A Real-time Regulation Method for Pumping Stations Driven by Both Data and Mechanism

Abstract

The present invention discloses a real-time regulation method for a pumping station driven by both data and mechanism, including determining key influencing factors that determine the number of operating units of the pumping station; determining the entire feasible region of the real-time regulation of the pumping station, constructing and solving a mechanism-driven model for the real-time regulation of the pumping station, obtaining the distribution of the number of operating units of the pumping station under the entire feasible region as the solution of the mechanism-driven model; determining the local feasible region of the real-time regulation of the pumping station, constructing and solving a data-driven model for the real-time regulation of the pumping station, obtaining the distribution of the number of operating units of the pumping station under the local feasible region as the solution of the data-driven model; constructing a fusion data set based on the solution of the mechanism-driven model and the solution of the data-driven model, training and validating a data-mechanism dual-driven model based on the fusion data set, and obtaining the final regulation decision result. The advantages are as follows: while pursuing the efficient operation of the pumping station, a scheduling decision that better meets the on-site regulation requirements is generated, and to a certain extent, the problem of frequent start-stop of units caused by the mechanism-driven model pursuing the efficient operation of the pumping station is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of optimal scheduling of pumping stations, and particularly to a real-time regulation method for pumping stations driven by both data and mechanism. Background Art

[0002] To complete the water transfer volume tasks in each period, it is necessary to comprehensively consider real-time water conditions and working conditions information to determine the operation plan of the pumping station units. On the basis of known pump performance parameters and under the premise of meeting relevant constraints, reasonably determining parameters such as the unit combination mode, flow rate, and rotational speed of the pumping station is of great significance for realizing the optimal operation of the pumping station. At present, water transfer projects with pumping stations as the core generally face problems such as low efficiency of pumping devices, waste of device energy, and high operation costs during operation. To achieve the goal of energy conservation and consumption reduction, domestic and foreign scholars have conducted a large number of studies on the optimal operation of pumping stations. At present, the methods for real-time regulation of pumping stations mainly include the following ways:

[0003] (1) Real-time regulation of pumping stations based on optimization algorithms

[0004] The real-time scheduling method of pumping stations based on optimization algorithms uses optimization algorithms to solve parameters such as the unit combination mode, flow rate, and rotational speed of the pumping station under the premise of meeting relevant constraints, so that the objective function of the pumping station operation reaches the maximum or minimum. Commonly used optimization algorithms include: genetic algorithm, ant colony algorithm, particle swarm algorithm, artificial neural network, etc. Although a large number of studies have confirmed the significant advantages of this method in improving the efficiency of pumping stations and reducing operation energy consumption, the application of this regulation method in actual projects is very limited. In actual applications, in order to pursue the efficient operation of the pumping station, this regulation method may cause frequent start-stop of the units, accelerating the aging of equipment while increasing the work pressure of on-site personnel.

[0005] (2) Real-time regulation of pumping stations based on manual experience

[0006] The real-time regulation method of pumping stations based on manual experience mainly relies on the experience and intuition of on-site dispatchers to determine when to start the pumping station units and how many units to start. It cannot be denied that this regulation method contains the rich wisdom and work experience of on-site dispatchers, but this method lacks scientificity and accuracy, and will sacrifice the operation efficiency of the pumping station to a certain extent, resulting in the pumping station continuously deviating from the high-efficiency area. In addition, the experience and intuition of on-site dispatchers may deviate according to different situations, resulting in the uncertainty of real-time regulation of pumping stations and making it difficult to adapt to unstable and uncertain environments. Summary of the Invention

[0007] The purpose of the present invention is to provide a real-time regulation method for pumping stations driven by both data and mechanism, so as to solve the foregoing problems existing in the prior art.

[0008] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0009] A real-time regulation method for a pumping station driven by both data and mechanism, comprising the following steps:

[0010] S1. Determination of key influencing factors:

[0011] Combining the actual situation of on-site dispatching of the pumping station and the theoretical research on real-time optimal regulation of the pumping station, summarize the corresponding relationship between the water regime data of the pumping station and the regulation strategy, and determine the key influencing factors that determine the number of operating pump units of the pumping station;

[0012] S2. Construction and solution of the mechanism-driven model:

[0013] Based on the key influencing factors that determine the number of operating pump units of the pumping station and the constraint conditions of real-time regulation and operation of the pumping station, determine the entire feasible region of real-time regulation of the pumping station; construct a real-time regulation mechanism-driven model of the pumping station by setting the objective function, decision variables, constraint conditions and solution algorithms, and input the entire feasible region of real-time regulation of the pumping station into the real-time regulation mechanism-driven model of the pumping station to obtain the distribution of the number of operating pump units of the pumping station under the entire feasible region as the solution of the mechanism-driven model;

[0014] S3. Construction and solution of the data-driven model:

[0015] Based on the historical operating water regime data and regulation decisions of the pumping station, determine the local feasible region of real-time regulation of the pumping station; construct a real-time regulation data-driven model of the pumping station based on the decision tree classification algorithm, and input the local feasible region of real-time regulation of the pumping station into the real-time regulation data-driven model of the pumping station to obtain the distribution of the number of operating pump units of the pumping station under the local feasible region as the solution of the data-driven model;

[0016] S4. Construction and solution of the data-mechanism dual-driven model:

[0017] Construct a fusion data set based on the solution of the mechanism-driven model and the solution of the data-driven model, train and verify a data-mechanism dual-driven model based on the decision tree based on the fusion data set, and obtain the final regulation decision result.

[0018] Preferably, step S1 specifically includes the following contents:

[0019] S101. Consult relevant literature on the optimal operation of the pumping station, determine the commonly used objective functions for the optimal operation of the pumping station, and determine the preliminary selected influencing factors according to the calculation formula of the objective function;

[0020] S102. Combine the on-site requirements and consider the water regime and project situation information that affect the on-site regulation decision;

[0021] S103. Combine the literature research and on-site requirements to determine that the key influencing factors that determine the number of operating pump units of the pumping station include the flow rate through the pumping station and the head of the pumping station.

[0022] Preferably, the determination of the entire feasible region for real-time regulation of the pump station in step S2 specifically includes the following contents:

[0023] S201. Query the design parameters of the pump station, including the minimum and maximum values of the key influencing factors that determine the number of operating pump units in the pump station;

[0024] S201. Discretize the key factors at a reasonable step size to generate a solution space;

[0025] S201. According to the constraint conditions for real-time regulation and operation of the pump station, eliminate the working conditions that do not meet the starting conditions, and obtain the entire feasible region for real-time regulation of the pump station.

[0026] Preferably, the constraint conditions for real-time regulation and operation of the pump station include the flow rate constraint of the pump unit, the head constraint of the pump unit, the downstream / upstream water level constraint of the pump station operation, the constraint on the number of operating units, and the pump flow-head characteristic curve constraint.

[0027] Preferably, the objective function of the mechanism-driven model for real-time regulation of the pump station is to maximize the operation efficiency of the pump station under the working conditions; the decision variable of the mechanism-driven model for real-time regulation of the pump station is the number of operating pump units in the pump station; the solution algorithm of the mechanism-driven model for real-time regulation of the pump station is the particle swarm optimization algorithm.

[0028] Preferably, the determination of the local feasible region for real-time regulation of the pump station in step S3 specifically includes the following contents:

[0029] S3011. Collect the water regime data since the annual operation of the pump station and the regulation and decision-making data since the annual operation of the pump station; all data are the scheduling results of the same time period;

[0030] S3012. Analyze the distribution characteristics of the historical operation data of the pump station. According to the minimum / maximum values of the flow rate and head of the operation data, float by a certain proportion to obtain the local flow rate and head operable intervals, and eliminate the working conditions that do not meet the starting conditions according to the operation constraint conditions to determine the local feasible region for real-time regulation of the pump station;

[0031] S3013. Analyze the distribution characteristics of the historical operation data of the pump station to determine the sample size and whether the distribution is balanced. Preferably, the construction and solution of the data-driven model for real-time regulation of the pump station in step S3 specifically include the following contents:

[0032] S3021. Divide the preprocessed historical regulation data of the pump station into a training set and a validation set at a certain proportion;

[0033] S3022. Build a data-driven model for real-time regulation of the pump station based on the decision tree classification algorithm;

[0034] S3023. Train and validate the real-time regulation data-driven model of the pumping station based on the training set and the validation set, and evaluate the performance of the model based on the evaluation metrics;

[0035] S3024. Input the real-time regulation local feasible region of the pumping station into the trained real-time regulation data-driven model of the pumping station to obtain the distribution of the number of operating units under the real-time regulation local feasible region of the pumping station as the solution of the data-driven model.

[0036] Preferably, the construction of the fusion data set in step S4 specifically includes the following contents.

[0037] S4011. Connect the solution of the mechanism-driven model and the solution of the data-driven model in series to form a fusion data set.

[0038] S4012. Perform normalization processing on the fusion data set.

[0039] S4013. Divide the normalized fusion data set into a training set and a validation set.

[0040] Preferably, the obtaining of the final regulation decision result in step S4 specifically includes the following contents.

[0041] S4021. Use a decision tree to construct a data-mechanism dual-driven model, input the training set into the data-mechanism dual-driven model for training, and input the validation set into the trained data-mechanism dual-driven model to output the model result.

[0042] S4022. Perform inverse normalization processing on the model result to obtain the final regulation decision result.

[0043] Preferably, after step S4022, it further includes

[0044] S4023. Evaluate the reliability of the data-mechanism dual-driven model, and use the entire feasible region of the model as the input to calculate the global regulation decision result of the model.

[0045] The beneficial effects of the present invention are as follows: 1. The method of the present invention combines the solution results of the mechanism-driven model and the historical operation scheduling decisions. While pursuing the efficient operation of the pumping station, it generates scheduling decisions that better meet the on-site regulation requirements. To a certain extent, it can avoid the frequent start-stop problems of the units caused by the mechanism-driven model's pursuit of the efficient operation of the pumping station, delay the aging of the equipment, and reduce the work pressure of on-site regulation personnel. 2. The method of the present invention takes into account the problem that the actual operating conditions are relatively concentrated and the historical samples are insufficient. It divides the local feasible region according to the historical operation range and uses a data-driven method to solve the solution of the local feasible region. It fuses the local solution based on the data-driven model and the global solution based on the mechanism-driven model, integrating the characteristics of the whole region and local features, and enhancing the comprehensiveness and insight of the data. 3. The method of the present invention aims at the problem of the fusion of the data-driven model and the mechanism-driven model, and performs data fusion at different scales at the dataset level to form a mechanism-data dual-driven fusion dataset. This method is applicable when integrating multiple models with similar performances, greatly improving the reliability and interpretability of the regulation decision-making, and obtaining a more comprehensive and optimized regulation plan. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0047] Figure 2 is the schematic diagram of the decision-making scheme of the mechanism-driven model in the embodiment of the present invention;

[0048] Figure 3 is the schematic diagram of the decision-making scheme of the data-driven model in the embodiment of the present invention;

[0049] Figure 4 is the schematic diagram of the decision-making scheme of the data-mechanism dual-driven model in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0050] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0051] Such as Figure 1As shown in the figure, in this embodiment, aiming at the problems of frequent start-stop of pump station units and low operating efficiency of pump stations existing in the existing pump station regulation methods, a real-time regulation method for pump stations driven by data and mechanism is proposed. This method uses the particle swarm optimization algorithm, with the highest operating efficiency of the pump station as the goal, constructs a real-time regulation model of the pump station mechanism, combines the historical dispatching decision-making scheme of the pump station operation, learns the historical dispatching decision and the optimal decision of the mechanism-driven model through the decision tree classification model, generates a new decision-making scheme, and realizes the reasonable and efficient operation of the pump station units. The present invention alleviates problems such as the difficult application of the mechanism-driven model for regulation, and can be used as an effective method for the real-time regulation of pump stations. The method of the present invention mainly includes the following four parts:

[0052] I. Determination of key influencing factors

[0053] Combined with the actual situation of on-site dispatching of the pump station and the theoretical research on real-time optimal regulation of the pump station, the corresponding relationship between the water condition data of the pump station and the regulation strategy is summarized, and the key influencing factors that determine the number of pump units started are determined. The specific contents are as follows:

[0054] (1). Consult relevant literatures on the optimal operation of the pump station, determine the commonly used objective functions (such as pump station efficiency) for the optimal operation of the pump station, and determine the primary influencing factors according to the calculation formula of the objective function.

[0055] (2). Considering the on-site dispatching requirements, on-site dispatching personnel mainly consider factors such as water conditions and working conditions when implementing regulation decisions, including the over-flow rate of the pump station, the pump head of the pump station, the operating status of the units, and the equipment wear condition, etc.

[0056] (3). Combining literature research and on-site requirements, it is determined that the key influencing factors that determine the number of pump units started in the pump station include the over-flow rate of the pump station and the pump head of the pump station.

[0057] II. Construction and solution of the mechanism-driven model

[0058] Based on the key influencing factors that determine the number of pump units started in the pump station and the constraint conditions for the real-time regulation operation of the pump station, the entire feasible region for the real-time regulation of the pump station is determined; a real-time regulation mechanism-driven model for the pump station is constructed by setting the objective function, decision variables, constraint conditions, and solution algorithm, and the entire feasible region for the real-time regulation of the pump station is input into the real-time regulation mechanism-driven model of the pump station to obtain the distribution of the number of pump units started under the entire feasible region of the pump station, which is used as the solution of the mechanism-driven model.

[0059] In this embodiment, determining the entire feasible region for the real-time regulation of the pump station specifically includes the following contents:

[0060] (1). Query the design parameter values of the pump station, including the minimum and maximum values of the key influencing factors.

[0061] (2). Discretize the key influencing factors according to a reasonable step size to generate a solution space.

[0062] (3) Based on the constraint conditions for the real-time regulation and operation of the pumping station, eliminate the working conditions that do not meet the starting conditions to obtain the entire feasible region for the real-time regulation of the pumping station. The real-time regulation and operation of the pumping station need to be within certain constraint conditions, including: the flow constraint of the pump unit, the head constraint of the pump unit, the downstream / upstream water level constraint of the pumping station operation, the constraint of the number of operating units, the flow-head characteristic curve constraint of the pump, etc.

[0063] In this embodiment, by taking the key influencing factors as decision variables and based on the particle swarm optimization algorithm, a mechanism-driven model for the real-time regulation of the pumping station is constructed, and then the optimal distribution of the number of starting units under different flow-head combinations is generated. The specific contents are as follows.

[0064] (1) Establish the objective function of the mechanism-driven model for the real-time regulation of the pumping station. In the present invention, the mechanism-driven model for the real-time regulation of the pumping station aims to maximize the operation efficiency of the pumping station under the calculated working conditions.

[0065] (2) Establish the decision variables of the mechanism-driven model for the real-time regulation of the pumping station. In the present invention, the mechanism-driven model for the real-time regulation of the pumping station takes the number of starting units of the pumping station as the decision variable.

[0066] (3) Establish the constraint conditions of the mechanism-driven model for the real-time regulation of the pumping station. In the present invention, under the condition that the operating flow of the pumping station is determined, the constraint conditions of the pump group flow, head, and water level need to be considered.

[0067] (4) Establish the solution algorithm of the mechanism-driven model for the real-time regulation of the pumping station. In the present invention, the solution is carried out according to the particle swarm optimization algorithm.

[0068] (5) According to the theoretical optimal solution of the mechanism-driven model for the real-time regulation of the pumping station, establish the distribution of the number of starting units of the pumping station in the entire feasible region, and generate the solution S1 of the mechanism-driven model.

[0069] III. Construction and solution of the data-driven model

[0070] Based on the historical operation water regime data and regulation decisions of the pumping station, determine the local feasible region for the real-time regulation of the pumping station; construct a data-driven model for the real-time regulation of the pumping station based on the decision tree classification algorithm, and input the local feasible region for the real-time regulation of the pumping station into the data-driven model for the real-time regulation of the pumping station to obtain the distribution of the number of starting units of the pumping station in the local feasible region as the solution of the data-driven model.

[0071] In this embodiment, determining the local feasible region for the real-time regulation of the pumping station specifically includes the following contents.

[0072] (1) Collect the water regime data since the annual operation of the pumping station, specifically including the head and flow rate of the pumping station; collect the regulation decisions since the annual operation of the pumping station, specifically including the number of starting units of the pumping station; all data are the scheduling results of the same time period.

[0073] (2) Analyze the distribution characteristics of historical operation data, find the minimum and maximum values of flow rate and head, and float up and down according to a certain proportion to obtain the operable range of local flow rate and head. Then, eliminate the working conditions that do not meet the starting conditions according to the operation constraint conditions, and establish the local operable domain of the pumping station based on historical data.

[0074] (3) Summarize the distribution of the number of operating pump units in the pumping station according to historical operation data, and analyze the sample size and its distribution.

[0075] In this embodiment, based on the preprocessed historical regulation data of the pumping station, a decision tree classification algorithm is used to generate a data-driven model for real-time regulation of the pumping station, and the performance of the model is evaluated. The specific contents are as follows.

[0076] (1) Divide the preprocessed historical regulation data of the pumping station into a training set and a validation set according to a certain proportion.

[0077] (2) Construct a data-driven model for real-time regulation of the pumping station based on the decision tree classification algorithm. Essentially, it is a classification model for the number of operating pump units based on two factors: flow rate and head.

[0078] (3) Conduct model performance evaluation, and evaluate the performance of the model based on evaluation indicators such as accuracy, precision, recall rate, and F1-score.

[0079] (4) Input the local feasible domain into the data-driven model to generate the distribution of the number of operating pump units under the local feasible domain of real-time regulation of the pumping station, which constitutes the solution S2 of the data-driven model.

[0080] IV. Construction and solution of the data-mechanism dual-driven model

[0081] Based on the solution of the mechanism-driven model and the solution of the data-driven model, a fusion data set is constructed. Based on the fusion data set, a data-mechanism dual-driven model based on decision tree is trained and verified to obtain the final regulation decision result.

[0082] In this embodiment, the construction of the fusion data set specifically includes the following contents.

[0083] (1) Connect the solution S1 of the mechanism-driven model and the solution S2 of the data-driven model in series to form the fusion data set S.

[0084] (2) Perform normalization processing on the fusion data set.

[0085] (3) Divide the normalized fusion data set into a training set and a validation set.

[0086] In this embodiment, the acquisition of the final regulation decision result specifically includes the following contents.

[0087] (1) Construct a data-mechanism dual-driven model using a decision tree. Input the training set into the data-mechanism dual-driven model for training, and input the validation set into the trained data-mechanism dual-driven model to output the model results.

[0088] (2) Perform denormalization processing on the model results to obtain the final regulation decision results.

[0089] (3) Evaluate the reliability of the data-mechanism dual-driven model, and use the entire feasible region of the model as the input to calculate the global regulation decision results of the model.

[0090] Example Two

[0091] In this example, taking a certain pumping station in the Eastern Route Project of the South-to-North Water Diversion as an example, the method flow provided by the present invention is used to establish the implementation regulation plan of the pumping station. The designed scale of this pumping station is 120 m 3 / s, with 5 units, among which 1 unit is a standby unit, and the single-unit flow rate is 30 m 3 / s. The flow rate range of the pumping station is 14.2 - 191.4 m 3 / s; the operating head of the pumping station is relatively low, with a designed head of 3.73 m and an average head of 1.6 m. The head range of the pumping station is 0 - 3.8 m, and the annual operating time is 5000 hours, and it pumps water for 210 days every year. The basic parameters of the pump units of this pumping station are shown in Table 1.

[0092] Table 1 Basic parameters of the pump units of the pumping station

[0093] Name Parameter Pump type ( / ) Rear-mounted bulb tubular pump Number of installed units (unit) 5 (4 in use and 1 standby) Single unit capacity (kW) 2000 Total installed capacity (kW) 10000 Pump impeller diameter (m) 3.2 Pump speed (r / min) 120 Motor speed (r / min) 750 Rated power of the supporting motor (KW) 2000 Total installed capacity of the motor (kW) 10000 Drive mode Gearbox drive

[0094] I. Obtaining key influencing factors

[0095] Combining literature research and on-site requirements, the key influencing factors determining the number of operating units of the pumping station are identified as: the flow rate through the pumping station and the head of the pumping station.

[0096] II. Construction and solution of the mechanism-driven model

[0097] 2.1 Determination of the entire feasible region for real-time regulation of the pumping station

[0098] The flow rate range of the pumping station is 14.2 - 191.4 m 3 / s, and the head range of the pumping station is 0 - 4.73 m. The flow rate and head are discretized with a step size of 0.1 m 3 / s and 0.1 m respectively, and the working conditions that do not meet the starting conditions are excluded to obtain the entire feasible region for real-time regulation of the pumping station.

[0099] 2.2 Model construction and solution

[0100] Taking the pump station head and pump station flow rate, the key influencing factors, as decision variables, with the highest pump station operation efficiency as the objective function, a real-time regulation mechanism-driven model for the pump station is constructed based on the particle swarm optimization algorithm to generate the optimal number of operating units distribution under different flow-head combinations. In this embodiment, there are 5 identical units in the Sihong Pump Station. According to the constraint ranges of the pump station flow rate and head, and based on the calculation results of the mechanism-driven model, the optimal number of operating units is 1, 2, 3, 4, and 5. The pump station regulation scheme based on the mechanism-driven model is as Figure 1 shown.

[0101] III. Construction and solution of the data-driven model

[0102] 3.1 Determination of the local feasible region for real-time regulation of the pump station

[0103] Collect the historical operation water regime data and regulation decisions of the pump station, specifically including: the flow rate through the pump station, head, and number of operating units. Since the pump station head cannot be directly monitored by a water level gauge, in this embodiment, the water levels of the inlet and outlet ponds of the pump station are collected according to the installation of the water level gauges, and the pump station head is calculated based on the water levels of the inlet and outlet ponds. The historical operation data of the pump station collected is shown in Table 2.

[0104] Table 2 Historical operation data

[0105]

[0106]

[0107] According to the operation data analysis, the historical regulation range of the pump station is as follows: there are 4 cases in the historical samples with the number of operating units being 2, 3, 4, and 5, the flow rate operation range is 37.2 - 139.5 m 3 / s, and the head operation range is 0.05 - 1.95 m, which accounts for a relatively small proportion and is concentrated in the entire feasible region of the pump station. Based on the historical sample data, the local feasible region of the pump station is set as the flow rate range: 35.0 - 150.0 m 3 / s, and the head range: 0.0 - 2.0 m, which are discretized at steps of 0.1 m 3 / s and 0.1 m respectively, and the inoperable working conditions are excluded to obtain the local operable region.

[0108] Count the number of operating units of the pump station. A total of 2417 historical regulation data of the pump station at different times are collected. Among them, there are 23 data with 2 operating units, 36 data with 3 operating units; 2081 data with 4 operating units, and 277 data with 5 operating units. The sample size is small and there is an obvious imbalance problem. To solve the problems of small sample size and concentrated distribution of the measured data, the construction of the data-driven model solution is carried out next.

[0109] 3.2 Model construction and solution

[0110] The preprocessed historical regulation data of the pump station (a total of 2,417 records) was randomly divided into a training set (2,176 records) and a validation set (241 records) at a ratio of 90% and 10%. A data-driven model for real-time regulation of the pump station was constructed based on the decision tree classification algorithm. Essentially, it is a classification model for the number of operating units constructed based on two factors: flow rate and head. The performance of the data-driven model was evaluated using relevant indicators, and the evaluation results showed that the data-driven model was reliable. See Table 3 for the evaluation results.

[0111] Table 3 Performance of the data-driven model

[0112]

[0113] The local feasible region was input into the data-driven model to generate the distribution of the number of operating units under the local feasible region of real-time regulation of the pump station, constituting the solution S2 of the data-driven model. See Figure 3 .

[0114] IV. Construction and solution of the data-mechanism dual-driven model

[0115] 4.1 Construction of the hybrid dataset

[0116] The solution S1 of the series mechanism-driven model (a total of 57,391 samples) and the solution S2 of the data-driven model (a total of 20,106 samples) were combined to form a fusion dataset S (a total of 77,497 samples). The fusion dataset S was normalized and divided into a training set and a validation set according to a certain ratio.

[0117] 4.2 Solution of the data-mechanism dual-driven model

[0118] The training set of the fusion dataset S was input into the data-mechanism dual-driven model for training, and the validation set of the fusion dataset S was input into the trained dual-driven model for validation. The model results were de-normalized to obtain the final regulation decision results of the model.

[0119] The model prediction results were compared with the historical regulation decisions and the mechanism-driven model decisions to analyze the rationality of the results. Some prediction results are shown in Table 4. The tested model was used to predict the number of operating units in the entire feasible region to obtain the decision results based on the data-mechanism dual-driven model, as shown in Figure 4 .

[0120] Table 4 Decision results of the number of operating units of the pump station under different regulation modes

[0121]

[0122]

[0123] By adopting the above technical solutions disclosed in the present invention, the following beneficial effects are obtained:

[0124] The present invention provides a real-time regulation method for pumping stations driven by both data and mechanism. The method of the present invention combines the solution results of the mechanism-driven model and historical operation scheduling decisions. While pursuing the efficient operation of the pumping station, it generates scheduling decisions that better meet the on-site regulation requirements. To a certain extent, it can avoid the frequent start-stop problems of the units caused by the mechanism-driven model's pursuit of the efficient operation of the pumping station, delay the aging of the equipment, and reduce the work pressure of on-site regulation personnel. Considering the problem that the actual operating conditions are relatively concentrated and the historical samples are insufficient, the method of the present invention divides the local feasible region according to the historical operation range, and uses the data-driven method to solve the solution of the local feasible region. It fuses the local solution based on the data-driven model and the global solution based on the mechanism-driven model, combines the characteristics of the global region and local features, and enhances the comprehensiveness and insight of the data. Aiming at the problem of the fusion of the data-driven model and the mechanism-driven model, the method of the present invention performs data fusion at different scales at the dataset level to form a mechanism-data dual-driven fusion dataset. This method is applicable when integrating multiple models with similar performances, greatly improving the reliability and interpretability of the regulation decision, and can obtain a more comprehensive and optimized regulation scheme.

[0125] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also fall within the protection scope of the present invention.

Claims

1. A real-time regulation method for pumping stations driven by both data and mechanism, characterized in that: It includes the following steps: S1. Determination of key influencing factors: Combining the actual situation of on-site dispatching of the pumping station and the theoretical research on real-time optimal regulation and control of the pumping station, summarize the corresponding relationship between the water regime data of the pumping station and the regulation and control strategies, and determine the key influencing factors that determine the number of operating units of the pumping station; S2. Construction and solution of the mechanism-driven model: Based on the key influencing factors that determine the number of operating units of the pumping station and the constraint conditions of real-time regulation and operation of the pumping station, determine the entire feasible region of real-time regulation of the pumping station; construct a mechanism-driven model for real-time regulation of the pumping station by setting the objective function, decision variables, constraint conditions, and solution algorithm, and input the entire feasible region of real-time regulation of the pumping station into the mechanism-driven model for real-time regulation of the pumping station to obtain the distribution of the number of operating units of the pumping station under the entire feasible region, which is used as the solution of the mechanism-driven model; S3. Construction and solution of the data-driven model: Based on the historical operating water regime data and regulation and control decisions of the pumping station, determine the local feasible region of real-time regulation of the pumping station; construct a data-driven model for real-time regulation of the pumping station based on the decision tree classification algorithm, and input the local feasible region of real-time regulation of the pumping station into the data-driven model for real-time regulation of the pumping station to obtain the distribution of the number of operating units of the pumping station under the local feasible region, which is used as the solution of the data-driven model; S4. Construction and solution of the data-mechanism dual-driven model: Construct a fusion data set based on the solution of the mechanism-driven model and the solution of the data-driven model, train and verify the data-mechanism dual-driven model based on the decision tree constructed from the fusion data set, and obtain the final regulation and control decision result.

2. The real-time regulation method of the pumping station driven by data-mechanism dual drive according to claim 1, characterized in that: Step S1 specifically includes the following content: S101. Consult relevant literature on the optimal operation of the pumping station, determine the commonly used objective functions for the optimal operation of the pumping station, and determine the preliminary selected influencing factors according to the calculation formula of the objective function; S102. Combine the on-site requirements and consider the water regime and project situation information that affect the on-site regulation and control decisions; S103. Combine the literature research and on-site requirements to determine that the key influencing factors that determine the number of operating units of the pumping station include the flow rate through the pumping station and the head of the pumping station.

3. The real-time regulation method for a pumping station driven by both data and mechanism as claimed in claim 1, wherein: In step S2, determining the entire feasible region of real-time regulation of the pumping station specifically includes the following content: S201. Query the design parameters of the pumping station, including the minimum and maximum values of the key influencing factors that determine the number of operating units of the pumping station; S201. Discretize the key factors at a reasonable step size to generate a solution space; S201. According to the constraint conditions of real-time regulation and operation of the pumping station, eliminate the working conditions that do not meet the starting conditions to obtain the entire feasible region of real-time regulation of the pumping station.

4. The real-time regulation method of the pumping station driven by data-mechanism dual drive according to claim 3, characterized in that: The constraint conditions for real-time regulation and operation of the pumping station include the flow rate constraint of the pump unit, the head constraint of the pump unit, the downstream / upstream water level constraint of the pumping station during operation, the constraint of the number of operating units, and the pump flow-head characteristic curve constraint.

5. The real-time regulation method of the pumping station driven by both data and mechanism according to claim 1, characterized in that: The objective function of the mechanism-driven model for real-time regulation of the pumping station is the highest operating efficiency of the pumping station under the working conditions; The decision variable of the mechanism-driven model for real-time regulation of the pumping station is the number of operating units of the pumping station; The solution algorithm of the mechanism-driven model for real-time regulation of the pumping station is the particle swarm optimization algorithm.

6. The real-time regulation method for the pump station driven by both data and mechanism according to claim 1, characterized in that: In step S3, determining the local feasible region of real-time regulation of the pumping station specifically includes the following content: S3011. Collect the water regime data since the operation year of the pumping station and the regulation and control decision data since the operation year of the pumping station; all data are the scheduling results of the same time period; S3012. Analyze the distribution characteristics of the historical operation data of the pumping station. According to the minimum / maximum values of the flow rate and head of the operation data, float by a certain proportion to obtain the operable ranges of the local flow rate and head, and eliminate the working conditions that do not meet the starting conditions according to the operation constraint conditions to determine the real-time regulation local feasible region of the pumping station; S3013. Analyze the distribution characteristics of the historical operation data of the pumping station to determine the sample size and whether the distribution is balanced.

7. The real-time regulation method for a pumping station driven by both data and mechanism as claimed in claim 1, wherein: In step S3, the construction and solution of the real-time regulation data-driven model of the pumping station specifically include the following contents: S3021. Divide the preprocessed historical regulation data of the pumping station into a training set and a validation set according to a certain proportion; S3022. Build a real-time regulation data-driven model of the pumping station based on the decision tree classification algorithm; S3023. Train and validate the real-time regulation data-driven model of the pumping station based on the training set and the validation set, and evaluate the performance of the model based on the evaluation index; S3024. Input the real-time regulation local feasible region of the pumping station into the trained real-time regulation data-driven model of the pumping station to obtain the distribution of the number of operating units under the real-time regulation local feasible region of the pumping station as the solution of the data-driven model.

8. The real-time regulation method of the pumping station driven by data-mechanism dual drive according to claim 1, characterized in that: In step S4, the construction of the fusion data set specifically includes the following contents: S4011. Connect the solution of the mechanism-driven model and the solution of the data-driven model in series to form a fusion data set; S4012. Perform normalization processing on the fusion data set; S4013. Divide the normalized fusion data set into a training set and a validation set.

9. The real-time regulation method for a pumping station driven by both data and mechanism according to claim 1, characterized in that: In step S4, obtaining the final regulation decision result specifically includes the following contents: S4021. Use a decision tree to build a data-mechanism dual-driven model, input the training set into the data-mechanism dual-driven model for training, and input the validation set into the trained data-mechanism dual-driven model to output the model result; S4022. Perform inverse normalization processing on the model result to obtain the final regulation decision result.

10. The real-time regulation method of the pumping station driven by both data and mechanism according to claim 9, characterized in that: After step S4022, it also includes: S4023. Evaluate the reliability of the data-mechanism dual-driven model, and use the entire feasible region of the model as the input to calculate the global regulation decision result of the model.

Images