Pump station outlet pressure regulation and control method for water supply network

Abstract

The invention belongs to the technical field of water supply network pressure regulation and control, and particularly relates to a pump station outlet pressure regulation and control method for a water supply network. Compared with constant pressure supply of a pump outlet in a mainstream method, the method has the advantages that various water supply time periods can be effectively distinguished through division of different water supply time periods, and a corresponding pump station outlet pressure value is provided for each time period. In other water supply periods except for the peak period, the pressure of the water outlet of the pump station more accurately meets the actual requirement, and therefore energy waste is effectively reduced. Compared with a method for keeping the pressure value of the most unfavorable point constant, since the outlet pressure of the pump station is accurately predicted by using historical data, the real-time pressure data of the most unfavorable point does not need to be collected and transmitted to the booster pump station in real time to guide the operation of the pump station, and the feedback of data such as pressure measurement information and the like has no requirement for timeliness; and frequent regulation and control are not needed. According to the method, the situation of energy waste of the pump station can be effectively reduced on the basis of reasonably controlling the cost.

Description

Technical Field

[0001] The invention belongs to the technical field of water supply network pressure regulation, and in particular relates to a method for regulating the outlet pressure of a pump station for a water supply network. Background Art

[0002] The water supply system is an important part of urban infrastructure, and the water pump is the key in the water supply system. The energy consumed by its operation accounts for the majority of the entire system. However, most pumping stations in my country still have the problem of energy waste. According to relevant industry statistics, the operating efficiency of urban water pump units is low, currently generally between 50% and 75%. As the main energy consumption link in the urban water supply system, improving its operating efficiency has an important impact on achieving the industry's energy-saving goals. Therefore, studying the optimization operation strategy of water pumps, improving the operating efficiency of pump stations and reducing operating energy consumption will help achieve the city's green and low-carbon development goals.

[0003] The current mainstream method for managing the pump group of a water supply pump station is to maintain a constant supply pressure at the pump outlet. That is, the outlet pressure of the water supply pump station is always a constant value, and the basis for the value of this constant value is that the most unfavorable point can also use water stably during the peak water consumption period. However, there is a big problem with this management method, that is, most of the time, the pressure at the most unfavorable point is excessively increased, which will cause a large waste of energy.

[0004] In order to solve this problem, some technicians in this field have proposed to keep the pressure value at the most unfavorable point constant, and then adjust the power of the water pump group of the water pump station in reverse in combination with the real-time water flow of the pipe network users. In this way, since the most unfavorable point at each time point is just able to use water stably, the energy waste of the pump station can be effectively avoided. However, to adopt such a technical solution, it is necessary to install a real-time pressure transmission device at the most unfavorable point, and transmit the collected pressure data to the booster pump station in real time to guide the operation of the pump station (in extreme cases, frequent regulation is required), which has high requirements for the real-time operation management capability of the pump station. In addition, it takes time for the pressure at the beginning of the pump station to conduct the pressure increase and decrease at the most unfavorable point, and the feedback of the pressure measurement information has a certain lag, which leads to the weak timeliness of the regulation strategy. Therefore, the cost of this technical solution is very high, and the real-time performance and operability are not strong, and the actual use is not ideal.

[0005] In summary, how to effectively reduce the energy waste of pumping stations on the basis of reasonable cost control has become a problem that needs to be solved urgently. Summary of the invention

[0006] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method for regulating the outlet pressure of a pump station for a water supply network, which can effectively reduce the energy waste of the pump station on the basis of reasonably controlling the cost.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0008] A method for regulating the outlet pressure of a pump station for a water supply network comprises the following steps:

[0009] S1. Construct a pump station outlet pressure prediction model; used to predict the pump station outlet pressure in the next N hours of time series based on the historical water supply flow sequence data, the pressure sequence data of the pressure measuring point and the water supply pressure sequence data required by the pump station;

[0010] S2. Obtain historical training data and perform preprocessing to obtain model training data; the historical training data includes historical water supply flow sequence data, pressure measurement point pressure sequence data, and water supply pressure sequence data required by the pump station;

[0011] S3, using the model training data obtained in S2, train the pump station outlet pressure prediction model constructed in S1;

[0012] S4, predicting the pump station outlet pressure of the time series of the next N hours through the trained pump station outlet pressure prediction model;

[0013] S5. Based on the prediction result of S4, the pump station outlet pressure is divided into multiple water supply periods; when dividing the water supply periods, according to the fluctuation of the predicted pump station outlet pressure, the maximum continuous time interval in which the fluctuation value is less than the preset value is taken as a water supply period;

[0014] S6. For each water supply period, the maximum pump station outlet pressure value in the period is used as the constant outlet pressure of the water supply period; the constant outlet pressures of each water supply period are integrated as a control scheme;

[0015] S7. According to the control scheme of S6, actual pump station outlet pressure control is performed to provide corresponding constant outlet pressures in different water supply periods.

[0016] Compared with the prior art, the present invention has the following beneficial effects:

[0017] 1. In this method, a pump station outlet pressure prediction model is constructed to predict the time series pump station outlet pressure in the next N hours. In actual use, based on the prediction results of the model, the pump station outlet pressure is divided into multiple different water supply periods, and according to the predicted fluctuation of the pump station outlet pressure, the maximum continuous time interval with a fluctuation value less than the preset value is taken as a water supply period. Afterwards, for each water supply period, the maximum pump station outlet pressure value in the period is taken as the constant outlet pressure of the water supply time, and the constant outlet pressure of each water supply time is integrated as a control scheme. Then, actual control is performed according to the control scheme.

[0018] Compared with the mainstream method of constant supply pressure at the pump outlet, this method can effectively distinguish various time periods of water use (such as peak time period, sub-peak time period, regular time period, valley time period, etc.) by dividing the water supply period, and provide corresponding pump station outlet pressure values ​​for each period. In other water supply periods except the peak time period, the water outlet pressure of the pump station is more in line with actual needs, thereby effectively reducing energy waste. Compared with keeping the pressure value of the most unfavorable point constant, since historical data (such as data from the past 10 days) is used to predict the pump station outlet pressure, there is no need to collect real-time pressure data at the most unfavorable point and transmit it to the booster pump station in real time to guide the operation of the pump station. There is no timeliness requirement for the feedback of data such as pressure measurement information, and there is no need for very frequent regulation. It is highly operational and has a low implementation cost.

[0019] 2. The pump station outlet pressure prediction model constructed by this method is trained based on the historical water supply flow sequence data, the pressure data sequence of the pressure measuring point, and the water supply pressure sequence data required by the pump station. The model can more accurately predict the pump station outlet pressure in the next N hours. This prediction ability helps to understand the changing trend of the pump station outlet pressure in advance and provide a reliable basis for subsequent regulation.

[0020] 3. Based on the prediction results, this method divides the pump station outlet pressure into multiple different water supply periods, and according to the predicted fluctuation of the pump station outlet pressure, the maximum continuous time interval with a fluctuation value less than the preset value is used as a water supply period. This division method can arrange the water supply period more reasonably and reduce the adverse effects on the water supply system caused by excessive pressure fluctuations.

[0021] In summary, this method can effectively reduce the energy waste of pumping stations on the basis of reasonable cost control.

[0022] Preferably, in S7, when a corresponding constant outlet pressure is provided in a water supply period, an optimal power control scheme of the water pump unit is solved by a preset power optimization model, and the water pump unit is controlled according to the obtained optimal power control scheme.

[0023] This setting, through the power optimization model, can accurately calculate the optimal power required by the pump unit under different water supply periods and constant outlet pressure. This helps to avoid the pump unit running at excessive power when it is not necessary, thereby further reducing energy consumption and operating costs.

[0024] In addition, long-term operation at too high a power will cause the pump unit to wear more severely and shorten the service life of the equipment. Through power optimization control, it can ensure that the pump unit operates within a reasonable power range, reduce the mechanical stress and wear caused by overload operation, and thus extend the service life of the equipment.

[0025] Preferably, the objective function of the power optimization model is:

[0026]

[0027] In the formula, c is the electricity price during the period; n is the number of speed regulating pumps; ω i To control the opening and closing coefficient of the i-th pump, the value is 0 or 1; Q i is the output flow of the i-th pump; S i is the speed ratio of the i-th pump; d 0i ,d 1i ,d 2i is the power fitting coefficient of the i-th pump.

[0028] Such a setting can ensure the optimal power control solution and reduce energy consumption as much as possible.

[0029] Preferably, the constraints of the power optimization model include single pump flow constraint, water pump head constraint, single pump shaft power constraint, number of units in operation constraint, and speed ratio constraint.

[0030] Such a setting can ensure the effectiveness of the optimal power control scheme.

[0031] Preferably, the single pump flow constraint is:

[0032] Q min <Q i <Q max ;

[0033] In the formula, Q i is the flow rate of the i-th pump in the pump station; Q min , Q max are the minimum flow and maximum flow of the i-th water pump respectively;

[0034] The pump head constraint is:

[0035] H min <H i <H max ;

[0036] In the formula, H i is the head of the i-th water pump; H min , R max are the minimum head and maximum head of the i-th water pump respectively.

[0037] Preferably, the single pump shaft power constraint is:

[0038] P min <P i <P max ;

[0039] Where Pi is the shaft power of the i-th water pump; P min , P max are the minimum and maximum shaft power of the ith pump in the pump station, respectively;

[0040] The number of machines in operation is constrained as follows:

[0041] 0<n<n max ;

[0042] In the formula, n is the number of pumps in operation in the pump station, n max The preset maximum number of powered-on units.

[0043] Preferably, the speed ratio constraint is:

[0044] S i ∈[S imin , 1];

[0045] Where S i is the speed ratio of the i-th pump; S imin is the minimum speed ratio of the i-th pump.

[0046] Preferably, in S7, when solving the optimal power control scheme of the water pump unit through the power optimization model, the solution is performed through an optimization method based on a hybrid improved cuckoo algorithm; the workflow of the optimization method based on the hybrid improved cuckoo algorithm includes:

[0047] Step 1: Randomly generate N bird nests as the initial population, record the initial optimal bird nest position, and perform parameter initialization settings;

[0048] Step 2: Use the Levy flight strategy to update the positions of other bird nests except the current best bird nest position, calculate the fitness value of each bird nest after the update, and then use the bird nest position with the largest fitness value among all the bird nests as the new best bird nest position;

[0049] Step 3: Use the update strategy of the GWO algorithm to update the location of each bird's nest again;

[0050] Step 4: For other bird nest locations other than the optimal solution of the GWO algorithm, random updates are performed in combination with the preset discovery probability Pa; wherein the random update process includes: generating a random number r for a bird nest location to be randomly updated, and comparing r with the discovery probability Pa; if r>Pa, updating the bird nest location by random walking; otherwise, the location of the bird nest remains unchanged;

[0051] Step 5: Calculate the fitness values ​​of all bird nest positions again, and record the bird nest position corresponding to the best fitness value as the optimal solution of the current round;

[0052] Step 6: Determine whether the termination condition is met. If so, output the optimal solution of the current round as the final result; if not, return to step 2 and perform the next round of iteration.

[0053] Such a setting,1. The hybrid improved cuckoo algorithm combines the randomness and long step length characteristics of the Levy flight strategy, which helps the algorithm to quickly explore and jump out of the local optimal solution in the search space, and improves the global search ability. The update strategy of the GWO algorithm is introduced, and the social level and hunting behavior of the gray wolf group are used to guide the search direction, which enhances the local search ability of the algorithm and enables the algorithm to more accurately approach the global optimal solution.

[0054] 2. The Levy flight strategy enables the algorithm to explore the solution space extensively in the early stage of the search and discover potential high-quality solution areas. The introduction of the GWO algorithm strengthens the detailed search for high-quality solution areas in the later stage of the search and improves the accuracy of the solution. This balance between exploration and utilization makes the algorithm more efficient and stable in solving complex optimization problems.

[0055] 3. Random updates are performed using the preset discovery probability Pa, which increases the randomness and diversity of the algorithm and helps prevent the algorithm from falling into premature convergence. The random walk method updates the nest position, allowing the algorithm to maintain a certain degree of flexibility during the search process and adapt to different search environments and problem characteristics.

[0056] 4. The optimal power control problem of a water pump unit often involves multiple variables and complex constraints, and traditional optimization methods are difficult to solve effectively. The hybrid improved cuckoo algorithm combined with the GWO algorithm can effectively deal with such complex optimization problems and find high-quality solutions through its powerful global search and local search capabilities.

[0057] In summary, through this method, the effectiveness of the optimal power control solution finally obtained can be guaranteed.

[0058] Preferably, in step 1, the initialization parameters include the abandonment probability Pa, the maximum number of iterations Tmax, and the parameters of the GWO algorithm;

[0059] In step 6, the termination condition is that the number of iterations reaches the maximum number of iterations Tmax.

[0060] Preferably, the pressure value in the power optimization model calculation optimization process solution meets the following conditions:

[0061] |H i -H i |≤0.2;

[0062] In the formula, H i is the constant outlet pressure of the ith water supply period in the control scheme; H jCalculate the pressure value in the optimization process solution for the corresponding water supply period.

[0063] In this way, the effectiveness of the optimal power control solution finally obtained can be guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to make the purpose, technical solution and advantages of the invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which:

[0065] Figure 1 A flowchart of the method is shown in FIG.

[0066] Figure 2 This is the flow chart of the hybrid improved cuckoo algorithm in this method;

[0067] Figure 3 Schematic diagram of the convergence of the optimization calculation of the Schwefel test function f1(x) in Example 2;

[0068] Figure 4 This is a schematic diagram comparing the optimization results of the f1(x) test function in Example 2;

[0069] Figure 5 This is a schematic diagram of the convergence of the optimization calculation of the Schwefe1 test function f2(x) in Example 2;

[0070] Figure 6 This is a schematic diagram of the convergence of the optimization calculation of the test function f2(x) in Example 2;

[0071] Figure 7 This is a schematic diagram of the analysis of the pressure value required for the pump station outlet in Example 3;

[0072] Figure 8 This is a statistical chart of pipe leakage and burst in the past three years in Example 3;

[0073] Fig. 9 This is a model diagram of a booster pump station in a town in Province S in Example 3;

[0074] Fig.10 This is a schematic diagram of the prediction of the water pressure at the highest daily water consumption in summer in Example 3;

[0075] Fig.11 This is a schematic diagram of water pressure prediction on the day with the highest water consumption in winter in Example 3;

[0076] Fig.12 The voltage regulation control situation of the highest day in summer in Example 3;

[0077] Fig.13 This is the voltage regulation control situation of the highest day in winter in the third embodiment;

[0078] Fig.14 This is a schematic diagram of the iterative results of the pump station optimization operation when H=48.03-51.03m in Example 3;

[0079] Figure 1 5 is a schematic diagram of the iterative results of the pump station optimization operation when H=51.03-54.03m in Example 3;

[0080] Figure 1 6 is a schematic diagram of the iterative results of the pump station optimization operation when H=54.03-57.03m in Example 3;

[0081] Figure 1 7 is a schematic diagram of the iterative results of the pump station optimization operation when H=57.03-60.03m in Example 3;

[0082] Fig.18 This is the relationship diagram between efficiency, flow rate and head in winter water use in Example 3;

[0083] Fig.19 This is the relationship diagram between efficiency, flow rate and head in summer water use in Example 3;

[0084] Fig. 20 This is a schematic diagram of the pressure change analysis at the most unfavorable point in Example 3;

[0085] Fig.21 This is a schematic diagram of the pressure change analysis of the No. 1 overpressure point in Example 3;

[0086] Fig. 22 This is a schematic diagram of the pressure change analysis of overpressure point No. 2 in Example 3. DETAILED DESCRIPTION

[0087] The following is a further detailed description through specific implementation methods:

[0088] Embodiment 1

[0089] like Figure 1 As shown, this embodiment discloses a method for regulating the outlet pressure of a pump station for a water supply network, comprising the following steps:

[0090] S1. Construct a pump station outlet pressure prediction model; it is used to predict the pump station outlet pressure in the next N hours of time series based on the historical water supply flow sequence data, the pressure sequence data of the pressure measuring point and the water supply pressure sequence data required by the pump station.

[0091] In the specific implementation, the pump station outlet pressure prediction model is constructed based on the LSTM neural network model. In the established neural network prediction model, the historical water supply flow sequence data, the pressure sequence data of the pressure measuring point and the water supply pressure sequence data required by the pump station are used as the input layer data, and the predicted water outlet pressure value is used as the output layer. The hidden layer uses black box operation to simulate and simulate the complex relationship between flow and pressure. The data relationship can be expressed as follows:

[0092] P pre =f LSTM (H i , P pump , Q);

[0093] Where P pre To predict the water supply pressure of the pump station m; f LSTM is the LSTM model; H i is the historical pressure value of the pressure measuring point m; P pump m is the water supply pressure value required for the operation history of the pumping station; Q is the historical water supply flow of the pumping station.

[0094] S2. Obtain historical training data and perform preprocessing to obtain model training data; the historical training data includes historical water supply flow sequence data, pressure measurement point pressure sequence data, and water supply pressure sequence data required by the pump station.

[0095] Combined with the specific actual situation of the background city, multiple historical operation data are used as input variables, and the outlet pressure of the pump station is set as the output variable. These different input variables can be summarized into two aspects. One is the water supply flow and pump station outlet pressure data of the previous eight days. The purpose of this type of data is to enable the model to learn the periodic laws of the pipe network operation; the other is the detection point pressure, which is to allow the model to perceive and verify the current predicted operation of the pipe network.

[0096] In the specific implementation, the historical data of pump station flow rate is selected from ten days of pump station flow rate data, of which the water volume data of the first eight days is used as the training set of the model, and the water volume data of the last two days is used as the test set. The historical data of water supply pressure required by the pump station is aimed at ensuring the constant pressure value of water supply at the most unfavorable point. The historical water supply flow data is used to calculate the pipeline water loss through hydraulic calculation, and the actual required pressure value of the pump station outlet is calculated as the variable input. The historical data of pressure at the pressure measuring point is selected. The pressure data consistent with the input water volume data time period is selected, and representative pressure monitoring points in the water supply network are selected. Due to the limited number of actual pressure detection points, the most unfavorable point of the network system (underpressure complaint point) and the low-lying overpressure point are used as representative detection points to provide pressure detection data as variable input.

[0097] Among them, preprocessing includes normalization. The factors affecting the outlet pressure of the pump station include the outlet flow rate and the pressure at the pressure measuring point, which are the input values ​​of the prediction model. Their respective dimensions and orders of magnitude are inconsistent. Similar to water volume prediction, these data need to be normalized.

[0098] S3, using the model training data obtained in S2, train the pump station outlet pressure prediction model constructed in S1;

[0099] S4. Predict the pump station outlet pressure in the future N hours of time series through the trained pump station outlet pressure prediction model. In the specific implementation, N is 24. In this way, the pump station outlet pressure is predicted for 24 hours a day, which can ensure the effectiveness of the adjustment frequency while ensuring the immediate effectiveness of the adjustment.

[0100] S5. Based on the prediction result of S4, the pump station outlet pressure is divided into multiple water supply periods; when dividing the water supply periods, according to the fluctuation of the predicted pump station outlet pressure, the maximum continuous time interval with a fluctuation value less than a preset value is taken as a water supply period. Through such a setting, various time periods for water use (such as peak time period, sub-peak time period, regular time period, valley time period, etc.) can be effectively distinguished.

[0101] S6. For each water supply period, the maximum pump station outlet pressure value in the period is used as the constant outlet pressure of the water supply period; the constant outlet pressures of each water supply period are integrated as a control scheme;

[0102] S7. According to the control scheme of S6, the actual pump station outlet pressure is controlled, and corresponding constant outlet pressure is provided in each water supply period, so as to meet the actual pressure required by the water supply network in the corresponding time period.

[0103] This method constructs a pump station outlet pressure prediction model to predict the pump station outlet pressure in the next N hours of time series. In actual use, based on the prediction results of the model, the pump station outlet pressure is divided into multiple water supply periods, and according to the predicted fluctuation of the pump station outlet pressure, the maximum continuous time interval with a fluctuation value less than the preset value is taken as a water supply period. Afterwards, for each water supply period, the maximum pump station outlet pressure value in the period is taken as the constant outlet pressure of the water supply time, and the constant outlet pressure of each water supply time is integrated as a control scheme. Then the actual control is carried out according to the control scheme. Compared with the mainstream method of constant supply pressure at the pump outlet, this method can effectively distinguish various time periods of water use (such as peak time period, sub-peak time period, regular time period, valley time period, etc.) through the division of water supply periods, and provide corresponding pump station outlet pressure values ​​for each period. In other water supply periods except the peak time period, the pump station outlet pressure is more in line with actual needs, thereby effectively reducing energy waste. Compared with keeping the pressure value of the most unfavorable point constant, since the historical data (such as the data of the past 10 days) is used to predict the outlet pressure of the pump station, there is no need to collect the real-time pressure data of the most unfavorable point and transmit it to the booster pump station in real time to guide the operation of the pump station. There is no timeliness requirement for the feedback of data such as pressure measurement information, and there is no need for very frequent regulation. It is highly operational and has a low implementation cost. This method can effectively reduce the energy waste of the pump station on the basis of reasonable cost control.

[0104] In addition, the pump station outlet pressure prediction model constructed by this method is trained based on historical water supply flow sequence data, pressure measurement point pressure sequence data, and water supply pressure sequence data required by the pump station. The model can more accurately predict the time series pump station outlet pressure in the next N hours. This prediction capability helps to understand the changing trend of the pump station outlet pressure in advance and provide a reliable basis for subsequent regulation. In addition, based on the prediction results, this method divides the pump station outlet pressure into multiple water supply periods, and according to the predicted fluctuation of the pump station outlet pressure, the maximum continuous time interval with a fluctuation value less than the preset value is used as a water supply period. This division method can arrange the water supply period more reasonably and reduce the adverse effects on the water supply system caused by excessive pressure fluctuations.

[0105] During specific implementation, in S7, when a water supply period provides a corresponding constant outlet pressure, the optimal power control scheme of the water pump unit is solved by a preset power optimization model, and the water pump unit is controlled according to the obtained optimal power control scheme. In this way, the power optimization model can accurately calculate the optimal power required by the water pump unit under different water supply periods and constant outlet pressures. This helps to avoid the water pump unit from running at too high a power when it is not necessary, thereby further reducing energy consumption and operating costs. In addition, running at too high a power for a long time will cause the water pump unit to wear more and shorten the service life of the equipment. Through power optimization control, it can ensure that the water pump unit operates within a reasonable power range, reduce the mechanical stress and wear caused by overload operation, and thus extend the service life of the equipment.

[0106] Among them, the objective function of the power optimization model is:

[0107]

[0108] In the formula, c is the electricity price during the period; n is the number of speed regulating pumps; ω i To control the opening and closing coefficient of the i-th pump, the value is 0 or 1; Q i is the output flow of the i-th pump; S i is the speed ratio of the i-th pump; d 0i ,d 1i ,d 2i is the power fitting coefficient of the i-th pump.

[0109] The constraints of the power optimization model include single pump flow constraint, pump head constraint, single pump shaft power constraint, number of units started constraint, and speed ratio constraint. In this way, the effectiveness of the optimal power control scheme can be guaranteed.

[0110] Single pump flow restriction

[0111] In combination with the operating characteristics of the water pump in the high efficiency section, the water supply flow rate of the water pump is limited within a certain range during the calculation process.

[0112] Q min <Q i <Q max ;

[0113] In the formula, Q i is the flow rate of the i-th pump in the pump station; Q min , Q max are the minimum flow rate and maximum flow rate of the i-th water pump respectively.

[0114] Pump head constraints

[0115] If the water pump is kept running in the high efficiency range, the water supply head of the water pump will fluctuate within a certain range.

[0116] H min <H i <H max ;

[0117] In the formula, H i is the head of the i-th water pump; R min , R max are the minimum head and maximum head of the i-th water pump respectively.

[0118] Single pump shaft power constraints

[0119] Each water pump has two states: on and off. When the water pump is on, and it is necessary to ensure that the water pump operates in the high efficiency section, the shaft power of the water pump will vary within a certain range.

[0120] P min <P i <P max ;

[0121] Where P i is the shaft power of the i-th water pump; P min , P max are the minimum and maximum shaft power of the i-th pump in the pump station respectively; the number of units in operation is constrained

[0122] There are three speed-regulating pumps of the same model in the pump station. Usually, one pump is used to supply water during the low-peak period of water consumption, and two pumps are used in parallel to supply water during the peak period of water consumption. This part constrains the number of pumps that are turned on, controls the actual number of pumps that are turned on to within two, and at the same time seeks the optimal solution for the number of pumps that are turned on in each period.

[0123] 0<n<n max ;

[0124] In the formula, n is the number of pumps in operation in the pump station, n max is the preset maximum number of powered-on machines. In this embodiment, n max =2.

[0125] Speed ​​ratio constraint

[0126] In actual operation, the operation of the water pump must be kept stable and cavitation must be prevented. To ensure safety, the water pump is usually not allowed to overload at a speed exceeding the rated speed. In addition, when the speed ratio is less than a certain value, the operating efficiency of the water pump will drop sharply. Therefore, the speed ratio of the water pump is generally controlled to be kept within an appropriate range.

[0127] S i ∈[S imin , 1];

[0128] Where S i is the speed ratio of the i-th pump; S iminis the minimum speed ratio of the i-th pump, and its value needs to be determined according to the specific power and efficiency of the speed regulating pump.

[0129] In specific implementation, when solving the optimal power control scheme of the water pump unit through the power optimization model, the solution is obtained through the optimization method based on the hybrid improved cuckoo algorithm; Figure 2 As shown in FIG. 1 , the workflow of the optimization method based on the hybrid improved cuckoo algorithm includes:

[0130] Step 1: Randomly generate N bird nests as the initial population, record the initial optimal bird nest position, and perform parameter initialization settings, where the initialization parameters include the abandonment probability Pa, the maximum number of iterations Tmax, and the parameters of the GWO algorithm.

[0131] Step 2: Use the Levy flight strategy to update the positions of other bird nests except the current best bird nest position, calculate the fitness value of each bird nest after the update, and then use the bird nest position with the largest fitness value among all the bird nests as the new best bird nest position;

[0132] Step 3: Use the update strategy of the GWO algorithm to update the location of each bird's nest again;

[0133] Step 4: For other bird nest locations other than the optimal solution of the GWO algorithm, random updates are performed in combination with the preset discovery probability Pa; wherein the random update process includes: generating a random number r for a bird nest location to be randomly updated, and comparing r with the discovery probability Pa; if r>Pa, updating the bird nest location by random walking; otherwise, the location of the bird nest remains unchanged;

[0134] Step 5: Calculate the fitness values ​​of all bird nest positions again, and record the bird nest position corresponding to the best fitness value as the optimal solution of the current round;

[0135] Step 6: Determine whether the termination condition is met. If so, output the optimal solution of the current round as the final result; if not, return to step 2 and perform the next round of iteration. The termination condition is that the number of iterations reaches the maximum number of iterations Tmax.

[0136] In specific implementation, the pressure value in the solution of the power optimization model calculation optimization process meets the following conditions:

[0137] |H i -H i |≤0.2;

[0138] In the formula, H i is the constant outlet pressure of the ith water supply period in the control scheme; H j Calculate the pressure value in the optimization process solution for the corresponding water supply period.

[0139] In this way, the effectiveness of the optimal power control solution finally obtained can be guaranteed.

[0140] The main advantage of the cuckoo algorithm is that it uses the Levy flight mechanism to generate motion trajectories. This highly random flight method allows the search process to quickly explore the solution space, thereby increasing the diversity of the population. However, the standard cuckoo algorithm relies entirely on Levy flights to update the nest position. This mechanism has certain deficiencies in local search capabilities, resulting in low optimization accuracy of the algorithm, weak local optimization capabilities in the later stage, and slow convergence speed.

[0141] The problem with the Gray Wolf Optimization Algorithm is that the alpha wolf in its core wolf pack does not always move toward the global optimal solution, and other wolf pack members tend to move toward the top three solutions of the leader wolf, so it is easy to fall into the local optimal solution, resulting in its relatively weak global search ability. In actual application, the Gray Wolf Algorithm does not perform well in solving complex optimization problems because the search strategy in the original algorithm is relatively simple and it is difficult to effectively find the global optimal solution.

[0142] Through analysis, it can be found that the cuckoo algorithm and the gray wolf optimization algorithm have complementary characteristics in their respective advantages and limitations. The gray wolf optimization algorithm has a strong local search capability, while the Levy flight strategy in the cuckoo algorithm can bring a more diverse population. In order to give full play to the advantages of the two algorithms, the position update strategy in the gray wolf optimization algorithm can be introduced into the cuckoo algorithm. The purpose of this is to reduce the blindness of the Levy flight search, enhance the local search capability of the cuckoo algorithm, and improve the optimization accuracy of the algorithm.

[0143] The hybrid improved cuckoo algorithm (GWO-CS) is designed by combining the characteristics of the two algorithms, aiming to balance the global search and local search capabilities to improve the optimization effect of the algorithm. By combining the advantages of the two algorithms, GWO-CS is able to perform fast searches on a global scale and fine-tune the solution space through the local search mechanism of the CS algorithm in the later stage. This hybrid improvement method enables the algorithm to better optimize the local optimal solution while searching for the global optimal solution, thereby improving the algorithm's optimization performance.

[0144] In the above optimization methods, the hybrid improved cuckoo algorithm combines the randomness and long step length characteristics of the Levy flight strategy, which helps the algorithm to quickly explore and jump out of the local optimal solution in the search space and improve the global search ability. The update strategy of the GWO algorithm is introduced, and the social level and hunting behavior of the gray wolf group are used to guide the search direction, which enhances the local search ability of the algorithm and enables the algorithm to more accurately approach the global optimal solution. The Levy flight strategy enables the algorithm to widely explore the solution space and discover potential high-quality solution areas in the early stage of the search. The introduction of the GWO algorithm strengthens the fine search of high-quality solution areas in the later stage of the search and improves the accuracy of the solution. This balance between exploration and utilization makes the algorithm more efficient and stable when solving complex optimization problems. In addition, random updates are performed through the preset discovery probability Pa, which increases the randomness and diversity of the algorithm and helps to avoid the algorithm from falling into premature convergence. The random walk method updates the nest position, so that the algorithm can maintain a certain flexibility during the search process and adapt to different search environments and problem characteristics. In addition, the optimal power control problem of the pump unit often involves multiple variables and complex constraints, which is difficult to solve effectively by traditional optimization methods. The hybrid improved cuckoo algorithm combined with the GWO algorithm can effectively deal with such complex optimization problems and find high-quality solutions through its powerful global search and local search capabilities. Through this method, the effectiveness of the optimal power control solution obtained can be guaranteed.

[0145] Embodiment 2

[0146] In order to better illustrate the performance of the optimization method of the hybrid improved cuckoo algorithm (GWO-CS algorithm) in this method, the following algorithm performance test description is carried out.

[0147] In order to test the performance of the GWO-CS algorithm, the Schwefel test function was selected for optimization test experiments, and a control group with the original algorithm was set up. The effect of improving the algorithm performance was analyzed through the optimization results to verify the correctness of the improvement strategy.

[0148] The Schwefel function is a typical deception problem. It has a global minimum point, which is far away from another local optimum. Once it falls into the local optimum, it is difficult to jump out. Therefore, this type of function can be used as an ideal test function to test the global optimization performance of the improved algorithm.

[0149] Select two typical functions and name them as f 1 (x) and f 2(x), the function expression and performance test results are as follows:

[0150] f 1 (x) = max{|x i |, 1≤i≤30}, -100≤x i≤100;

[0151] min(f 1 ) = f 1 (0, ..., 0) = 0;

[0152] from Figure 3 It can be seen that compared with the original CS algorithm, GWO-CS has a significant improvement in convergence speed. Figure 4 It can be seen from the comparison that the traditional algorithm falls into the local optimal solution during the optimization process. The improved GWO-CS algorithm has significantly improved its performance in this aspect, and its global optimization results are obviously better than those of the traditional CS algorithm.

[0153] Another Schwefel class test function f 2 The expression of (x) and the test results are as follows:

[0154]

[0155] min(f 2 ) = f 2 (0, ..., 0) = 0;

[0156] Similar to the test result analysis of the previous function, Figure 5 It can be seen that GWO-CS has a significant advantage in the convergence speed of the optimization process, and Figure 6 In the comparison, the traditional CS algorithm also exposes the problem of falling into the local optimal solution, while the globality of the optimal solution of the GWO-CS algorithm is significantly better than that of the traditional CS algorithm.

[0157] Test results analysis

[0158] By comparing the optimization solution results, it can be found that the improved GWO-CS algorithm shows obvious advantages in convergence speed. In contrast, the original CS algorithm is easily disturbed by local optimal points and it is difficult to jump out of the local optimal solution. However, the GWO-CS algorithm can quickly converge to the global optimal point and has higher solution accuracy. When optimizing the test function, it was observed that the GWO-CS algorithm showed obvious advantages over the original CS algorithm in terms of convergence speed and solution accuracy, and it was not easy to fall into the local optimal solution. It showed good optimization effects on both Schwefel test functions and had good robustness. Therefore, it can be considered that the hybrid improved cuckoo algorithm provides a more effective optimization method for the pump station operation optimization problem.

[0159] Embodiment 3

[0160] In order to better illustrate the effect of this method, a specific example is used for illustration.

[0161] Basic Situation

[0162] A certain urban area in S Province covers an area of ​​46 square kilometers and has a registered population of 87,000. Its water supply is transferred and supplied by an intermediate pressure pump station with a water supply scale of 20,000 tons / day. The pump station is supplied with water by three speed-regulating pumps of the same model (two for use and one for backup), with a rated flow of 550m 3 / h, head 65-72m, all controlled by variable frequency.

[0163] Current situation and existing problems

[0164] After field investigation, it was found that the pump stations in the study area had the following operating problems:

[0165] Problem 1: The pump station consumes a lot of energy in operation. The existing extensive pressure regulation method results in a relatively high surplus water pressure during the actual operation of the pump station. This head that exceeds the actual demand not only increases the energy loss of the pump station operation, but also causes the pipe network system to bear excessive pressure.

[0166] Problem 2: The pressure distribution of the water supply network system is unreasonable. The current pump station outlet pressure adopts the traditional control method of ensuring water supply at the most unfavorable point during the highest daily peak. This method lacks flexibility. The network pressure is significantly affected by the fluctuation of user water consumption. The network system pressure is high during the low-peak period of water consumption in the early morning, which aggravates network leakage and is prone to pipe burst accidents.

[0167] ① High energy consumption of pumping station

[0168] The pump station adopts the traditional mode of ensuring the most unfavorable point pressure supply when the highest daily peak is reached in pressure management, and controls the outlet pressure of the pump station to be constant at 81m for water supply. In many periods of time, this value is much greater than the actual required value of the pipe network. The required pressure value of the pump station outlet is estimated by taking the highest water consumption day in the statistical historical data as follows: Figure 7 As shown in the figure, it can be seen that due to the extensive management mode adopted by the booster pump station in pressure control, there is a lot of excess pressure in the pipeline network during the low-peak water consumption period, resulting in a lot of unnecessary energy waste.

[0169] ②Irrational pipe network pressure control

[0170] Since the city where the project is located is a mountainous city with large terrain differences, under the traditional pressure management mode, the extensive constant pump station outlet pressure will face the risk of overpressure and pipe burst in the low-lying high-pressure area when the water consumption increases. Figure 8It is a statistical record of pipe burst repairs from 2021 to 2023. It can be seen that the annual number of pipe burst repairs is more than 100 times, the leakage rate calculated by the difference between production and sales is high, and it has increased with the increase in water consumption in recent years. The areas with high incidence of pipe burst accidents are mostly concentrated in two low-lying areas. It can be considered that the violent pressure fluctuations in the pipeline system are the main reason for the frequent pipe burst accidents in the area, indicating that the extensive pipe network pressure management mode in this area is outdated.

[0171] Applied Research Framework

[0172] The water use data of the highest water use days in summer and winter and the 30th day of each month are used as the background data source for the engineering case study. According to the results of the on-site investigation, the actual working water inlet pressure of the pump station is 0.24MPa, and the outlet water supply pressure of the pump station is constantly controlled at 0.81MPa throughout the day. According to the actual operation data of the pump station in 2023 provided by the staff, the highest daily water use in winter and summer is selected and organized as shown in Table 1:

[0173] Table 1

[0174]

[0175] Take a booster pump station with a water consumption of 20,000 tons / day in a mountain town in S Province as an example. Fig. 9 This is the topological model diagram of the pump station and its pipe network system. The pump outlet point and the most unfavorable point of the pipe network are marked in the model diagram. The water pressure requirement at the most unfavorable point is not less than 26m. The booster pump station uses two variable frequency speed regulating pumps for daily operation, and operates in two modes: the original outlet constant pressure and time-sharing pressure regulation. The EPANET2.0 hydraulic simulation is called through the optimization algorithm to explore the energy-saving effect of the optimization scheme.

[0176] Pump curve parameter determination

[0177] (1) Performance curve fitting determination

[0178] The water delivery capacity of the booster pump station in the study area is 20,000 tons / day. There are three water pumps in the pump station (two in use and one in reserve), all of which are variable frequency controlled speed regulating pumps with a rated flow of 780m 3 / h, the rated head is 0.72MPa, the rated efficiency is 82%, the rated speed is 1480r / min, and the water pump sample data is collected to obtain the discrete points on the QH curve and the QP curve, as shown in Table 2.

[0179] Table 2

[0180]

[0181] The basic performance QH curve, QP and Q-η curve of the water pump at rated speed are fitted by the least squares method using the discrete data in the above table.

[0182] The fitting parameters obtained in the curve fitting process are summarized in Table 3.

[0183] Table 3

[0184]

[0185] Determination of time-based voltage regulation strategy

[0186] (1) Pump station outlet pressure prediction

[0187] The pressure prediction method in this method is used to predict the water outlet pressure on the highest day in summer and the highest day in winter. The water outlet pressure of the pump station within ten days is taken as the original data, the first nine days are used as the training set, and the last day, which is the day with the highest water consumption, is used as the prediction set. The prediction results are as follows: Fig.10 and Fig.11 shown.

[0188] According to the data in the figure, it can be seen that the change trend of the pressure prediction curve is basically consistent with the actual pressure value, whether in summer or winter, and there is no obvious difference in the prediction performance in different seasons. It can be considered that the prediction results of the model can meet the requirements required for the subsequent application of the method.

[0189] (2) Time-based voltage regulation strategy

[0190] Based on the predicted values ​​of the pump station outlet pressure on the highest day in summer and the highest day in winter, the pump station outlet pressure control plan for the day is determined, so that the supply pressure in each period is guaranteed while the outlet pressure in some periods is reduced. Fig.12 and Fig.13 shown.

[0191] The pressure values ​​set for each time period can cover the water pump outlet pressure curve, and it is found that due to the difference in water consumption in winter and summer, the results of the pump station time-based pressure regulation strategy will also be quite different. Therefore, due to the different daily water consumption, the pump station outlet pressure curve will also change accordingly. In order to ensure the accuracy and applicability of the daily time-based pressure regulation strategy, it is recommended that the pressure regulation period and pressure regulation value can be adjusted daily according to actual needs.

[0192] Once the time-phased pressure regulation strategy is determined, the water supply pressure value for each period is determined. Next, an optimization algorithm is needed to optimize the pump start-stop combination and speed ratio matching in the pump station in order to obtain the optimal combination solution with the lowest energy consumption.

[0193] Verification of the applicability of GWO-CS algorithm for pump station optimization operation

[0194] In order to explore the applicability of the GWO-CS algorithm for solving the mathematical model of pump station optimization operation, the following algorithm environment is built: the solution step is set to 24h, the prediction result of the VMD-LSTM model is used as the input flow, and the 24h pressure value determined by the pump station time-division pressure regulation strategy is used as the pressure input, and then the running water pump combination and the speed ratio of the running speed regulating pump corresponding to each time period are optimized. In order to improve the applicability of the output optimization solution, and considering the characteristics that the pressure difference in each time period in the time-division pressure regulation strategy does not exceed 3m, a pressure optimization range is set every 3m in the program to appropriately narrow the algorithm optimization range.

[0195] The parameter settings of the GWO-CS algorithm for optimizing the pump station operation problem are as follows: the population size is set to 30, and the maximum number of iterations is set to 100. The optimization iteration diagrams under different pressure range requirements obtained by Matlab2022b software programming are as follows: Figures 14 to 17 As shown. From the iteration diagram of the pump station optimization scheme under different pressure range requirements in the figure, it can be seen that the pump station operation optimization based on the GWO-CS algorithm has high calculation accuracy and fast iteration speed, and the optimal solution converges quickly under different pressure range requirements. According to the different matching flow rates under the corresponding pressure range and the different start-stop schemes of the pump units, the difficulty of the optimization process will be different. In the optimization process of the pump station operation scheme under different conditions based on the GWO-CS algorithm, the optimal solution is converged within 40 iterations, and the average calculation time is 0.52s, showing good optimization convergence ability and calculation speed. It can be considered that the GWO-CS algorithm is applicable and feasible when solving pump station optimization operation problems.

[0196] Pumping station optimization operation scheme based on GWO-CS algorithm

[0197] In order to achieve the normal operation of the pump station and reduce the operating costs, it is necessary to formulate a scientific and reasonable daily operation plan for the pump station. This optimization plan needs to meet the water supply demand in each time period, while effectively reducing the daily operating costs. The computer technology is used for optimization calculations, and the constraints such as flow, head, and number of pump units are considered in the calculation to find the best daily operation plan for the pump station, ensuring that the operation plan of the pump station is optimal in terms of operating efficiency and economic benefits, and avoiding the uncertainty of blindly adjusting the operation based on experience.

[0198] The GWO-CS algorithm is now used to determine the optimal operation plan of the pump station in each time period, that is, to find the most effective pump operation combination and the speed ratio of the speed regulating pump in each time period. In addition, the efficiency of the operating water pump under each plan is calculated. The results of the optimization plan are shown in Table 4 and Table 4.

[0199] Table 4 Operation of the optimized pressure regulation control scheme for booster pump stations with water consumption in winter

[0200]

[0201]

[0202] Table 5 Operation of the optimized pressure regulation control scheme for booster pump stations with water consumption in summer

[0203]

[0204] By observing the daily operation optimization plan, it can be noticed that the water consumption varies in different time periods every day. The GWO-CS algorithm can determine the best pump start combination and the speed ratio of the speed regulating pump according to the water consumption in each time period. The pump control distribution in the optimization plan is more reasonable. It not only optimizes the operation status of the pump while meeting the water supply demand, but also avoids frequent start and stop of the pump unit, thereby reducing the loss of the pump and extending its service life.

[0205] Through optimization calculation, computer technology can be used to assist in formulating the daily operation plan of the pump station to achieve economical and efficient operation. This method not only improves the scientificity and rationality of the operation of the pump station, but also effectively reduces the operating costs of the pump station.

[0206] Energy saving effect analysis

[0207] (1) Analysis of pump unit working efficiency

[0208] For pump units, the key to saving energy consumption is to ensure that the water pump operates in the high-efficiency range of the frequency conversion, so it is very important to adjust the frequency conversion parameters reasonably. Therefore, with the user's water volume and head supply as the constraint condition, the minimum speed ratio of the pump unit is limited, and the GWO-CS algorithm is used to calculate the optimal speed ratio of the pump unit operating within the constraint condition range. The optimization results are sorted out and plotted into the efficiency-flow-head relationship as shown in the figure. Fig.18 and Fig.19 In the efficiency-flow-head relationship diagram, the "higher" the point position is, the higher the efficiency of the pump during that period. If the points at high positions are more "dense", it means that the pump has better working efficiency within certain working conditions. If the points at high positions are more dispersed, it means that the pump has better working efficiency within a larger working condition range.

[0209] from Fig.18 and Fig.19It can be seen that under the control strategy of the optimization scheme, regardless of winter or summer, the working efficiency of each pump unit is above 72%, and most of the points are at a high level, indicating that the pump unit maintains a relatively high efficiency working state in most periods. The proportion of the efficient operation time of the internal water pump to the entire day's operation time is statistically analyzed, and the results are shown in Table 6.

[0210] Table 6

[0211]

[0212] From the statistical results in Table 6, it can be seen that under the optimization scheme, the proportion of time periods with pump efficiency greater than 75% in a single day exceeds 80%; the proportion of time periods with high efficiency greater than 80% in winter and more than 20% in summer is close to 30% throughout the day. Overall, under the optimization scheme, the pump's operating efficiency remains at a high level most of the time throughout the day, indicating that the pump's energy consumption has been improved and the energy consumption at most times has been effectively utilized at a high level.

[0213] (2) Analysis of electricity consumption costs

[0214] After using the GWO-CS algorithm to solve the optimal pump operation combination and speed ratio of the speed regulating pump in each period, the total power of the pump in each period under the corresponding scheme is calculated according to the water consumption in different seasons. Then, the calculated total power can be compared with the power consumption under the original scheme. The effect of the improved scheme can be evaluated based on the comparison results. If the total power under the new scheme is lower than that of the original scheme, it means that the energy consumption of running the water pump has been reduced, thus achieving the goal of energy saving. On the contrary, if the total power of the new scheme is higher, then the effectiveness of the scheme needs to be re-evaluated.

[0215] In this way, the total power calculated by the running water pump of the optimization scheme given by the GWO-CS algorithm under different seasonal water use conditions can be objectively evaluated, and compared with the power consumption of the original scheme. The comparative analysis results are shown in Tables 7 and 8.

[0216] Table 7 Power consumption analysis of voltage regulation optimization strategy in different time periods in winter

[0217]

[0218]

[0219] Table 8 Power consumption analysis of the optimized voltage regulation strategy in different time periods in summer

[0220]

[0221] From the analysis of the results in the above table, it can be seen that under the condition of high flow, the optimization scheme can greatly reduce energy consumption by turning on two variable frequency pumps for full variable frequency operation, and the energy saving is relatively large; however, under the condition of low flow, the full variable frequency operation of a single pump cannot fully utilize the advantages of variable frequency, and the energy saving is relatively small. The highest hourly energy saving ratio in winter is up to 24.2%, and the average daily energy saving ratio is 13.4%; the highest hourly energy saving ratio in summer is 32.2%, and the average daily energy saving ratio is 18.1%. According to the investigation, the city charges the basic electricity fee for the water supply pump station at 0.6583 yuan / kwh, and does not set the peak electricity fee in Pinggu. It is estimated that the annual electricity saving can reach 181,187.99 yuan.

[0222] According to the above pump station daily operation optimization results, it can be considered that by applying the GWO-CS optimization algorithm to adjust the control strategy of the pump station, the problem of blindly adjusting the operation based on experience is avoided, the operation efficiency of the pump station is significantly improved, and the economic operation of the pump station is achieved, which provides a reference for further improving the scientific and green automation control goals of water supply projects.

[0223] Analysis of operating pressure of pipe network system

[0224] In the original plan, the outlet pressure of the pump station is constant 24 hours a day. The pipe network at the most unfavorable point will have excess pressure waste when the water consumption is low at night. At the same time, affected by the fluctuation of water consumption, the pressure at this point will also fluctuate greatly within a day. The pump station outlet pressure is regulated and managed based on the predicted water volume as a guide, and the nighttime pressure is appropriately reduced. The pressure can also be adjusted appropriately according to the fluctuation of water consumption. The basic information of the pressure detection point is shown in Table 9.

[0225] Table 9

[0226]

[0227] After adjusting the pump station operation strategy, the pressure value at the most unfavorable point is measured again and compared with the measured value under the original scheme before adjustment. Figure 20 to Figure 22 shown.

[0228] (1) Analysis of pressure drop at the most unfavorable point in the pipe section

[0229] from Fig. 20 It can be seen that after adjusting the operation strategy of the booster pump station, the pressure value at the most unfavorable point has been significantly reduced at night, and the overall fluctuation trend is more stable than before. The pressure detected on that day was 2.46m lower than that under the original operation plan. It can be considered that after the optimized operation of the pump station, unnecessary pressure waste at the most unfavorable point has been reduced, and its pressure fluctuation has been more stable.

[0230] (2) Analysis of pressure reduction in overpressure area

[0231] Based on the collected historical statistical records, pressure analysis was conducted in two high-risk areas with low terrain characteristics where pipe burst accidents frequently occurred. Both areas had obvious overpressure during the night time. The pressure value of the No. 1 overpressure detection point fluctuated greatly. The average pressure of the two areas on the monitoring day significantly exceeded the pressure value required for calculation. After adjusting the pump station operation strategy, the pressure values ​​of the two overpressure points were measured again and compared with the measured values ​​under the original scheme before adjustment. Fig.21 and Fig. 22 As shown in the figure, after the pressure of the pipeline network was adjusted by optimizing the operation of the pressure pump station, the pressure values ​​of the two overpressure points at night were significantly reduced, the pressure fluctuation was more stable, and the pressure values ​​of the two nodes were generally reduced. The pressure value of the No. 1 overpressure detection point was reduced by an average of 3.49m per day, and the pressure value of the No. 2 overpressure detection point was reduced by an average of 4.16m per day. The pressure values ​​of the two places were closer to the actual required pressure values, which not only eliminated a lot of unnecessary pressure waste, but also reduced the risk of pipe burst.

[0232] summary

[0233] Taking the actual engineering booster pump station and water supply network as an example, the actual operation status of the pump station and the problems existing in the network are analyzed, and the pump station optimization operation strategy based on the GWO-CS algorithm is applied to the urban pump station engineering example. The optimization scheme of the pump combination and speed ratio distribution of the pump station at 24 moments is obtained on a single day as the research scale. The optimization scheme is compared with the traditional scheme to explore the energy-saving effect of the optimization scheme in actual engineering.

[0234] The main conclusions obtained in the engineering case application research are as follows: in terms of energy saving and consumption reduction, the pump efficiency in each period of the optimization scheme is maintained at a high level, reflecting its great energy saving potential. The energy consumption reduction effect of the optimization scheme is further verified by comparing it with the energy consumption of the historical scheme. The research results show that an average of 13.4% energy consumption can be saved per day in winter; an average of 18.1% energy consumption can be saved per day in summer, and the estimated annual electricity bill can be saved by 181,187.99 yuan. In terms of reducing the operating pressure of the pipeline network during the low-peak period of water use, the pipeline pressure of the water supply system operating under the time-sharing pressure regulation strategy did not appear under-pressure. Under the premise of ensuring normal water supply demand, the pipeline pressure of some low-peak periods of water use was appropriately reduced. The analysis of several representative pressure measuring points in the system found that the pressure at the most unfavorable point during the low-peak period of water use was reduced by 7m at the most, and the average daily reduction was 2.46m; the pressure in the overpressure area of ​​the pipeline network was reduced by 6-7m at the most during the low-peak period, and the average daily reduction was 3-4m. In general, the problem of overpressure operation of the pipeline system has been effectively improved, which is of great significance for reducing the leakage rate of the pipeline network and reducing the maintenance rate of pipe bursts.

[0235] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit the technical solution. Those skilled in the art should understand that those modifications or equivalent substitutions of the technical solution of the present invention that do not depart from the purpose and scope of the technical solution should be included in the scope of the claims of the present invention.

Claims

1. A method for regulating the outlet pressure of a pump station for a water supply network, characterized in that: The following steps are involved: S1. Construct a pump station outlet pressure prediction model; used to predict the pump station outlet pressure in the next N hours of time series based on the historical water supply flow sequence data, the pressure sequence data of the pressure measuring point and the water supply pressure sequence data required by the pump station; S2. Obtain historical training data and perform preprocessing to obtain model training data; the historical training data includes historical water supply flow sequence data, pressure measurement point pressure sequence data, and water supply pressure sequence data required by the pump station; S3, using the model training data obtained in S2, train the pump station outlet pressure prediction model constructed in S1; S4, predicting the pump station outlet pressure of the time series of the next N hours through the trained pump station outlet pressure prediction model; S5, based on the prediction result of S4, dividing the pump station outlet pressure into multiple water supply periods; When dividing the water supply period, according to the predicted fluctuation of the pump station outlet pressure, the maximum continuous time interval in which the fluctuation value is less than the preset value is taken as a water supply period; S6. For each water supply period, the maximum pump station outlet pressure value in the period is used as the constant outlet pressure of the water supply period; the constant outlet pressures of each water supply period are integrated as a control scheme; S7. According to the control scheme of S6, actual pump station outlet pressure control is performed to provide corresponding constant outlet pressures in different water supply periods.

2. The method for regulating the outlet pressure of a pump station for a water supply network according to claim 1, characterized in that: In S7, when a corresponding constant outlet pressure is provided in a water supply period, the optimal power control scheme of the water pump unit is solved by a preset power optimization model, and the water pump unit is controlled according to the obtained optimal power control scheme.

3. The method for regulating the outlet pressure of a pump station for a water supply network according to claim 2, characterized in that: The objective function of the power optimization model is: In the formula, c is the electricity price during the period; n is the number of speed regulating pumps; ω i To control the opening and closing coefficient of the i-th pump, the value is 0 or 1; Q i is the output flow of the i-th pump; S i is the speed ratio of the i-th pump; d 0i d 1i d 2i is the power fitting coefficient of the i-th pump.

4. The method for regulating the outlet pressure of a pump station for a water supply network according to claim 2, characterized in that: The constraints of the power optimization model include single pump flow constraint, pump head constraint, single pump shaft power constraint, number of units in operation constraint, and speed ratio constraint.

5. The method for regulating the outlet pressure of a pump station for a water supply network according to claim 4, characterized in that: The single pump flow constraint is: Q min <Q i <Q max ; In the formula, Q i is the flow rate of the i-th pump in the pump station; Q min , Q max are the minimum flow and maximum flow of the i-th water pump respectively; The pump head constraint is: H min <H i <H max ; In the formula, H i is the head of the i-th water pump; H min , H max are the minimum head and maximum head of the i-th water pump respectively.

6. The method for regulating the outlet pressure of a pump station for a water supply network according to claim 4, characterized in that: The power constraint of a single pump shaft is: P min <P i <P max ; Where P i is the shaft power of the i-th water pump; P 1in , P max are the minimum and maximum shaft power of the ith pump in the pump station, respectively; The number of machines in operation is constrained as follows: 0<n<n max ; In the formula, n is the number of pumps in operation in the pump station, n max The preset maximum number of powered-on units.

7. The method for regulating the outlet pressure of a pump station for a water supply network according to claim 4, characterized in that: The speed ratio constraint is: S i ∈[S imin ,1]; Where S i is the speed ratio of the i-th pump; S imin is the minimum speed ratio of the i-th pump.

8. The method for regulating the outlet pressure of a pump station for a water supply network according to claim 2, characterized in that: In S7, when solving the optimal power control scheme of the water pump unit through the power optimization model, the solution is obtained through an optimization method based on a hybrid improved cuckoo algorithm; The workflow of the optimization method based on the hybrid improved cuckoo algorithm includes: Step 1: Randomly generate N bird nests as the initial population, record the initial optimal bird nest position, and perform parameter initialization settings; Step 2: Use the Levy flight strategy to update the positions of other bird nests except the current best bird nest position, calculate the fitness value of each bird nest after the update, and then use the bird nest position with the largest fitness value among all the bird nests as the new best bird nest position; Step 3: Use the update strategy of the GWO algorithm to update the location of each bird's nest again; Step 4: For other bird nest positions other than the optimal solution of the GWO algorithm, random updates are performed in combination with the preset discovery probability Pa; wherein the random update process includes: generating a random number r for a bird nest position to be randomly updated, and comparing r with the discovery probability Pa; if r>Pa, updating the bird nest position by random walking; otherwise, the position of the bird nest remains unchanged; Step 5: Calculate the fitness values ​​of all bird nest positions again, and record the bird nest position corresponding to the best fitness value as the optimal solution of the current round; Step 6: Determine whether the termination condition is met. If so, output the optimal solution of the current round as the final result; if not, return to step 2 and perform the next round of iteration.

9. The method for regulating the outlet pressure of a pump station for a water supply network according to claim 8, characterized in that: In step 1, the initialization parameters include the abandonment probability Pa, the maximum number of iterations Tmax, and the parameters of the GWO algorithm; In step 6, the termination condition is that the number of iterations reaches the maximum number of iterations Tmax.

10. The method for regulating the outlet pressure of a pump station for a water supply network according to claim 8, characterized in that: The pressure value in the solution of the power optimization model calculation optimization process meets the following conditions: |H i -H i |≤0.2; In the formula, H i is the constant outlet pressure of the ith water supply period in the control scheme; H j Calculate the pressure value in the optimization process solution for the corresponding water supply period.

Images