Intelligent optimized operation decision-making method for sewage treatment lift pump based on pressure maintaining constraint

Abstract

In order to solve the problems that in the urban sewage treatment process, a water pump control system is generally high in energy consumption, inaccurate in regulation and control, poor in system stability and the like, the invention designs a sewage treatment lift pump intelligent optimization operation decision-making method based on pressure maintaining constraint. The method comprises the steps that firstly, a pump frequency intelligent optimization decision model based on pressure maintaining constraint is established, and joint optimization of pressure constraint, flow and energy consumption is achieved; secondly, optimization regulation and control are carried out on setting of the frequency of the lifting pump through a particle swarm algorithm, and intelligent frequency regulation and system energy efficiency optimization of a pump station are achieved. And finally, in combination with sewage treatment lift pump operation data and the optimization decision model, optimization regulation and control of lift pump frequency combination are realized. The result shows that the model can quickly and accurately make a decision on the frequency of the lift pump, and can respond to load fluctuation and external environment change in real time in the actual operation process, so that the stability and energy efficiency of the system are ensured.

Description

Technical Field

[0001] The present invention designs a method for intelligent optimization operation decision of sewage treatment lifting pump based on pressure maintenance constraint. By collecting sewage treatment lifting pump operation data and pump frequency intelligent optimization decision model based on pressure maintenance constraint, the pump station system optimizes energy efficiency to the greatest extent while ensuring pressure stability. By adjusting the operating frequency of the lifting pump, the method can automatically adapt to changes in system load and intelligently adjust the working state of the pump, thereby improving the overall operating efficiency and energy-saving effect of the system. This method can effectively solve the problems of low energy efficiency and unstable operation of traditional pump station control systems when load fluctuates and operating status changes, and provides a highly efficient and intelligent optimization and control solution for the pump station system. Background Art

[0002] With the acceleration of industrialization, sewage treatment systems play an important role in urban water supply, drainage and various industrial applications. However, in the sewage treatment process, existing water pump control systems generally have problems such as high energy consumption, inaccurate regulation and poor system stability. These problems not only lead to energy waste, but may also affect the long-term operation safety and efficiency of the water pump system. In traditional water pump control methods, empirical rules or traditional PID control strategies are usually relied on, which are often difficult to cope with complex working environments and dynamically changing load requirements. In order to solve the above problems, intelligent control technology has been gradually applied to pump station systems in recent years. By introducing adaptive adjustment algorithms, real-time monitoring and intelligent adjustment of pump station systems can be achieved, and the operating status of pump stations can be optimized, thereby improving the efficiency of the system and reducing energy consumption. In this context, this patent proposes a sewage treatment lift pump intelligent optimization operation decision method based on pressure maintenance constraints. This method establishes a pump frequency intelligent optimization decision model based on pressure maintenance constraints through real-time data acquisition, combined with constraints such as pressure, flow and energy consumption. In this model, the frequency of the lift pump is the main decision variable. The goal is to maintain the pressure stability of the system by adjusting the operating frequency of the pump while optimizing energy efficiency. This technology can not only effectively reduce energy consumption, but also improve the stability and operation efficiency of the pump station. Therefore, the present invention is of great significance in optimizing and regulating the pump station system.

[0003] The present invention designs a method for intelligent optimization operation decision of sewage treatment lifting pump based on pressure maintenance constraint. It mainly establishes a pump frequency intelligent optimization decision model based on pressure maintenance constraint from three aspects: pressure model, flow model and energy consumption model, and uses particle swarm optimization algorithm to update the model parameters to realize intelligent frequency regulation of pump station and system energy efficiency optimization. This method can automatically adapt to different working conditions and load changes, realize adaptive regulation and efficient operation of pump station, and provide reference for optimization and control of sewage treatment pump station system. Summary of the invention

[0004] The present invention obtains a sewage treatment lifting pump intelligent optimization operation decision method based on pressure maintenance constraint. The method establishes an intelligent pump frequency control optimization model based on pressure maintenance constraint, realizes the joint optimization of pressure, flow and energy consumption, and then uses particle swarm optimization algorithm to dynamically optimize model parameters. The method can make adaptive adjustments according to field data, optimize the operating frequency of the pump, and make the pump station minimize energy consumption under the premise of ensuring the stability of system pressure, providing a reference for the optimization and control of sewage treatment pump station system.

[0005] The present invention adopts the following technical solutions and implementation steps:

[0006] A method for intelligent optimization operation decision of a sewage treatment lift pump based on pressure maintenance constraint is provided, which includes: collecting operation data of the sewage treatment lift pump, establishing a pump frequency intelligent optimization decision model based on pressure maintenance constraint, designing a pump frequency setting optimization strategy based on particle swarm algorithm, and optimizing and controlling the sewage treatment lift pump, including the following steps:

[0007] (1) Collecting sewage treatment lift pump operation data

[0008] Taking the operation process of the pump station as the research object, the main pipe flow of the pump station, the pump frequency, the total pressure of the pump station, the head of the pump unit and the total power consumption of the pump station are selected as the relevant variables for pump station regulation. The main pipe flow of the pump station at time t is collected as the vector q(t)=[q1(t),q2(t),…,q k (t)] T , T represents matrix transpose, the number of arrays collected k = 1, 2...50, the value range of the total flow is [1000 cubic meters / hour, 8000 cubic meters / hour], and the boost pump frequency at time t is the matrix F(t) = [f1(t), f2(t), ..., f i (t)], the number of lift pumps i = 1, 2, 3, 4, the frequency of the i-th lift pump is the vector f i (t) = [f 1,i (t),f 2,i (t),…,f k,i (t)] T , where f k,i (t) is the kth frequency value of the i-th boost pump, and the frequency value range is [34 Hz, 45 Hz]. The total pressure of the pumping station at time t is the vector s(t) = [s1(t), s2(t),…, s k (t)] T The value range of total pressure is [0 N / m2, 200 N / m2], and the operating head of the lifting pump unit at time t is the matrix H(t) = [h1(t), h2(t), …, h i (t)] T , the lift of the i-th lift pump is vector hi (t) = [h 1,i (t),h 2,i (t),…,h k,i (t)] T , where h k,i (t) is the kth lift value of the i-th lift pump, and the lift value range is [5 meters, 45 meters]. The total power consumption of the pumping station at time t is the vector p(t) = [p1(t), p2(t), …, p k (t)] T , the power value is in kilowatts;

[0009] (2) Establishing an intelligent optimization decision model for pump frequency based on pressure maintenance constraints

[0010] The intelligent optimization decision model for pump frequency based on pressure maintenance constraint includes pump pressure constraint, flow model and energy consumption model, which aims to keep the total pressure and total pipe flow of the pump station system stable within the target range and minimize the overall energy consumption of the pump station.

[0011] 1) Pump pressure constraint

[0012] Based on the relationship between the total pressure of the pumping station and the pumping frequency, flow rate and head, ensure that the total pressure of the pumping station is maintained in a stable range [144 Newtons / square meter, 146 Newtons / square meter];

[0013] According to the collected pump station main flow value q k (t) and the pump frequency vector value f k,i (t), the approximate linear relationship between the main flow of the pump station and the booster pump frequency at time t is expressed as:

[0014]

[0015] Among them, the constant r1 related to flow rate and pump frequency is 60;

[0016] According to the collected operating head of the lifting pump unit and the main pipe flow of the pump station, the relationship between the head and the main pipe flow at time t is expressed as:

[0017]

[0018] Among them, the maximum head of each lifting pump is h max =45 meters, maximum flow rate of the main pipe of the pump station q max = 8000 cubic meters / hour. The mathematical relationship between the total pressure and flow rate of the pump station is expressed using the additional pressure loss term. At time t, the pipeline friction loss model in fluid mechanics can be approximately expressed as:

[0019] l(t)=r2·(q k (t)) 2(3)

[0020] Among them, the constant r2=0.0005 related to pipeline characteristics and fluid characteristics, etc., based on the mathematical relationship between the total pressure of the pump station and the pump frequency, flow rate and head, the total pressure value of the pump station s total (t) is expressed as:

[0021] s total (t) = 1000 9.180 h i (t)+l(t) (4)

[0022] In order to ensure that the total pressure of the pumping station is within the target range [144 Newtons / square meter, 146 Newtons / square meter], a pressure constraint penalty function is defined. The pressure constraint penalty function at time t is expressed as:

[0023]

[0024] 2) Traffic model

[0025] The flow model is used to ensure that the water supply flow of the pump station remains within a reasonable range and meets the actual production water supply demand. k (t) Constraints:

[0026] q min ≤q k (t)≤q max (6)

[0027] Among them, the minimum flow rate of the pump station main pipe q min = 1000 cubic meters / hour, define the flow penalty function, the penalty function of the flow at time t is expressed as:

[0028]

[0029] 3) Energy consumption model

[0030] A mathematical model is established for the power consumption of each lifting pump and the corresponding pump frequency. According to the collected pump frequency and total energy consumption, the mathematical relationship between the total power consumption and pump frequency at time t is expressed as:

[0031]

[0032] Among them, the constant r3 related to the characteristics of the lift pump (such as head, pump efficiency, etc.) is 0.05;

[0033] 4) Set the fitness function

[0034] Considering the pressure stability, flow rationality and energy efficiency minimization goals, the fitness function is designed as the objective function. The fitness function at time t is expressed as:

[0035] F(x i (t)) = a·P(s total (t))+b·Q(q k (t))+c·E total (t) (9)

[0036] Among them, x i (t) is the operating parameter of the i-th lift pump, and the frequency combination vector is x(t) = [x1(t), x2(t),…, x i (t)], the weight coefficient of the pressure stability target is a = 0.4, the weight coefficient of the flow rationality target is b = 0.2, and the weight coefficient of the energy efficiency minimization target is c = 0.4;

[0037] (3) Design a boost pump frequency setting optimization strategy based on particle swarm algorithm

[0038] In the boost pump frequency setting optimization strategy based on particle swarm algorithm, the particle optimization process is to gradually approach the optimal solution by updating the frequency combination and speed of the particles;

[0039] ① The frequency combination vector of the particle at time t is x(t), and the velocity of the particle is vector v(t) = [v1(t), v2(t), …, v i (t)], v(t) determines the direction and step of frequency adjustment of each boost pump. Set the current time t = 1, and initialize the frequency combination vector x(1) = [x1(1), x2(1), …, x i (1)], where x1(1) = f 1,1 (t), x2(1) = f 1,2 (t), x i (1) = f 1,i (t), initialize the velocity vector v(1) of the particle to each element value v0, and initialize the optimal frequency value L of the i-th boost pump i (1) = 34, initialize the frequency value G of the i-th boost pump in the optimal frequency combination i (1) = 34;

[0040] ②According to the initial value L i (1) and G i (1) Adjust the particle speed. The speed value v of the particle of the i-th lift pump at time t+1 i The update rule for (t+1) is expressed as:

[0041] v i (t+1)=w·v i (t)+m1·n1·(L i (t)-f i (t))+m2·n2·(G i(t)-f i (t)) (10)

[0042] Among them, the inertia weight w = 0.7, the learning factor m1 = 0.5, m2 = 0.5, and the value range of random numbers n1 and n2 is [0, 1];

[0043] ③ Update the particle frequency combination according to the frequency combination of the particle at time t and the particle speed at the next moment. The particle frequency value x of the i-th boost pump at time t+1 is i The update rule for (t+1) is expressed as:

[0044] x i (t+1)=x i (t)+v i (t+1) (11)

[0045] Among them, the frequency combination vector of the particle at time t+1 is x(t+1)=[x1(t+1),x2(t+1),…,x i (t+1)];

[0046] ④ According to formula (9), calculate the fitness function corresponding to the particle frequency combination at the current time t. When the optimal frequency value L of the i-th boost pump at time t+1 is i The update rule for (t+1) is expressed as:

[0047]

[0048] ⑤ When the i-th boost pump frequency value G in the optimal frequency combination at time t+1 i The update rule for (t+1) is expressed as:

[0049] G i (t+1)=argmin(F(L i (t))) (13)

[0050] ⑥If t<Z, the number of iterations Z=100, then t=t+1, return to step ② to continue training, if t=Z, or L i (t) and G i If the value of (t) does not change in 20 consecutive updates, the training is completed;

[0051] (4) Optimization and control of sewage treatment lift pumps

[0052] The mathematical relationship between pump frequency and pump station operation related parameters is described based on the operation data of sewage treatment lift pumps. The pump frequency and the main pipe flow of the pump station are used as model inputs. Formulas (1)-(8) are used to establish a pump frequency intelligent optimization decision model based on pressure maintenance constraints from three parts: pump pressure constraints, flow model and energy consumption model. The fitness function is defined and the decision model is trained using the particle swarm algorithm according to formulas (10)-(13). The fitness function is minimized to ensure that the pump station can reduce energy consumption as much as possible while satisfying the pressure constraints and main pipe flow constraints. Finally, the optimal frequency combination vector x′(t) is output to achieve optimized regulation of sewage treatment lift pumps.

[0053] The creativity of the present invention is mainly reflected in:

[0054] (1) Aiming at the problems of unstable pressure and excessive energy consumption in sewage treatment pump station systems, the present invention proposes an intelligent optimization operation decision method for sewage treatment booster pumps based on pressure maintenance constraints to achieve optimal control of booster pump frequency in sewage treatment pump station systems;

[0055] (2) The present invention establishes an intelligent optimization decision model for pump frequency based on pressure maintenance constraints, performs joint control optimization from the three parts of pressure constraints, flow constraints and energy consumption constraints, and uses a particle swarm optimization algorithm to update parameters to achieve automatic optimization of pump frequency, thereby minimizing energy consumption while ensuring stable pressure of the pump station system. The invention enables the system to adaptively adjust the pump frequency under different working conditions, while ensuring the optimal operating state of the pump station under various load conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram of the process of realizing the optimization and control of the sewage treatment lifting pump according to the present invention;

[0057] Figure 2 This is a graph showing the experimental results of the present invention on the intelligent optimization operation decision of the lift pump. DETAILED DESCRIPTION

[0058] The present invention obtains a sewage treatment lift pump intelligent optimization operation decision method based on pressure maintenance constraint; the lift pump frequency and the pump station main pipe flow are selected as input variables of the pump frequency intelligent optimization decision model based on pressure maintenance constraint, and the optimal frequency combination of the lift pump is used as the output variable;

[0059] The experimental data comes from the daily report of equipment analysis of a certain urban sewage treatment plant; after removing abnormal experimental samples, 200 sets of available data remain, of which 150 sets are used as training samples and the remaining 50 sets are used as test samples.

[0060] The present invention adopts the following technical solutions and implementation steps:

[0061] A method for intelligent optimization operation decision of a sewage treatment lift pump based on pressure maintenance constraint is provided, which includes: collecting operation data of the sewage treatment lift pump, establishing a pump frequency intelligent optimization decision model based on pressure maintenance constraint, designing a pump frequency setting optimization strategy based on particle swarm algorithm, and optimizing and controlling the sewage treatment lift pump, including the following steps:

[0062] (1) Collecting sewage treatment lift pump operation data

[0063] Taking the operation process of the pump station as the research object, the main pipe flow of the pump station, the pump frequency, the total pressure of the pump station, the head of the pump unit and the total power consumption of the pump station are selected as the relevant variables for pump station regulation. The main pipe flow of the pump station at time t is collected as the vector q(t)=[q1(t),q2(t),…,q k (t)] T , T represents matrix transpose, the number of arrays collected k = 1, 2...50, the value range of the total flow is [1000 cubic meters / hour, 8000 cubic meters / hour], and the boost pump frequency at time t is the matrix F(t) = [f1(t), f2(t), ..., f i (t)], the number of lift pumps i = 1, 2, 3, 4, the frequency of the i-th lift pump is the vector f i (t) = [f 1,i (t),f 2,i (t),…,f k,i (t)] T , where f k,i (t) is the kth frequency value of the i-th boost pump, and the frequency value range is [34 Hz, 45 Hz]. The total pressure of the pumping station at time t is the vector s(t) = [s1(t), s2(t),…, s k (t)] T The value range of total pressure is [0 N / m2, 200 N / m2], and the operating head of the lifting pump unit at time t is the matrix H(t) = [h1(t), h2(t), …, h i (t)] T , the lift of the i-th lift pump is vector h i (t) = [h 1,i (t),h 2,i (t),…,h k,i (t)] T , where h k,i (t) is the kth lift value of the i-th lift pump, and the lift value range is [5 meters, 45 meters]. The total power consumption of the pumping station at time t is the vector p(t) = [p1(t), p2(t), …, p k (t)] T , the power value is in kilowatts;

[0064] (2) Establishing an intelligent optimization decision model for pump frequency based on pressure maintenance constraints

[0065] The intelligent optimization decision model for pump frequency based on pressure maintenance constraint includes pump pressure constraint, flow model and energy consumption model, which aims to keep the total pressure and total pipe flow of the pump station system stable within the target range and minimize the overall energy consumption of the pump station.

[0066] 1) Pump pressure constraint

[0067] Based on the relationship between the total pressure of the pumping station and the pumping frequency, flow rate and head, ensure that the total pressure of the pumping station is maintained in a stable range [144 Newtons / square meter, 146 Newtons / square meter];

[0068] According to the collected pump station main flow value q k (t) and the pump frequency vector value f k,i (t), the approximate linear relationship between the main flow of the pump station and the booster pump frequency at time t is expressed as:

[0069]

[0070] Among them, the constant r1 related to flow rate and pump frequency is 60;

[0071] According to the collected operating head of the lifting pump unit and the main pipe flow of the pump station, the relationship between the head and the main pipe flow at time t is expressed as:

[0072]

[0073] Among them, the maximum head of each lifting pump is h max =45 meters, maximum flow rate of the main pipe of the pump station q max = 8000 cubic meters / hour. The mathematical relationship between the total pressure and flow rate of the pump station is expressed using the additional pressure loss term. At time t, the pipeline friction loss model in fluid mechanics can be approximately expressed as:

[0074] l(t)=r2·(q k (t)) 2 (16)

[0075] Among them, the constant r2=0.0005 related to pipeline characteristics and fluid characteristics, etc., based on the mathematical relationship between the total pressure of the pump station and the pump frequency, flow rate and head, the total pressure value of the pump station s total (t) is expressed as:

[0076] s total (t) = 1000 9.180 h i (t)+l(t) (17)

[0077] In order to ensure that the total pressure of the pumping station is within the target range [144 Newtons / square meter, 146 Newtons / square meter], a pressure constraint penalty function is defined. The pressure constraint penalty function at time t is expressed as:

[0078]

[0079] 2) Traffic model

[0080] The flow model is used to ensure that the water supply flow of the pump station remains within a reasonable range and meets the actual production water supply demand. k (t) Constraints:

[0081] q min ≤q k (t)≤q max (19)

[0082] Among them, the minimum flow rate of the pump station main pipe q min = 1000 cubic meters / hour, define the flow penalty function, the penalty function of the flow at time t is expressed as:

[0083]

[0084] 3) Energy consumption model

[0085] A mathematical model is established for the power consumption of each lifting pump and the corresponding pump frequency. According to the collected pump frequency and total energy consumption, the mathematical relationship between the total power consumption and pump frequency at time t is expressed as:

[0086]

[0087] Among them, the constant r3 related to the characteristics of the lift pump (such as head, pump efficiency, etc.) is 0.05;

[0088] 4) Set the fitness function

[0089] Considering the pressure stability, flow rationality and energy efficiency minimization goals, the fitness function is designed as the objective function. The fitness function at time t is expressed as:

[0090] F(x i (t)) = a·P(s total (t))+b·Q(q k (t))+c·E total (t) (22)

[0091] Among them, x i (t) is the operating parameter of the i-th lift pump, and the frequency combination vector is x(t) = [x1(t), x2(t),…, x i(t)], the weight coefficient of the pressure stability target is a = 0.4, the weight coefficient of the flow rationality target is b = 0.2, and the weight coefficient of the energy efficiency minimization target is c = 0.4;

[0092] (3) Design a boost pump frequency setting optimization strategy based on particle swarm algorithm

[0093] In the boost pump frequency setting optimization strategy based on particle swarm algorithm, the particle optimization process is to gradually approach the optimal solution by updating the frequency combination and speed of the particles;

[0094] ① The frequency combination vector of the particle at time t is x(t), and the velocity of the particle is vector v(t) = [v1(t), v2(t), …, v i (t)], v(t) determines the direction and step of frequency adjustment of each boost pump. Set the current time t = 1, and initialize the frequency combination vector x(1) = [x1(1), x2(1), …, x i (1)], where x1(1) = f 1,1 (t), x2(1) = f 1,2 (t), x i (1) = f 1,i (t), initialize the velocity vector v(1) of the particle to each element value v0, and initialize the optimal frequency value L of the i-th boost pump i (1) = 34, initialize the frequency value G of the i-th boost pump in the optimal frequency combination i (1) = 34;

[0095] ②According to the initial value L i (1) and G i (1) Adjust the particle speed. The speed value v of the particle of the i-th lift pump at time t+1 i The update rule for (t+1) is expressed as:

[0096] v i (t+1)=w·v i (t)+m1·n1·(L i (t)-f i (t))+m2·n2·(G i (t)-f i (t)) (23)

[0097] Among them, the inertia weight w = 0.7, the learning factor m1 = 0.5, m2 = 0.5, and the value range of random numbers n1 and n2 is [0, 1];

[0098] ③ Update the particle frequency combination according to the frequency combination of the particle at time t and the particle speed at the next moment. The particle frequency value x of the i-th boost pump at time t+1 is iThe update rule for (t+1) is expressed as:

[0099] x i (t+1)=x i (t)+v i (t+1) (24)

[0100] Among them, the frequency combination vector of the particle at time t+1 is x(t+1)=[x1(t+1),x2(t+1),…,x i (t+1)];

[0101] ④ According to formula (9), calculate the fitness function corresponding to the particle frequency combination at the current time t. When the optimal frequency value L of the i-th boost pump at time t+1 is i The update rule for (t+1) is expressed as:

[0102]

[0103] ⑤ When the i-th boost pump frequency value G in the optimal frequency combination at time t+1 i The update rule for (t+1) is expressed as:

[0104] G i (t+1)=argmin(F(L i (t))) (26)

[0105] ⑥If t<Z, the number of iterations Z=100, then t=t+1, return to step ② to continue training, if t=Z, or L i (t) and G i If the value of (t) does not change in 20 consecutive updates, the training is completed;

[0106] (4) Optimization and control of sewage treatment lift pumps

[0107] The mathematical relationship between pump frequency and pump station operation related parameters is described based on the operation data of the sewage treatment lift pump. The pump frequency and the main pipe flow of the pump station are used as the model input. The pump frequency intelligent optimization decision model based on pressure maintenance constraint is established from the three parts of pump pressure constraint, flow model and energy consumption model using formulas (14)-(21), and the fitness function is defined. The decision model is trained using the particle swarm algorithm according to formulas (23)-(26) to minimize the fitness function, ensuring that the pump station can reduce energy consumption as much as possible while satisfying the pressure constraint and the main pipe flow constraint. Finally, the optimal frequency combination vector x′(t) is output to achieve the optimization control of the sewage treatment lift pump.

Claims

1. A method for intelligent optimization operation decision of sewage treatment lifting pump based on pressure maintenance constraint, characterized in that The following steps are involved: (1) Collecting sewage treatment lift pump operation data Taking the operation process of the pump station as the research object, the main pipe flow of the pump station, the pump frequency, the total pressure of the pump station, the head of the pump unit and the total power consumption of the pump station are selected as the relevant variables for pump station regulation. The main pipe flow of the pump station at time t is collected as the vector q(t)=[q1(t),q2(t),…,q k (t)] T , T represents matrix transpose, the number of arrays collected k = 1, 2...50, the value range of the total flow is [1000 cubic meters / hour, 8000 cubic meters / hour], and the boost pump frequency at time t is the matrix F(t) = [f1(t), f2(t), ..., f i (t)], the number of lift pumps i = 1, 2, 3, 4, the frequency of the i-th lift pump is the vector f i (t) = [f 1,i (t),f 2,i (t),…,f k,i (t)] T , where f k,i (t) is the kth frequency value of the i-th boost pump, and the frequency value range is [34 Hz, 45 Hz]. The total pressure of the pumping station at time t is the vector s(t) = [s1(t), s2(t),…, s k (t)] T The value range of total pressure is [0 N / m2, 200 N / m2], and the operating head of the lifting pump unit at time t is the matrix H(t) = [h1(t), h2(t), …, h i (t)] T , the lift of the i-th lift pump is vector h i (t) = [h 1,i (t),h 2,i (t),…,h k,i (t)] T , where h k,i (t) is the kth lift value of the i-th lift pump, and the lift value range is [5 meters, 45 meters]. The total power consumption of the pumping station at time t is the vector p(t) = [p1(t), p2(t), …, p k (t)] T , the power value is in kilowatts; (2) Establishing an intelligent optimization decision model for pump frequency based on pressure maintenance constraints The intelligent optimization decision model for pump frequency based on pressure maintenance constraint includes pump pressure constraint, flow model and energy consumption model, which aims to keep the total pressure and total pipe flow of the pump station system stable within the target range and minimize the overall energy consumption of the pump station. Specifically, 1) pump pressure constraint Based on the relationship between the total pressure of the pumping station and the pumping frequency, flow rate and head, ensure that the total pressure of the pumping station is maintained in a stable range [144 Newtons / square meter, 146 Newtons / square meter]; According to the collected pump station main flow value q k (t) and the pump frequency vector value f k,i (t), the approximate linear relationship between the main flow of the pump station and the booster pump frequency at time t is expressed as: Among them, the constant r1 related to flow rate and pump frequency is 60; According to the collected operating head of the lifting pump unit and the main pipe flow of the pump station, the relationship between the head and the main pipe flow at time t is expressed as: Among them, the maximum head of each lifting pump is h max =45 meters, maximum flow rate of the main pipe of the pump station q max = 8000 cubic meters / hour. The mathematical relationship between the total pressure and flow rate of the pump station is expressed using the additional pressure loss term. At time t, the pipeline friction loss model in fluid mechanics can be approximately expressed as: l(t)=r2·(q k (t)) 2 (3) Among them, the constant r2=0.0005 related to pipeline characteristics and fluid characteristics, etc., based on the mathematical relationship between the total pressure of the pump station and the pump frequency, flow rate and head, the total pressure value of the pump station s total (t) is expressed as: s total (t)=1000·9.180·h i (t)+l(t) (4) In order to ensure that the total pressure of the pumping station is within the target range [144 Newtons / square meter, 146 Newtons / square meter], a pressure constraint penalty function is defined. The pressure constraint penalty function at time t is expressed as: 2) Traffic model The flow model is used to ensure that the water supply flow of the pump station remains within a reasonable range and meets the actual production water supply demand. k (t) Constraints: q min ≤q k (t)≤q max (6) Among them, the minimum flow rate of the pump station main pipe q min = 1000 cubic meters / hour, define the flow penalty function, the penalty function of the flow at time t is expressed as: 3) Energy consumption model A mathematical model is established for the power consumption of each lifting pump and the corresponding pump frequency. According to the collected pump frequency and total energy consumption, the mathematical relationship between the total power consumption and pump frequency at time t is expressed as: Among them, the constant r3 related to the characteristics of the lift pump (such as head, pump efficiency, etc.) is 0.05; 4) Set the fitness function Considering the pressure stability, flow rationality and energy efficiency minimization goals, the fitness function is designed as the objective function. The fitness function at time t is expressed as: F(x i (t))=a·P(s total (t))+b·Q(q k (t))+c·E total (t) (9) Among them, x i (t) is the operating parameter of the i-th lift pump, and the frequency combination vector is x(t) = [x1(t), x2(t),…, x i (t)], the weight coefficient of the pressure stability target is a = 0.4, the weight coefficient of the flow rationality target is b = 0.2, and the weight coefficient of the energy efficiency minimization target is c = 0.4; (3) Design a boost pump frequency setting optimization strategy based on particle swarm algorithm In the boost pump frequency setting optimization strategy based on particle swarm algorithm, the particle optimization process is to gradually approach the optimal solution by updating the frequency combination and speed of the particles; ① The frequency combination vector of the particle at time t is x(t), and the velocity of the particle is vector v(t) = [v1(t), v2(t), …, v i (t)], v(t) determines the direction and step of frequency adjustment of each boost pump. Set the current time t = 1, and initialize the frequency combination vector x(1) = [x1(1), x2(1), …, x i (1)], where x1(1) = f 1,1 (t), x2(1) = f 1,2 (t), x i (1) = f 1,i (t), initialize the velocity vector v(1) of the particle to each element value v0, and initialize the optimal frequency value L of the i-th boost pump i (1) = 34, initialize the frequency value G of the i-th boost pump in the optimal frequency combination i (1) = 34; ②According to the initial value L i (1) and G i (1) Adjust the particle speed. The speed value v of the particle of the i-th lift pump at time t+1 i The update rule for (t+1) is expressed as: v i (t+1)=w·v i (t)+m1·n1·(L i (t)-f i (t))+m2·n2·(G i (t)-f i (t)) (10) Among them, the inertia weight w = 0.7, the learning factor m1 = 0.5, m2 = 0.5, and the value range of random numbers n1 and n2 is [0, 1]; ③ Update the particle frequency combination according to the frequency combination of the particle at time t and the particle speed at the next moment. The particle frequency value x of the i-th boost pump at time t+1 is i The update rule for (t+1) is expressed as: x i (t+1)=x i (t)+v i (t+1) (11) Among them, the frequency combination vector of the particle at time t+1 is x(t+1)=[x1(t+1),x2(t+1),…,x i (t+1)]; ④ According to formula (9), calculate the fitness function corresponding to the particle frequency combination at the current time t. When the optimal frequency value L of the i-th boost pump at time t+1 is i The update rule for (t+1) is expressed as: ⑤ When the optimal frequency combination at time t+1, the frequency value G of the i-th boost pump i The update rule for (t+1) is expressed as: G i (t+1)=argmin(F(L i (t))) (13) ⑥If t<Z, the number of iterations Z=100, then t=t+1, return to step ② to continue training, if t=Z, or L i (t) and G i If the value of (t) does not change in 20 consecutive updates, the training is completed; (4) Optimization and control of sewage treatment lift pumps The mathematical relationship between pump frequency and pump station operation related parameters is described based on the operation data of sewage treatment lift pump. The pump frequency and pump station main flow are used as model inputs. Formulas (1)-(8) are used to establish a pump frequency intelligent optimization decision model based on pressure maintenance constraints from three parts: pump pressure constraint, flow model and energy consumption model. The fitness function is defined and the decision model is trained using the particle swarm algorithm according to formulas (10)-(13). The fitness function is minimized and the optimal frequency combination vector x′(t) is output to achieve optimal control of sewage treatment lift pump.

Images