A multi-period scheduling optimization method for gas compression stations

Abstract

The present invention discloses a multi-period scheduling optimization method for a gas compressor station, including: designing a compressor station scheduling optimization module based on a two-level hierarchical structure, the module completes the establishment of a compressor start-stop decision model with the minimum gas compressor station operation cost as the objective function, and the establishment of a compressor load distribution model with the minimum compressor energy consumption as the objective function; designing a compressor station scheduling optimization solution module based on DE-GWO and Gurobi, designing a DE-GWO optimization algorithm to solve the first-stage start-stop decision model, calling the Gurobi solver to solve the second-stage load distribution model, and finally obtaining a gas compressor station multi-period scheduling optimization model. The present invention proposes an optimization model based on a two-level hierarchical structure for the mixed integer nonlinear programming problem of gas compressor station scheduling optimization, and designs a solution method for the two proposed stages, which effectively improves the compressor operation efficiency and reduces the operation cost of the compressor station.

Description

Technical Field

[0001] The invention belongs to the field of scheduling optimization and energy, and specifically relates to a multi-period scheduling optimization method for a gas compressor station. Background Art

[0002] In order to meet the growing gas demand of urban residents, businesses, factories and other customers, the construction of gas transportation-related infrastructure has been vigorously promoted in recent years. By the end of 2020, the total mileage of long-distance gas pipelines has increased by 57.38% compared with 2012. In thousands of kilometers of long-distance pipelines, multiple compressor stations will be set up to boost pressure and provide power to maintain the pipeline pressure within the required range. The gas compressor station is mainly composed of multiple compressors in parallel, and the compressor is a high-energy consumption equipment. With the continuous increase in pipeline mileage, the number of compressor stations and compressors is also increasing, and the energy consumption cost is rising. At present, the operating energy consumption cost of compressors accounts for 45% of the operating cost of professional companies. Therefore, how to formulate a transmission and distribution scheduling strategy to reduce the operating cost of compressors has also become a key focus of gas suppliers. At present, the optimization methods for gas compressor station scheduling mainly include mixed integer nonlinear programming (MINLP), dynamic programming, and intelligent optimization algorithms. MINLP mainly involves the optimization of compressor station modeling. The dynamic programming method can find the global optimal solution to the problem, but the computational complexity is high. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a multi-period scheduling optimization method for a gas compressor station. The method adopts an intelligent optimization algorithm and has the characteristics of high computational efficiency, high global search capability and strong flexibility in the process of solving the compressor scheduling optimization method. The method covers the design of a compressor station scheduling optimization model establishment module based on a two-level hierarchical structure and a compressor scheduling optimization solution module based on DE-GWO (differential evolution-grey wolf optimization algorithm) and Gurobi solver. The method can establish and solve the compressor start-stop decision model and load distribution model in stages, realize the formulation of the compressor station scheduling optimization plan, and reduce the operating cost of the compressor station.

[0004] In order to achieve the above object, the present invention adopts the following technical scheme:

[0005] A multi-period scheduling optimization method for a gas compressor station comprises the following steps:

[0006] Step 1: Design a compressor station dispatch optimization model based on a two-level hierarchical structure, and complete the establishment of mathematical models for the two stages of dispatch optimization, which are a start-stop decision model and a load distribution model;

[0007] Step 2: Design a compressor station scheduling optimization solution method based on DE-GWO and Gurobi. For the compressor station scheduling optimization model established in step 1, design the DE-GWO optimization algorithm to solve the start-stop decision model in the first stage, and call the Gurobi solver to solve the load distribution model in the second stage.

[0008] Furthermore, the step 1 comprises:

[0009] Step 1.1, establish the start-stop decision model of the first stage compressor, including: according to the information of the superior dispatching center, the multi-day gas load forecast, the inlet pressure, the outlet pressure, the compression ratio, the temperature and other information are known, the compressor efficiency is fixed as a constant, the compressor speed variable is not introduced, the decision variable is the integer variable start-stop state of the compressor, and the compressor start-stop decision model is established with the minimum operation cost of the compressor station as the objective function;

[0010] Step 1.2, establish the load distribution model of the second-stage compressor, including: taking the compressor flow and speed as decision variables, taking the minimum compressor energy consumption as the objective function, determining the flow and speed of each compressor, and calculating its working efficiency; the load distribution model optimizes the compressor flow distribution for each hour, so that it operates under more efficient conditions and reduces the energy consumption of each compressor.

[0011] Furthermore, the step 2 comprises:

[0012] Step 2.1, the differential evolution algorithm is combined with the gray wolf optimization algorithm to form an improved gray wolf algorithm of hybrid differential vector, namely DE-GWO. In the process of iteratively updating the individual positions of the wolf pack by the gray wolf optimization algorithm, the operations of mutation, crossover and selection of the differential evolution algorithm are combined to improve the diversity of the new population and enhance the ability to explore the solution space; the start and stop scheme of the compressor in each cycle is solved by the DE-GWO algorithm;

[0013] Step 2.2: For each cycle, call the Gurobi solver once to solve the load distribution model and obtain the flow distribution result of the compressor in each cycle.

[0014] Furthermore, in step 1.1, the operating cost of the gas compressor station includes the sum of the energy consumption of all compressors in the station, the cost of state conversion when the compressor is turned on, and the cost of gas supply exceeding demand.

[0015] Furthermore, it is applicable to gas compressor stations in different cities and regions.

[0016] The beneficial effects of the present invention compared with the prior art are:

[0017] (1) The present invention establishes a multi-period compressor scheduling optimization model based on a two-level hierarchical structure. The model is based on the multi-day forecast results of gas load, takes into account the costs of compressor energy consumption, multi-period start-stop, and oversupply, and optimizes the multi-day compressor scheduling plan to reduce the operating cost of the compressor unit.

[0018] (2) The present invention proposes a method for solving a multi-period compressor scheduling optimization model based on DE-GWO and Gurobi. This includes proposing an improved DE-GWO optimization algorithm to solve the first-stage compressor start-stop decision model, introducing the mutation, crossover, and selection mechanisms of the differential evolution algorithm to improve the wolf pack update rule of the gray wolf optimization algorithm; and using the Gurobi solver to solve the second-stage compressor load distribution model. The two-stage model solution is used to achieve compressor station scheduling optimization, thereby reducing gas transportation costs while meeting customer gas demand.

[0019] In summary, the present invention first takes the minimum operation cost of the compressor station as the objective function to establish the start-stop decision model of the compressor, and takes the minimum energy consumption of the compressor as the objective function to establish the load distribution model of the compressor. The two models together constitute a two-level hierarchical structure. Then, the DE-GWO optimization algorithm is designed to solve the multi-cycle start-stop decision model of the first stage, and the Gurobi solver is used to solve the load distribution model of each cycle respectively. Finally, the formulation of the compressor station scheduling optimization plan is realized, and the gas transportation cost is reduced on the basis of meeting the gas demand of customers. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A block diagram of a multi-period scheduling optimization method for a gas compressor station according to the present invention;

[0021] Figure 2 It is a compressor station dispatching flow chart based on a two-level hierarchical structure;

[0022] Figure 3 Solve the flow chart for the compressor station dispatch optimization model;

[0023] Figure 4 This is the flow chart of the DE-GWO algorithm. DETAILED DESCRIPTION

[0024] The present invention is described in further detail below in conjunction with the accompanying drawings.

[0025] like Figure 1 As shown, the present invention relates to a multi-period scheduling optimization method for a gas compressor station, including the design of a compressor station scheduling optimization method module 1 based on a two-level hierarchical structure, and the design of a compressor station scheduling optimization solution method module 2 based on DE-GWO and Gurobi, which can realize the formulation of a multi-period scheduling optimization plan for a compressor station and reduce the gas transportation cost on the basis of meeting the gas demand of customers.

[0026] like Figure 1 As shown, a multi-period scheduling optimization method for a gas compressor station of the present invention comprises the following steps:

[0027] Step 1: Design a compressor station dispatch optimization model based on a two-level hierarchical structure, and complete the establishment of two-stage mathematical models for dispatch optimization. The mathematical models are the start-stop decision model and the load distribution model. The process is as follows: Figure 2 As shown, the specific implementation is as follows:

[0028] Step 1.1 Establish the start-stop decision model of the first-stage compressor, fix the compressor efficiency as a constant, do not introduce the compressor speed variable, the decision variable is the integer variable start-stop state of the compressor, take the minimum compressor operating cost as the objective function, and solve the start-stop state of the compressor.

[0029] The operating cost of a compressor station mainly includes the following parts: the first part is the sum of the energy consumption of all compressors in the station; the second part is the cost of state conversion when the compressor is turned on; the third part is the cost of gas supply exceeding demand. The objective function is the compressor operating cost, which can be expressed by the following formula:

[0030] ,

[0031] in, is the compressor station operating cost, is the sum of the energy consumption of the compressor operation. is the compressor start-up cost, is the cost of oversupply of compressors, is the rated power of the i-th compressor, is the start / stop status of the i-th compressor on day t, P i is the compressor power, is the start-stop energy consumption coefficient of the i-th compressor, is the oversupply energy consumption coefficient of the i-th compressor, is the volume flow rate of the i-th compressor on day t, is the gas demand on day t, n is the total number of compressors, and T is the total number of cycles.

[0032] The compression ratio of the parallel compressor is the same as the overall compression ratio of the compressor station, and the flow rate of all compressors is equal to the overall flow rate of the compressor station. At the same time, the flow rate of the compressor station must be greater than the gas demand of the customer. Based on the above constraints and objective function, the start-stop decision model of the first stage compressor can be expressed as follows:

[0033] ,

[0034] in, is the compression ratio of the compressor station, is the start / stop status of the i-th compressor, is the compression ratio of the ith compressor, is the compressor station flow rate on day t, is the gas demand on day t.

[0035] Step 1.2 In the first stage, the start and stop status of the compressors at each hour in the time T is determined. Therefore, the load distribution model in the second stage optimizes the compressor flow distribution for each hour. The time variable t can be omitted when constructing the model. The objective function is the sum of the energy consumption of the compressor units P total As shown below:

[0036] ,

[0037] The start and stop status of the compressor has been calculated in the first stage. Based on the compression ratio, inlet pressure and other parameters provided by the dispatch center, the compressor load distribution model can be simplified. The compressor speed can be expressed as a function of the flow rate. At the same time, when the compressor pressure head is determined, the minimum allowable flow rate and maximum allowable flow rate of the compressor under this working condition can be solved. The compressor working efficiency is further calculated, and the compressor energy consumption is expressed as volume flow rate. The compressor load distribution model can be simplified as follows, where the start and stop status of the compressor is represented by an integer variable.

[0038] ,

[0039] in, is the compressor energy consumption, Q i is the volume flow rate of the ith compressor, in m 3 / s, Q0 is the volume flow rate of the compressor station, is the minimum permissible volume flow rate of the ith compressor, in m 3 / s, is the maximum allowable volume flow rate of the ith compressor, in m 3 / s.

[0040] Through the above two stages, a multi-period scheduling optimization model for gas compressor stations was established.

[0041] Step 2: Design a compressor scheduling optimization solution method based on DE-GWO and Gurobi. For the compressor station scheduling optimization model established in step 1, design a DE-GWO optimization algorithm to solve the start-stop decision model in the first stage, and call the Gurobi solver to solve the load distribution model in the second stage. The flow chart is as follows: Figure 3 As shown, the specific implementation is as follows:

[0042] Step 2.1 Design the first-stage compressor scheduling optimization algorithm based on DE-GWO. Obtain data such as compression ratio, inlet pressure, temperature, and customer demand forecast from the superior scheduling center, and fix the compressor efficiency to a specific value. Use the above data as the input of the DE-GWO algorithm, and obtain the start-stop decision plan of the compressor in multiple cycles after calculation. The specific design of the DE-GWO algorithm is as follows:

[0043] The exploration capability of the DE (Differential Evolution, DE) algorithm and the development capability of GWO are combined to ensure that the algorithm has stronger global optimization capabilities. In the process of iteratively updating the individual positions of the wolf pack by the GWO (Grey Wolf Optimizer, GWO) algorithm, the mutation, crossover and selection operations of the DE algorithm are combined to improve the diversity of the new population and enhance the exploration capability of the solution space. The DE-GWO algorithm flow chart is shown below: Figure 4 As shown:

[0044] Step 2.1.1 Hyperparameter initialization. The initialization objects include the number of gray wolf populations, the dimension of the gray wolf position vector, and the maximum number of iterations. The dimension of the gray wolf position vector is the dimension d of the solution space. Since the present invention is aimed at multi-cycle and multi-compressor scheduling optimization, the dimension d of the solution space is set to the product of the number of cycles T and the number of compressors N, that is, the solution [1, N] represents the start and stop status of the compressor group in the first cycle, and the solution [N+1, 2N] represents the start and stop status of the compressor group in the second cycle, and so on;

[0045] Step 2.1.2. Population initialization. Randomly initialize the position vectors of all individuals in the gray wolf population and calculate their fitness values;

[0046] Step 2.1.3. Select the three wolves with the highest fitness values ​​in the wolf pack, denoted as α, β, and δ;

[0047] Step 2.1.4. Update the wolf pack positions X of other individual wolves in the wolf pack according to the positions of the three wolves α, β, and δ;

[0048] Step 2.1.5. Perform mutation operation on the wolf pack position X to generate a mutant wolf pack V, and further cross it with the original wolf pack X to generate a new wolf pack U. Compare the fitness function values ​​f(X) and f(U) of the corresponding individuals of the wolf pack X and the new wolf pack U, and select the individuals with lower fitness function to form a new wolf pack;

[0049] Step 2.1.6. Update the DE-GWO algorithm control parameters , and , the first parameter is the parameter that controls the algorithm's exploration and exploitation capabilities, and the last two parameters are the parameter vectors for the wolf pack to update its position;

[0050] Step 2.1.7. Repeat steps 2.1.3 to 2.1.6 until the maximum number of iterations is met;

[0051] Step 2.1.8 outputs the position vector of α wolf and uses it as the optimal solution of the compressor unit start-stop decision model.

[0052] Step 2.2 Based on the start and stop status of the compressor in the T period, the Gurobi solver is called once for each period to solve the load distribution model and obtain the flow distribution results of the compressor groups in each period. The decision variables of the load distribution model of the second-stage compressor are small, and the commercial solver Gurobi can be directly called to solve it and obtain the global optimal solution. Gurobi supports parallel computing and distributed computing, so its solution speed is faster. Compared with the metaheuristic algorithm, it can ensure that the solution result is the global optimal.

[0053] In summary, the present invention includes the design of a compressor station scheduling optimization model module based on a two-level hierarchical structure and the design of a compressor scheduling optimization solution module based on DE-GWO and Gurobi, which can realize the formulation of a multi-period scheduling optimization plan for a compressor station and reduce the gas transportation cost on the basis of meeting the gas demand of customers.

[0054] The contents not described in detail in the specification of the present invention belong to the prior art known to the professional and technical personnel in this field.

[0055] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A multi-period scheduling optimization method for a gas compressor station, characterized in that: The steps include: Step 1: Design a compressor station dispatch optimization model based on a two-level hierarchical structure, and complete the establishment of mathematical models for the two stages of dispatch optimization. The mathematical models are a start-stop decision model and a load distribution model, including: Step 1.1, establish the start-stop decision model of the first stage compressor, including: according to the information of the superior dispatching center, the multi-day gas load forecast, the inlet pressure, the outlet pressure, the compression ratio, and the temperature information are known, the compressor efficiency is fixed as a constant, and the compressor speed variable is not introduced. The decision variable is the integer variable start-stop state of the compressor, and the compressor start-stop decision model is established with the minimum operation cost of the compressor station as the objective function; Step 1.2, establishing a load distribution model for the second-stage compressor, including: taking the compressor flow rate and speed as decision variables, taking the lowest compressor energy consumption as the objective function, determining the flow rate and speed of each compressor, and calculating its working efficiency; the load distribution model optimizes the compressor flow distribution for each hour, so that it operates under a more efficient working condition and reduces the energy consumption of each compressor; Step 2: Design a compressor station scheduling optimization solution method based on DE-GWO and Gurobi. For the compressor station scheduling optimization model established in step 1, design a DE-GWO optimization algorithm to solve the start-stop decision model of the first stage, and call the Gurobi solver to solve the load distribution model of the second stage, including: Step 2.1, the differential evolution algorithm is combined with the gray wolf optimization algorithm to form an improved gray wolf algorithm of hybrid differential vector, namely DE-GWO. In the process of iteratively updating the individual positions of the wolf pack by the gray wolf optimization algorithm, the operations of mutation, crossover and selection of the differential evolution algorithm are combined to improve the diversity of the new population and enhance the ability to explore the solution space; the start and stop scheme of the compressor in each cycle is solved by the DE-GWO algorithm; Step 2.2: For each cycle, call the Gurobi solver once to solve the load distribution model and obtain the flow distribution result of the compressor in each cycle.

2. A multi-period scheduling optimization method for a gas compressor station according to claim 1, characterized in that: In step 1.1, the operating cost of the compressor station includes the sum of the energy consumption of all compressors in the station, the cost of state conversion when the compressor is turned on, and the cost of gas supply exceeding demand.

3. A multi-period scheduling optimization method for a gas compressor station according to claim 1, characterized in that: Applicable to gas compressor stations in different cities and regions.

Images