.

[0241] Embodiment 4:

[0242] Based on any one of the above-mentioned Embodiment 1 - Embodiment 3, this embodiment proposes an intelligent scheduling system for coordinated operation of a variable-speed multi-pump parallel system, which is used to execute the intelligent scheduling method for coordinated operation of the variable-speed multi-pump parallel system; it includes a simulation unit, a construction unit, a solution unit, a prediction unit, and an output unit;

[0243] The simulation unit is used to call numerical methods to simulate and calculate the variable-speed performance characteristics of parallel centrifugal pumps and the pipeline loss characteristics of the variable-speed multi-pump parallel system, and obtain digital characteristic curves;

[0244] The building unit is used to establish an intelligent regulation mathematical model for a variable-speed multi-pump parallel system with the minimum total shaft power as the objective function according to the digital characteristic curve under the given demand targets of the variable-speed multi-pump parallel system, namely the demand flow rate Q a and the demand head H a , and under various constraint conditions

[0245] The solving unit is used to solve the intelligent regulation mathematical model of the variable-speed multi-pump parallel system according to the improved particle swarm optimization algorithm based on the golden sine algorithm to obtain operation data

[0246] The prediction unit is used to call a feedforward artificial neural network to establish a surrogate model and regressively predict the operation data

[0247] The output unit is used to obtain the optimal regulation scheme for the variable-speed multi-pump parallel system according to the prediction result of the surrogate model

[0248] This embodiment also provides an electronic device, including a memory and a processor; a computer program is stored on the memory; when the computer program is executed on the processor, the above-mentioned intelligent scheduling method for the coordinated operation of the variable-speed multi-pump parallel system is implemented

[0249] This embodiment also provides a computer-readable storage medium, on which a computer instruction is stored; when the computer instruction is executed on the above-mentioned electronic device, the above-mentioned intelligent scheduling method for the coordinated operation of the variable-speed multi-pump parallel system is implemented

[0250] Other parts of this embodiment are the same as any one of the above-mentioned Embodiment 1 - Embodiment 3, so they will not be described in detail

[0251] The above is only a preferred embodiment of the present invention, and does not impose any form of limitation on the present invention. Any simple modification and equivalent change made to the above embodiments based on the technical essence of the present invention shall fall within the protection scope of the present invention

Claims

1. An intelligent scheduling method for coordinated operation of a variable-speed multi-pump parallel system, characterized in that, First, call the numerical method to simulate and calculate the performance characteristics of parallel centrifugal pumps and the pipeline loss characteristics of the variable-speed multi-pump parallel system, and obtain the digital characteristic curves. Secondly, according to the digital characteristic curves, with the minimum total shaft power as the objective function, establish an intelligent control mathematical model for the variable-speed multi-pump parallel system according to the set demand objectives and constraint conditions of the variable-speed multi-pump parallel system. Then, improve the particle swarm optimization algorithm according to the golden sine algorithm, solve the intelligent control mathematical model of the variable-speed multi-pump parallel system, and obtain the operation data. Call the feedforward artificial neural network to establish a surrogate model and regressively predict the operation data. Finally, according to the prediction results of the surrogate model, obtain the optimal control scheme for the variable-speed multi-pump parallel system; The intelligent scheduling method for the coordinated operation of the variable-speed multi-pump parallel system specifically includes the following steps: Step S1: Call the numerical method to simulate and calculate the variable-speed performance characteristics of parallel centrifugal pumps and the pipeline loss characteristics of the variable-speed multi-pump parallel system, and obtain the digital characteristic curves; Step S2: According to the digital characteristic curves, with the minimum total shaft power as the objective function, establish an intelligent control mathematical model for the variable-speed multi-pump parallel system according to the set demand objectives and constraint conditions of the variable-speed multi-pump parallel system; Step S3: Improve the particle swarm optimization algorithm according to the position update strategy, adaptive strategy, and global optimal-guided dynamic reverse learning, solve the intelligent control mathematical model of the variable-speed multi-pump parallel system, and obtain the operation data; Step S4: Call the feedforward artificial neural network to establish a surrogate model, regressively predict the operation data, and obtain the prediction results; Step S5: According to the prediction results of the surrogate model, combine with the scheduling decision variables to obtain the optimal coordinated operation scheduling scheme for the variable-speed multi-pump parallel system; The specific steps of the said Step S3 include the following steps: Step S31: Construct a position update strategy according to the golden sine algorithm; the position update strategy is used to update the position of the particle; Step S32: Construct an adaptive strategy according to the set weight factor and learning factor; the adaptive strategy is used to adaptively update the weight factor and learning factor; Step S33: Based on the reverse solution of the value of the u -th individual in the current population on the v -dimensional space, construct a globally optimal-guided dynamic reverse learning strategy; the globally optimal-guided dynamic reverse learning strategy is used to, after the particle swarm optimization algorithm iteration is completed, complete dynamic reverse learning according to the globally optimal-guided dynamic reverse learning strategy, screen out the top N individuals with the optimal fitness, and merge the reverse population with the initial population to form a new population; Step S34: Improve the particle swarm optimization algorithm according to the position update strategy, adaptive strategy, and global optimal-guided dynamic reverse learning strategy, and solve the intelligent control mathematical model of the variable-speed multi-pump parallel system according to the improved particle swarm optimization algorithm to obtain the operation data.

2. The intelligent scheduling method for coordinated operation of a variable-speed multi-pump parallel system according to claim 1, characterized in that The specific steps of the said Step S1 include the following steps: Step S11: According to the factory characteristic curve of a single centrifugal pump, or by establishing a three-dimensional model of a single centrifugal pump, the characteristic curve of a single centrifugal pump is calculated by using a numerical simulation method based on performance prediction; the characteristic curve of a single centrifugal pump is fitted by a polynomial fitting method to obtain the single-pump flow rate Q - head H , single-pump flow rate Q - shaft power P and single-pump flow rate Q - efficiency η mathematical expressions of the characteristic curves; Step S12: According to the similarity theory of centrifugal pumps, calculate the characteristic curve of a single centrifugal pump under variable speed conditions, and establish the mathematical expressions of the single-pump flow Q -head H , single-pump flow Q -shaft power P and single-pump flow Q -efficiency η characteristic curves. Step S13: Calculate the characteristic curve of the variable-speed multi-pump parallel system according to the characteristics of flow rate superposition and constant head of parallel centrifugal pumps, and establish the total flow rate of the variable-speed multi-pump parallel system Q t - Total shaft power P t And total flow rate Q t - Head of parallel pumps H t Mathematical expressions of the characteristic curve Step S14: Construct a pipeline loss characteristic curve according to the net head of the pipeline, the pipe resistance coefficient of the branch pipe before the parallel node, the pipe resistance coefficient of the main pipeline of the variable-speed multi-pump parallel system, the flow rate of the branch pipeline, and the total flow rate of the parallel centrifugal pumps.

3. The intelligent scheduling method for coordinated operation of a variable-speed multi-pump parallel system according to claim 2, characterized in that The specific steps of the said Step S2 include the following steps: Step S21: Construct an objective function according to the characteristic curve of a single centrifugal pump, the characteristic curve of a single centrifugal pump under variable speed conditions, the characteristic curve of the variable-speed multi-pump parallel system, the pipeline loss characteristic curve, the total shaft power of the variable-speed multi-pump parallel system, the shaft power of each centrifugal pump, the state factor of the centrifugal pump, the centrifugal pump speed ratio, the number of operating units of the parallel centrifugal pumps, and the single-pump flow rate; Step S22: Construct constraint conditions; the construction of constraint conditions includes: constructing the head constraint of the variable-speed multi-pump parallel system according to the pipeline loss characteristic curve, the head characteristic curve of the parallel pumps, and the required head, constructing the start-stop number constraint according to the maximum number of operating units, constructing the flow constraint of the variable-speed multi-pump parallel system according to the water supply demand flow, constructing the speed ratio constraint according to the rated speed of the centrifugal pump, and constructing the operating condition constraint of the centrifugal pump according to the rated flow of the centrifugal pump and the set constraint interval; Step S23: Establish an intelligent regulation mathematical model of the variable-speed multi-pump parallel system according to the objective function and the constraint conditions.

4. The intelligent scheduling method for coordinated operation of a variable-speed multi-pump parallel system according to claim 1, wherein The specific steps of step S4 include the following steps: Step S41: Call a feedforward artificial neural network with a single hidden layer to construct a surrogate model; Step S42: Adjust the regression coefficients of the surrogate model, use the operating data as the input layer, and train through a single hidden layer to obtain the prediction results.

5. The intelligent scheduling method for coordinated operation of a variable-speed multi-pump parallel system according to claim 3, characterized in that, The objective function established in step S21 is: Among them, P t is the total shaft power of the variable-speed multi-pump parallel system; P i is the shaft power of the i th centrifugal pump; W i is the state factor of the i th centrifugal pump; b 1, b 2, b 3, b 4 are the fitting coefficients of the polynomial fitting of the centrifugal pump flow-power curve; k is the speed ratio of the centrifugal pump; m o is the number of operating centrifugal pumps in the variable-speed multi-pump parallel system; Q is the single-pump flow rate.

6. An intelligent scheduling system for coordinated operation of a variable-speed multi-pump parallel system, which is used to execute the intelligent scheduling method for coordinated operation of the variable-speed multi-pump parallel system as described in claim 1; characterized in that, It includes a simulation unit, a construction unit, a solution unit, a prediction unit, and an output unit; The simulation unit is used to numerically simulate and calculate the performance characteristics of the variable-speed multi-pump parallel centrifugal pump and the pipeline loss characteristics of the variable-speed multi-pump parallel system to obtain the digital characteristic curve; The building unit is used to establish an intelligent control mathematical model for a variable-speed multi-pump parallel system with the minimum total shaft power as the objective function according to the digital characteristic curve, under the given demand targets of the variable-speed multi-pump parallel system, i.e., the required flow rate Q a and the required head H a , and under various constraint conditions; The solution unit is used to solve the intelligent regulation mathematical model of the variable-speed multi-pump parallel system according to the improved particle swarm optimization algorithm based on the golden sine algorithm to obtain the operating data; The prediction unit is used to call a feedforward artificial neural network to establish a surrogate model and perform regression prediction on the operating data; The output unit is used to obtain the optimal regulation scheme of the variable-speed multi-pump parallel system according to the prediction results of the surrogate model.

7. An electronic device, characterized in that, It includes a memory and a processor; a computer program is stored on the memory; when the computer program is executed on the processor, the intelligent scheduling method for the coordinated operation of the variable-speed multi-pump parallel system according to any one of claims 1-5 is implemented.

8. A computer-readable storage medium, characterized in that, A computer instruction is stored on the computer-readable storage medium; when the computer instruction is executed on the electronic device according to claim 7, the intelligent scheduling method for the coordinated operation of the variable-speed multi-pump parallel system according to any one of claims 1-5 is implemented.

Images