Water distribution system prediction control method
By collecting data in the water distribution system and training neural network models, establishing system simulation and abstract models, and building model prediction controllers, the problem of unintuitive and unsafe water distribution system control in the existing technology is solved, and simple and efficient prediction control is achieved.
Patent Information
- Application Number
- CN202510221225.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-27
- Publication Date
- 2025-06-03
AI Technical Summary
The existing water distribution system control methods cannot make intuitive and safe control decisions for the distribution system, and ignore the integrity of the system.
A water distribution system prediction control method is adopted, by collecting system data, training neural network models, establishing system simulation models and system abstract models, building model prediction controllers, and generating predictive control data.
This method does not require detailed physical models and expert experience, and can generate effective predictive models in complex dynamic systems, providing intuitive and secure predictive control data, simplifying the system modeling process.
Smart Images
Figure CN120085541A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of control of water distribution systems, and particularly to a predictive control method for water distribution systems. Background Art
[0002] A water distribution system is a complex dynamic system. Existing control methods for water distribution systems simply control the water distribution system based on the detection data of sensors, completely ignoring the integrity of the water distribution system, resulting in the inability to make intuitive and safe control decisions for the distribution system. Summary of the Invention
[0003] Based on this, the purpose of the present application is to provide a predictive control method for a water distribution system, which can overcome the deficiencies of the prior art.
[0004] To achieve the above purpose, the technical solution adopted by the present application is as follows:
[0005] A predictive control method for a water distribution system, comprising:
[0006] Collecting system data of the water distribution system at multiple moments, where the system data includes current state data of the water distribution system, current control data of the water distribution system, and next state data of the water distribution system;
[0007] Training a neural network model according to the current state data, the current control data, and the next state data to obtain a system simulation model of the water distribution system;
[0008] Dividing the current state data of multiple water distribution systems into multiple first state partitions; randomly selecting several current state data from each first state partition, and inputting the selected several current state data and the corresponding current control data into the system simulation model to obtain several simulated state data;
[0009] Covering the simulated state data with reachable sets, and establishing a reachable set table as a system abstraction model of the water distribution system according to the corresponding relationship between each first state partition and the reachable sets;
[0010] Constructing a model predictive controller according to the system simulation model and the system abstraction model, and inputting the real-time state data of the water distribution system into the model predictive controller to obtain predictive control data of the water distribution system.
[0011] As an implementation, the step of training a neural network model according to the current state data, the current control data, and the next state data to obtain a system simulation model of the water distribution system includes:
[0012] Input a plurality of the current state data and a plurality of the current control data into the neural network model to obtain a plurality of neural network model output values;
[0013] Construct a loss function according to each of the neural network model output values and the corresponding next state data;
[0014] Train the neural network model by minimizing the loss function to obtain the system simulation model.
[0015] As an implementation manner, the step of constructing a loss function according to each of the neural network model output values and the corresponding next state data includes:
[0016] Obtain the loss function through the following formula:
[0017]
[0018] where, is the output value of the loss function, N is the number of neural network model output values, ξ p,t+1 is the next state data, is the neural network model output value.
[0019] As an implementation manner, the step of using the reachable set to cover the simulated state data and establishing a reachable set table according to the corresponding relationship between each of the first state partitions and the reachable set as the system abstraction model of the water distribution system includes:
[0020] Divide the simulated state data into multiple second state partitions;
[0021] Obtain the reachable probability of each of the second state partitions according to the number of simulated state data of each of the second state partitions;
[0022] Exclude the second state partitions with reachable probabilities less than a preset probability threshold to obtain a reachable set covering the simulated state data of the remaining second state partitions.
[0023] As an implementation manner, the step of constructing a model predictive controller according to the system simulation model and the system abstraction model and inputting the real-time state data of the water distribution system into the model predictive controller to obtain the predictive control data of the water distribution system includes:
[0024] Construct a tracking constraint function according to the output of the system simulation model and the output of the system abstraction model;
[0025] Construct an input constraint function according to the input of the system simulation model and the input of the system abstraction model;
[0026] Construct an unsafe area avoidance constraint function based on the output of the system simulation model and the unsafe state;
[0027] Obtain the tracking phase cost function according to the tracking constraint function, the input constraint function, and the unsafe area avoidance constraint function;
[0028] Construct the model predictive controller according to the tracking phase cost function, the prediction horizon, and the terminal cost function.
[0029] As an implementation, the step of constructing the model predictive controller according to the tracking phase cost function, the prediction horizon, and the terminal cost function includes:
[0030] Obtain the model predictive controller through the following formula:
[0031]
[0032] where minJ(R t ,v t ) is the minimum value output by the model predictive controller, N MPC is the prediction horizon, l(R t ,v t ) is the tracking phase cost function, and l f (R t ) is the terminal cost function.
[0033] As an implementation, the step of constructing the tracking constraint function according to the output of the system simulation model and the output of the system abstraction model includes:
[0034] Obtain the tracking constraint function through the following formula:
[0035]
[0036] where l track is the tracking constraint function, N MPC is the prediction horizon, is the output of the system abstraction model, and R t is the output of the system simulation model.
[0037] As an implementation, the step of constructing the input constraint function according to the input of the system simulation model and the input of the system abstraction model includes:
[0038] Obtain the input constraint function through the following formula:
[0039]
[0040] where l input is the input constraint function, NMPC is the prediction horizon, v t+1 is the input of the system abstraction model, v t is the input of the system simulation model.
[0041] As an implementation manner, the step of constructing the unsafe region avoidance constraint function according to the output of the system simulation model and the unsafe state includes:
[0042] The unsafe region avoidance constraint function is obtained through the following formula:
[0043]
[0044] where l collision is the unsafe region avoidance constraint function, N MPC is the prediction horizon, is the output of the system abstraction model, R unsafe is the unsafe state, and d(·) is the distance between the output of the system abstraction model and the unsafe state.
[0045] As an implementation manner, the step of obtaining the tracking phase cost function according to the tracking constraint function, the input constraint function, and the unsafe region avoidance constraint function includes:
[0046] l(R t , v t ) = l track + l input + l collision
[0047] where l(R t , v t ) is the tracking phase cost function, l track is the tracking constraint function, l input is the input constraint function, and l collision is the unsafe region avoidance constraint function.
[0048] Compared with the traditional technology, the beneficial effects of the water distribution system predictive control method described in this application are:
[0049] This application does not require comprehensive system theory knowledge and a large amount of expert experience, but relies on a data-driven approach to train the dynamic relationship of the neural network system abstraction model. This data-driven method can generate an effective prediction model using limited data in the absence of a detailed physical model, solving the strong dependence on detailed physical models and expert experience in traditional system modeling methods, making the modeling process of complex dynamic systems simpler and more efficient, and enabling intuitive and safe predictive control data to be predicted.
[0050] For better understanding and implementation, the present application will be described in detail below with reference to the accompanying drawings. Description of the Drawings
[0051] Figure 1 It is a flowchart of the prediction control method for the water distribution system according to an embodiment of the present application;
[0052] Figure 2 It is a schematic diagram of the partition where the state space is divided into three dimensions according to an embodiment of the present application;
[0053] Figure 3 It is a schematic diagram of the partition numbers according to an embodiment of the present application;
[0054] Figure 4 It is a schematic diagram of the remaining second state partition according to an embodiment of the present application;
[0055] Figure 5 It is a schematic diagram of the control process of the water distribution system according to an embodiment of the present application. Detailed Embodiments
[0056] To make the objectives, technical solutions and advantages of the present application clearer, the embodiments of the present application will be further described in detail below with reference to the accompanying drawings.
[0057] It should be clear that the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts shall fall within the scope of protection of the embodiments of the present application.
[0058] When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. In the description of the present application, it should be understood that the terms "first", "second", "third", etc. are only used to distinguish similar objects, and do not have to be used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances. The singular forms of "a", "the" and "said" used in the present application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. The words "if" / "when" used herein can be interpreted as "when...", "when...", or "in response to a determination".
[0059] In addition, in the description of this application, unless otherwise specified, "a plurality of" means two or more. "And / or" describes the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, both A and B exist simultaneously, and B exists alone. The character " / " generally represents an "or" relationship between the associated objects before and after.
[0060] Please refer to Figure 1 , which is a flowchart of a predicted control method for a water distribution system according to an embodiment of this application, including:
[0061] S1: Collect system data of the water distribution system at multiple moments. The system data includes the current state data of the water distribution system, the current control data of the water distribution system, and the next state data of the water distribution system.
[0062] Among them, the system data is data of discrete time, and its representation form is as follows:
[0063]
[0064] ξ p,t represents the state of the system at the current moment, ξ p,t+1 represents the state of the system at the next moment, v t represents the input of the system at the current moment, ω t represents the disturbance of the system at the current moment, f p is a continuous function representing the state transition function of the system, y t represents the output of the system at the current moment, g p represents the output mapping. Usually, f p is difficult to obtain, especially for those systems with complex dynamics. Therefore, it is assumed that in this work, f p is unknown and there is no prior knowledge of f p ; at the same time, it is assumed that the system state ξ p (ξ p,t and ξ p,t+1 ) and the input v t can be directly obtained through certain sensors. Although the dynamics of the system are considered unknown, its trajectory is accessible and can be sampled. The data collected in system P is expressed in the following form:
[0065] S2: Train a neural network model according to the current state data, the current control data, and the next state data to obtain a system simulation model of the water distribution system.
[0066] The neural network model adopted in this application is an LSTM neural network. Let ξ p,t+1 and (ξ p,t , vt ) Respectively as the output and input of the LSTM neural network to train the LSTM neural network.
[0067] Among them, step S2 includes:
[0068] S21: Input a number of the current state data and a number of the current control data into the neural network model to obtain a number of neural network model output values.
[0069] S22: Construct a loss function according to each neural network model output value and the corresponding next state data:
[0070] Obtain the loss function through the following formula:
[0071]
[0072] Among them, is the output value of the loss function, N is the number of neural network model output values, ξ p,t+1 is the next state data, is the neural network model output value.
[0073] S23: Train the neural network model by minimizing the loss function to obtain the system simulation model.
[0074] S3: Divide the current state data of multiple water distribution systems into multiple first state partitions; randomly select a number of current state data from each first state partition, and input the selected number of current state data and the corresponding current control data into the system simulation model to obtain a number of simulated state data.
[0075] S4: Use the reachable set to cover the simulated state data, and establish a reachable set table according to the corresponding relationship between each first state partition and the reachable set as the system abstraction model of the water distribution system.
[0076] Among them, the system simulation model and the system abstraction model need to meet the following conditions: (In the conditions, is the system simulation model and is the system abstraction model. Let
[0077] (1) Make Among them and represent a finite interval (set) of the system simulation model and the system abstraction model. This formula means that when taking a point in a certain interval of the system simulation model, there is a similar point in this interval of the system abstraction model, and the distance is very close, satisfying
[0078] (2) For any that conforms to two points, if there is ξ in the system p,t+1 = f p (ξ p,t , v t ), then is in and satisfies
[0079] (3) For two systems and in the complete state space, it satisfies
[0080] (4) For two systems and in the complete state space and the allowable input space, it satisfies P
[0081]
[0082] When , the obtained relationship is called the exact simulation relationship.
[0083] Therefore, in step S3, several selected current state data need to satisfy where 1 , β is a custom parameter, and both are usually relatively small values. However, in order to ensure the calculation speed, they should not be taken too small.
[0084] Among them, the Monte Carlo method can also be introduced into the neural network to ensure that the probability distribution of the distance meets the requirements
[0085] Step S4 includes:
[0086] S41: Divide the simulated state data into multiple second state partitions.
[0087] R i = {ξ p |PGξ p - c i P ∞ < 1}
[0088] where G is a symmetric matrix and c i is the set center. The function of G is to perform a linear transformation on the points in the finite state space and map them to a new coordinate system. Different regions have different centers but share the same matrix G, R i , i = {1, 2,..., M}, M ∈ N ZPartitions that make up X, which are disjoint from each other, can thus be expressed as:
[0089]
[0090] where N Z represents a positive integer, X represents the complete state space (finite) of the system, and R i represents the i-th finite partition of the complete state space X. The above formula indicates that the complete state space of the system can be divided into several disjoint finite partitions. In addition, the input space is discretized in the same way.
[0091] The complete set of allowed inputs can be represented by the set V, where V i represents the i-th finite partition in the complete set of allowed inputs and is expressed as:
[0092]
[0093] Please refer to Figure 2 to divide the complete state space into multiple three-dimensional partitions, and the number of each partition is as shown in Figure 3 .
[0094] S42: Obtain the reachability probability of each of the second state partitions according to the number of simulated state data of each of the second state partitions.
[0095] Among them, the quotient of the number of simulated state data of each of the second state partitions divided by the total number of simulated state data is the reachability probability of each of the second state partitions.
[0096] S43: Exclude the second state partitions whose reachability probability is less than a preset probability threshold to obtain a reachable set covering the simulated state data of the remaining second state partitions.
[0097] Through step S43, a reachable set with the second state partitions whose reachability probability is less than the preset probability threshold trimmed can be obtained.
[0098] Step S43 can be expressed by the following formula:
[0099]
[0100] where represents the set of all results obtained when the i-th finite partition R i of the state is input under the j-th finite partition V i of the input.
[0101] Among them, after excluding the second state partitions whose probability is less than the preset probability threshold, the remaining second state partitions are as shown in Figure 4As shown. In the figure, the gray-filled part in the upper right corner is the excluded second state partition, and the other parts are the remaining second state partitions.
[0102] S5: Construct a model predictive controller according to the system simulation model and the system abstraction model, and input the real-time state data of the water distribution system into the model predictive controller to obtain the predictive control data of the water distribution system.
[0103] Among them, step S5 includes:
[0104] S51: Construct a tracking constraint function according to the output of the system simulation model and the output of the system abstraction model.
[0105] The tracking constraint function is obtained through the following formula:
[0106]
[0107] Among them, l track is the tracking constraint function, N MPC is the prediction horizon, is the output of the system abstraction model, R t is the output of the system simulation model.
[0108] S52: Construct an input constraint function according to the input of the system simulation model and the input of the system abstraction model.
[0109] The input constraint function is obtained through the following formula:
[0110]
[0111] Among them, l input is the input constraint function, N MPC is the prediction horizon, v t+1 is the input of the system abstraction model, v t is the input of the system simulation model.
[0112] S53: Construct an unsafe region avoidance constraint function according to the output of the system simulation model and the unsafe state.
[0113] The unsafe region avoidance constraint function is obtained through the following formula:
[0114]
[0115] Among them, l collision is the unsafe region avoidance constraint function, N MPC is the prediction horizon, is the output of the system abstraction model, R unsafeis an unsafe state, and d(·) is the distance between the output of the system abstraction model and the unsafe state.
[0116] S54: Obtain the tracking phase cost function according to the tracking constraint function, the input constraint function, and the unsafe region avoidance constraint function:
[0117] Obtain the model predictive controller through the following formula:
[0118]
[0119] where minJ(R t , v t ) is the minimum value output by the model predictive controller, N MPC is the prediction horizon, l(R t , v t ) is the tracking phase cost function, and l f (R t ) is the terminal cost function.
[0120] S55: Construct the model predictive controller according to the tracking phase cost function, the prediction horizon, and the terminal cost function.
[0121] Obtain the model predictive controller through the following formula:
[0122] l(R t , v t ) = l track + l input + l collision
[0123] where l(R t , v t ) is the tracking phase cost function, l track is the tracking constraint function, l input is the input constraint function, and l collision is the unsafe region avoidance constraint function.
[0124] Please refer to Figure 5 , according to all the steps of this application, the control process of this application for the water distribution system can be obtained as Figure 5 shown: First, collect limited data in the original system, and use this data to establish a learning model through a neural network to capture the dynamic relationship of the system. Second, build an abstraction model with the help of this learning model to improve the interpretability of the system in the form of a reachable set table, and the system constraints can be represented in the abstraction model through boundaries and reachability conditions. Finally, design a model predictive control method based on the abstraction model to find the optimal control input.
[0125] Compared with the prior art, the present application does not require comprehensive system theory knowledge and a large amount of expert experience. Instead, it relies on a data-driven approach to abstract the dynamic relationships of the neural network system model through training. This data-driven method can generate an effective prediction model using limited data in the absence of a detailed physical model, solving the strong dependence on detailed physical models and expert experience in traditional system modeling methods, making the modeling process of complex dynamic systems simpler and more efficient, and enabling the prediction of intuitive and safe predictive control data. Moreover, the present application also considers the impact of system constraints on predictive control data and can output predictive control data with constraints, which can be used to obtain the optimal predictive control data for the real-time state data of the water distribution system.
[0126] The device embodiments described above are merely illustrative. The components described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the present application solution. A person of ordinary skill in the art can understand and implement it without creative work.
[0127] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0128] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the selected functions in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device that implements the functions in the flowFigure 1 one process or multiple processes and / or blocks Figure 1 selected functions in one block or multiple blocks.
[0129] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the selected functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.
[0130] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and a memory.
[0131] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM) and / or non-volatile memory such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer-readable media.
[0132] Computer-readable media includes permanent and non-permanent, removable and non-removable media and can store information by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.
[0133] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, commodity or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or also includes elements inherent in such process, method, commodity or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the existence of another identical element in the process, method, commodity or device comprising the element.
[0134] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.
Claims
1. A predictive control method for a water distribution system, characterized in that: include: Collecting system data of the water distribution system at multiple times, the system data including current state data of the water distribution system, current control data of the water distribution system and next state data of the water distribution system; Training a neural network model according to the current state data, the current control data and the next state data to obtain a system simulation model of the water distribution system; Dividing the current state data of the plurality of water distribution systems into a plurality of first state partitions; randomly selecting a plurality of current state data from each first state partition, inputting the selected plurality of current state data and corresponding current control data into the system simulation model to obtain a plurality of simulated state data; Using reachable sets to cover the simulation state data, and establishing a reachable set table as a system abstract model of the water distribution system according to the corresponding relationship between each of the first state partitions and the reachable sets; A model predictive controller is constructed according to the system simulation model and the system abstract model, and the real-time status data of the water distribution system is input into the model predictive controller to obtain predictive control data of the water distribution system.
2. The water distribution system predictive control method according to claim 1, characterized in that: The step of training a neural network model according to the current state data, the current control data and the next state data to obtain a system simulation model of the water distribution system comprises: Inputting a plurality of the current state data and a plurality of the current control data into the neural network model to obtain a plurality of neural network model output values; Constructing a loss function according to each of the neural network model output values and the corresponding next state data; The neural network model is trained by minimizing the loss function to obtain the system simulation model.
3. The water distribution system predictive control method according to claim 2, characterized in that: The step of constructing a loss function according to each of the neural network model output values and the corresponding next state data comprises: The loss function is obtained by the following formula: in, is the output value of the loss function, N is the number of output values of the neural network model, ξ p,t+1 is the next state data, Output values for the neural network model.
4. The water distribution system predictive control method according to claim 1, characterized in that: The step of using a reachable set to cover the simulation state data and establishing a reachable set table as a system abstract model of the water distribution system according to the corresponding relationship between each of the first state partitions and the reachable set includes: dividing the simulation state data into a plurality of second state partitions; Obtaining a reachability probability of each second-state partition according to the amount of simulation state data of each second-state partition; The second state partitions whose reachability probability is less than a preset probability threshold are excluded to obtain a reachable set of simulation state data covering the remaining second state partitions.
5. The water distribution system predictive control method according to claim 1, characterized in that: The step of constructing a model predictive controller according to the system simulation model and the system abstract model, inputting the real-time status data of the water distribution system into the model predictive controller, and obtaining predictive control data of the water distribution system comprises: Constructing a tracking constraint function according to the output of the system simulation model and the output of the system abstract model; constructing an input constraint function according to the input of the system simulation model and the input of the system abstract model; Constructing an unsafe area avoidance constraint function according to the output of the system simulation model and the unsafe state; Obtaining a tracking phase cost function according to the tracking constraint function, the input constraint function and the unsafe area avoidance constraint function; The model predictive controller is constructed according to the tracking phase cost function, the prediction time domain and the terminal cost function.
6. The water distribution system predictive control method according to claim 5, characterized in that: The step of constructing the model predictive controller according to the tracking phase cost function, the prediction time domain and the terminal cost function comprises: The model predictive controller is obtained by the following formula: Among them, minJ(R t ,v t ) is the minimum value of the model predictive controller output, N MPC is the prediction time domain, l(R t ,v t ) is the cost function of the tracking stage, l f (R t ) is the terminal cost function.
7. The water distribution system predictive control method according to claim 6, characterized in that: The step of constructing a tracking constraint function according to the output of the system simulation model and the output of the system abstract model comprises: The tracking constraint function is obtained through the following formula: Among them, l track is the tracking constraint function, N MPC For the prediction time domain, is the output of the system abstract model, R t It is the output of the system simulation model.
8. The water distribution system predictive control method according to claim 6, characterized in that: The step of constructing an input constraint function according to the input of the system simulation model and the input of the system abstract model comprises: The input constraint function is obtained through the following formula: Among them, l input is the input constraint function, N MPC is the prediction time domain, v t+1 is the input of the system abstract model, v t It is the input of the system simulation model.
9. The water distribution system predictive control method according to claim 6, characterized in that: The step of constructing an unsafe area avoidance constraint function according to the output of the system simulation model and the unsafe state comprises: The unsafe area avoidance constraint function is obtained through the following formula: Among them, l collision is the unsafe area avoidance constraint function, N MPC For the prediction time domain, is the output of the system abstract model, R unsafe is the unsafe state, and d(·) is the distance between the output of the system abstract model and the unsafe state.
10. The water distribution system predictive control method according to claim 6, characterized in that: The step of obtaining a tracking phase cost function according to the tracking constraint function, the input constraint function and the unsafe area avoidance constraint function comprises: l(R t ,v t )=l track +l input +l collision Among them, l(R t ,v t ) is the cost function of the tracking stage, l track is the tracking constraint function, l input is the input constraint function, l collision Avoid constraint functions for unsafe areas.