Irrigation pipeline system optimization design method based on intelligent algorithm

Abstract

The invention discloses an irrigation pipeline system optimization design method based on an intelligent algorithm, and relates to the field of irrigation pipeline optimization design. Pipe network nodes of an irrigation pipeline system are classified into water source nodes and water demand nodes, a state set of a pipe network arrangement form is defined as S, a node pipeline connection action is defined as a, the state of the pipe network arrangement form is defined as st after the t-th pipeline connection action at is conducted, and for the state st, the probability of obtaining a new state st + 1 of the pipe network arrangement form after the action at is executed is defined as P (st + 1st, at). An executable action set is A (st), a pipe network arrangement decision strategy pi is constructed by using an MDP algorithm, an expected return obtained after the action at is executed in a state st is calculated according to a Bellman equation, and the execution action at corresponding to the maximum expected return is obtained to serve as a node pipeline connection action in the state st, and executing the node pipeline connection action at in the state st until any one of the water-requiring nodes has a connection relationship, and obtaining a pipe network arrangement form of the irrigation pipeline system.

Description

Technical Field

[0001] The invention relates to the field of irrigation pipeline optimization design, and in particular to an irrigation pipeline system optimization design method based on an intelligent algorithm. Background Art

[0002] Farm irrigation systems are generally composed of three parts: water source projects, backbone water supply networks, and field water distribution networks. Among them, the investment in water supply and distribution networks generally accounts for 70% of the total system investment. Whether the planning and design of this part is reasonable will directly affect the irrigation efficiency and the safety and reliability of operation. Irrigation network optimization refers to the search for the lowest cost, energy-saving and safest design solution with the help of relevant optimization theory and computer technology under the condition of meeting specific constraints of the network (including node pressure, design flow, etc.), so as to achieve the optimal or relatively optimal system goals.

[0003] At present, for the optimization design of agricultural irrigation pipe networks, most of them use intelligent algorithms such as enumeration method, genetic algorithm, simulated annealing algorithm, artificial neural network, etc. The above design methods usually seek the optimal solution based on the constraints that restrict the optimization of pipe irrigation system, such as investment cost, optimal pipe diameter, shortest pipe, and minimum energy consumption, so as to obtain the global optimal solution or local optimal solution for the entire irrigation pipe system. However, it is difficult to implement or the accuracy is not enough for multi-objective optimization calculation of complex pipe systems. Therefore, we propose an optimization design method for irrigation pipe system based on intelligent algorithm. Summary of the invention

[0004] The main purpose of the present invention is to provide an irrigation pipe system optimization design method based on an intelligent algorithm, which can effectively solve the problems in the background technology.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] An irrigation pipe system optimization design method based on intelligent algorithm, comprising:

[0007] The network nodes of the irrigation pipeline system are classified into water source nodes and water demand nodes, wherein the water demand nodes are divided into intermediate water demand nodes with irrigation water inflow and outflow and terminal water demand nodes with only irrigation water inflow according to the situation of irrigation water flowing through the nodes; wherein, for any of the water demand nodes, only one node pipeline connection action is performed in the irrigation water inflow direction, and the number of the terminal water demand nodes is at least one

[0008] Define the state set of the pipe network layout as S, the node pipe connection action as a, and perform the pipe connection action a for the tth time t The state of the rear pipe network layout is s t , for state s t , perform action at Then obtain the new state s of the pipe network layout t+1 The probability of P(s t+1 s t ,a t ), the set of executable actions is A(s t );

[0009] The MDP algorithm is used to construct the pipeline network layout decision strategy π, and the Bellman equation is used to calculate the t Next, perform action a t The expected return obtained after that is used to obtain the execution action a corresponding to the maximum expected return. t As in state t The node pipeline connection action below; wherein the Bellman equation defines the expected return as follows:

[0010]

[0011] In the formula, Q(s t ,a t ) is represented as in state s t Next, perform action a t The expected return after γ is the discount factor, and γ∈(0,1); R(s t ,a t ) represents the pipe network layout in state s t Next, perform action a t The reward value obtained; π(a t+ 1s t+1 ) is represented as in state s t+1 Next, perform action a t+1 probability.

[0012] The layout of the pipe network is in state s t Next, perform action a t The reward value R(s) obtained t ,a t ) is calculated as:

[0013]

[0014] In the formula, Indicated as in state s t Next, perform action a t The pipe network layout cost function; Indicated as in state s t Next, perform action a t The pipe network layout length function; μ and ν are constant coefficients, and μ+ν=1, 0<μ<1, 0<ν<1.

[0015] In status tNext, perform action a t The expressions of the pipe network layout cost function and pipe network layout length function are:

[0016]

[0017] In the formula, α is the equalization factor, e is the benchmark rate of return, %; λ is the economic calculation period of the project; p is the maintenance rate, %; E1 is the unit kilowatt cost of the pump station, yuan / kW; ε is the reserve coefficient of the pump station unit; Q i is the flow rate of the pump station, L / s; H i is the head of the water distribution pump station, m; η i is the efficiency of the water distribution pump station, %; k is the economic effect index of the pump station, and 0<k<1; f(D i ) is the unit length cost of the pipe section, RMB; D i is the diameter of the pipe section, m; L i is the length of the pipe section, m; E is the electricity price, yuan / kW·h; T is the working time of the pump station, h; n is the number of pipe sections; To perform action a t The length of the pipe section; n y It is the number of water demand nodes that have been connected to the node pipeline.

[0018] Execution in state s t The node pipeline connection action a t A connection relationship exists to any of the water demand nodes, and a pipe network layout of the irrigation pipe system is obtained.

[0019] The present invention has the following beneficial effects:

[0020] Compared with the prior art, the network nodes of the irrigation pipeline system are classified into water source nodes and water demand nodes, wherein the water demand nodes are divided into intermediate water demand nodes with irrigation water inflow and outflow and terminal water demand nodes with only irrigation water inflow according to the situation of irrigation water flowing through the nodes, the state set of the pipeline layout form is defined as S, the node pipeline connection action is a, and the pipeline connection action a is performed at the tth time. t The state of the rear pipe network layout is s t , for state s t , perform action a t Then obtain the new state s of the pipe network layout t+1 The probability of P(s t+1 s t ,a t ), the set of executable actions is A(s t ), using the MDP algorithm to construct the pipeline network layout decision strategy π, and according to the Bellman equation to calculate the state s t Next, perform action at The expected return obtained after that is used to obtain the execution action a corresponding to the maximum expected return. t As in state t The node pipeline connection action under is executed in state s t The node pipeline connection action a t There is a connection relationship with any of the water demand nodes, and the pipe network layout of the irrigation pipe system is obtained. It can seek the best solution from multiple objectives to obtain the global optimal solution or local optimal solution for the irrigation pipe system. It is suitable for multi-objective optimization calculation of complex pipe systems and improves the accuracy of multi-objective optimization calculation of pipe systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flowchart of an irrigation pipe system optimization design method based on an intelligent algorithm of the present invention;

[0022] Figure 2 It is a schematic diagram of all node-pipeline connection actions and connection structures that can be implemented in the embodiment. DETAILED DESCRIPTION

[0023] The present invention will be further described below in conjunction with specific implementation methods, wherein the accompanying drawings are only used for exemplary descriptions and represent only schematic diagrams rather than actual drawings, and should not be understood as limiting the present invention. In order to better illustrate the specific implementation methods of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product.

[0024] An optimization design method for irrigation pipe system based on intelligent algorithm.

[0025] The network nodes of the irrigation pipeline system are classified into water source nodes and water demand nodes, wherein the water demand nodes are divided into intermediate water demand nodes with irrigation water inflow and outflow and terminal water demand nodes with only irrigation water inflow according to the situation of irrigation water flowing through the nodes; wherein, for any water demand node, only one node pipeline connection action is performed in the irrigation water inflow direction, and the number of terminal water demand nodes is at least one;

[0026] Define the state set of the pipe network layout as S, the node pipe connection action as a, and perform the pipe connection action a for the tth time t The state of the rear pipe network layout is s t , for state s t , perform action a t Then obtain the new state s of the pipe network layout t+1 The probability of P(s t+1 s t ,a t ), the set of executable actions is A(s t );

[0027] The MDP algorithm is used to construct the pipeline network layout decision strategy π, and the Bellman equation is used to calculate the t Next, perform action a t The expected return obtained after that is used to obtain the execution action a corresponding to the maximum expected return. t As in state t The node pipeline below connects the actions; the Bellman equation defines the expected return as follows:

[0028]

[0029] In the formula, Q(s t ,a t ) is represented as in state s t Next, perform action a t The expected return after γ is the discount factor, and γ∈(0,1); R(s t ,a t ) represents the pipe network layout in state s t Next, perform action a t The reward value obtained; π(a t+ 1s t+1 ) is represented as in state s t+1 Next, perform action a t+1 probability.

[0030] The layout of the pipe network is in state s t Next, perform action a t The reward value R(s) obtained t ,a t ) is calculated as:

[0031]

[0032] In the formula, Indicated as in state s t Next, perform action a t The pipe network layout cost function; Indicated as in state s t Next, perform action a t The pipe network layout length function; μ and ν are constant coefficients, and μ+ν=1, 0<μ<1, 0<ν<1.

[0033] In status t Next, perform action a t The expressions of the pipe network layout cost function and pipe network layout length function are:

[0034]

[0035] In the formula, α is the equalization factor, e is the benchmark rate of return, %; λ is the economic calculation period of the project; p is the maintenance rate, %; E1 is the unit kilowatt cost of the pump station, yuan / kW; ε is the reserve coefficient of the pump station unit; Q i is the flow rate of the pump station, L / s; H i is the head of the water distribution pump station, m; η i is the efficiency of the water distribution pump station, %; k is the economic effect index of the pump station, and 0<k<1; f(D i ) is the unit length cost of the pipe section, RMB; D i is the diameter of the pipe section, m; L i is the length of the pipe section, m; E is the electricity price, yuan / kW·h; T is the working time of the pump station, h; n is the number of pipe sections; To perform action a t The length of the pipe section; n y It is the number of water demand nodes that have been connected to the node pipeline.

[0036] Execution in state s t The node pipeline connection action a t There is a connection relationship to any water demand node, and the pipe network layout of the irrigation pipe system is obtained.

[0037] In this embodiment, the technical solution is described by assuming that the number of water source nodes is 1 and the number of water demand nodes is 5. Figure 2 As shown,

[0038] When t=1, the state of the pipe network layout is s1, which is the node pipe connection action starting from the water source node. For state s1, the executable action set includes: water source node → water demand node 1; water source node → water demand node 2; water source node → water demand node 3; water source node → water demand node 4; water source node → water demand node 5; water source node → water demand node 1 and water demand node 2; water source node → water demand node 1 and water demand node 3; water source node → water demand node 1 and water demand node 4; water source node → water demand node 1 and water demand node 5; water source node → water demand node 2 and water demand node 3; water source node → water demand node 2 and water demand node 4; water source node → water demand node 2 and water demand node 5; water source node → water demand node 1, water demand node 2 and water demand node 3; water source node → water demand node 1, water demand node 2 and water demand node 4; water source node → water demand node 1, water demand node 2 and water demand node 5; water source node → water demand node 2, water demand node 3 and water demand node 4; water source node → water demand node 2, water demand node 3 and water demand node 5; water source node → water demand node 3, water demand node 4 and water demand node 5; water source node → water demand node 1, water demand node 2, water demand node 3 and water demand node 4; water source node → water demand node 1, water demand node 2, water demand node 3 and water demand node 5; water source node → water demand node 1, water demand node 2, water demand node 3, water demand node 4 and water demand node 5; water source node → water demand node 1, water demand node 2, water demand node 3, water demand node 4 and water demand node 5;

[0039] If external constraints are not considered, for state s1, the probability of executing any node pipeline connection action in the above action set is the same. Since only one node pipeline connection action is performed for any water demand node in the irrigation water inflow direction, when the execution action is water source node → water demand node 1, water demand node 2, water demand node 3, water demand node 4 and water demand node 5, at this time, water demand node 1-water demand node 5 all become terminal water demand nodes, and the pipe network layout decision strategy π is water source node → water demand node 1, water demand node 2, water demand node 3, water demand node 4 and water demand node 5, thereby obtaining the pipe network layout form of the irrigation pipe system;

[0040] When the execution action is water source node → water demand node 1, at this time, for state s2, since water demand node 1 has performed a node pipe connection action in the irrigation water inflow direction, it can no longer be used as an irrigation water inflow node, but can be used as an irrigation water outflow node. The executable action set includes: water demand node 1 → water demand node 2; water demand node 1 → water demand node 3; water demand node 1 → water demand node 4; water demand node 1 → water demand node 5; water demand node 1 → water demand node 2 and water demand node 3; water demand node 1 → water demand node 2 and water demand node 4; water demand node 1 → water demand node 2 and water demand node 5; water demand node Water node 1→water demand node 3 and water demand node 4; water demand node 1→water demand node 3 and water demand node 5; water demand node 1→water demand node 4 and water demand node 5; water demand node 1→water demand node 2, water demand node 3 and water demand node 4; water demand node 1→water demand node 2, water demand node 3 and water demand node 5; water demand node 1→water demand node 3, water demand node 4 and water demand node 5; water demand node 1→water demand node 2, water demand node 3, water demand node 4 and water demand node 5; Similarly, if no external constraints are considered, for state s2, the probability of executing any node pipeline connection action in the above action set is the same;

[0041] When the execution action is water source node → water demand node 1 → water demand node 2, at this time, for state s3, since water demand node 2 has performed a node pipe connection action in the irrigation water inflow direction, it can no longer be used as an irrigation water inflow node, but can be used as an irrigation water outflow node. The executable action set includes: water demand node 2 → water demand node 3; water demand node 2 → water demand node 4; water demand node 2 → water demand node 5; water demand node 2 → water demand node 3 and water demand node 4; water demand node 2 → water demand node 3 and water demand node 5; water demand node 2 → water demand node 4 and water demand node 5; water demand node 2 → water demand node 3, water demand node 4 and water demand node 5;

[0042] When the execution action is water source node → water demand node 1 → water demand node 2 → water demand node 3, at this time, for state s4, since water demand node 3 has performed a node pipe connection action in the irrigation water inflow direction, it can no longer be used as an irrigation water inflow node, but can be used as an irrigation water outflow node. The executable action set includes: water demand node 3 → water demand node 4; water demand node 3 → water demand node 5; water demand node 3 → water demand node 4 and water demand node 5;

[0043] When the execution action is water source node → water demand node 1 → water demand node 2 → water demand node 3 → water demand node 4, at this time, for state s5, since water demand node 4 has performed a node pipe connection action in the irrigation water inflow direction, it can no longer be used as an irrigation water inflow node, but can be used as an irrigation water outflow node, then the executable action set only includes: water demand node 4 → water demand node 5; at this time, water demand nodes 1-water demand node 4 are all intermediate water demand nodes with irrigation water inflow and outflow, and water demand node 5 is the terminal water demand node with only irrigation water inflow;

[0044] According to the above connection action principle, any water demand node has a connection relationship, and a pipe network layout form of the irrigation pipe system can be obtained, that is, water source node → water demand node 1 → water demand node 2 → water demand node 3 → water demand node 4 → water demand node 5;

[0045] As mentioned above, although there are many different pipeline layouts, for any state s t For example, when the number of water demand nodes is constant, without external constraints, the set of executable actions is a certain value, and in state s t Next, perform action a t After that, the new state s of the pipe network layout is obtained t+1 The probability P(s t+1 s t ,a t ) and in state s t+1 Next, perform action a t+1 The probability of can be determined. Therefore, in order to determine the best layout strategy, the pipeline network layout decision strategy π is constructed by using the MDP algorithm. According to the Bellman equation, the pipeline network layout decision strategy π is calculated in state s. t Next, perform action a t The expected return obtained after that is used to obtain the execution action a corresponding to the maximum expected return. t As in state t The node pipeline connection action under state s is calculated by the pipeline network layout cost function and the pipeline network layout length function. t Next, perform action a t The reward value R(s) obtained t ,a t ), the reward value R(s) t ,a t ) is brought into the Bellman equation to calculate the state s t Next, perform action a t The expected return obtained after that, the execution action a corresponding to the maximum expected return t That is, in state s t The best node pipeline connection action under;

[0046] In addition, it should be noted that when there is a constraint relationship between water demand nodes, such as in actual situations, if water demand node 1 cannot perform a pipeline connection action with water demand node 2 through a pipeline, then the pipeline connection action with water demand node 1→water demand node 2 can be deleted from the executable action set to redefine the executable action set and the corresponding probability of executing the action.

[0047] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An irrigation pipe system optimization design method based on intelligent algorithm, characterized in that: include: Classifying the network nodes of the irrigation pipeline system into water source nodes and water demand nodes, wherein the water demand nodes are divided into intermediate water demand nodes with irrigation water inflow and outflow and terminal water demand nodes with only irrigation water inflow according to the situation of irrigation water flowing through the nodes; Define the state set of the pipe network layout as S, the node pipe connection action as a, and perform the pipe connection action a for the tth time t The state of the rear pipe network layout is s t , for state s t , perform action a t Then obtain the new state s of the pipe network layout t+1 The probability of P(s t+1 s t ,a t ), the set of executable actions is A(s t ); The MDP algorithm is used to construct the pipeline network layout decision strategy π, and the Bellman equation is used to calculate the t Next, perform action a t The expected return obtained after that is used to obtain the execution action a corresponding to the maximum expected return. t As in state t The node pipeline connection action below; Execution in state s t The node pipeline connection action a t A connection relationship exists to any of the water demand nodes, and a pipe network layout of the irrigation pipe system is obtained.

2. The irrigation pipe system optimization design method based on intelligent algorithm according to claim 1 is characterized by: The Bellman equation defines the expected return as follows: In the formula, Q(s t ,a t ) is represented as in state s t Next, perform action a t The expected return after γ is the discount factor, and γ∈(0,1); R(s t ,a t ) represents the pipe network layout in state s t Next, perform action a t The reward value obtained; π(a t+1 s t+1 ) is represented as in state s t+1 Next, perform action a t+1 probability.

3. The irrigation pipe system optimization design method based on intelligent algorithm according to claim 2 is characterized by: The layout of the pipe network is in state s t Next, perform action a t The reward value R(s) obtained t ,a t ) is calculated as: In the formula, Indicated as in state s t Next, perform action a t The pipe network layout cost function; Indicated as in state s t Next, perform action a t The pipe network layout length function; μ and ν are constant coefficients, and μ+ν=1, 0<μ<1, 0<ν<1.

4. The irrigation pipe system optimization design method based on intelligent algorithm according to claim 3 is characterized by: In status t Next, perform action a t The expressions of the pipe network layout cost function and pipe network layout length function are: In the formula, α is the equalization factor, e is the benchmark rate of return, %; λ is the economic calculation period of the project; p is the maintenance rate, %; E1 is the unit kilowatt cost of the pump station, yuan / kW; ε is the reserve coefficient of the pump station unit; Q i is the flow rate of the pump station, L / s; H i is the head of the water distribution pump station, m; η i is the efficiency of the water distribution pump station, %; k is the economic effect index of the pump station, and 0<k<1; f(D i ) is the unit length cost of the pipe section, RMB; D i is the diameter of the pipe section, m; L i is the length of the pipe section, m; E is the electricity price, yuan / kW·h; T is the working time of the pump station, h; n is the number of pipe sections; To perform action a t The length of the pipe section; n y It is the number of water demand nodes that have been connected to the node pipeline.

5. The irrigation pipe system optimization design method based on intelligent algorithm according to claim 1 is characterized by: For any of the water demanding nodes, only one node pipe connection action is performed in the irrigation water inflow direction, and the number of the terminal water demanding nodes is at least one.

Images