Generator set multi-target maintenance optimization scheduling method considering renewable energy access

Abstract

The invention discloses a generator set multi-target maintenance optimization scheduling method considering renewable energy access. The method comprises the following steps: establishing an uncertain model of renewable energy output and load; considering an economy target, a reliability target and a renewable energy consumption target, selecting a minimum total cost, a minimum standby capacity sum of squares and a minimum wind curtailment and light curtailment amount for comprehensive evaluation, and constructing a multi-target unit maintenance optimization model; converting a system power balance constraint and a reserve capacity constraint into a deterministic constraint with clear equivalence; selecting a membership function of each target based on a fuzzy theory, and converting the multi-target unit maintenance optimization model into a single-target unit maintenance optimization model; and solving the single-target unit maintenance optimization model converted based on the fuzzy theory by using an improved ant colony algorithm, and obtaining the maintenance and operation conditions of each unit while obtaining the maximum satisfaction degree. The method is applied to a power grid to arrange maintenance scheduling of the generator set, and economic benefits, reliability and renewable energy consumption are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of generator set maintenance optimization scheduling, and specifically designs a generator set multi-objective maintenance optimization scheduling method that takes into account the economy, reliability and renewable energy access and consumption problems of a power system. Background Art

[0002] With the clean energy transformation of the power system, the proportion of wind power and solar power generation in total power generation has increased rapidly. Due to the uncertainty and volatility of renewable energy, the large-scale access of renewable energy to the grid poses a huge challenge to the formulation of power generation unit maintenance plans.

[0003] In order to solve the unit maintenance optimization problem caused by the uncertainty of renewable energy, robust optimization methods and stochastic optimization methods are usually used. However, robust optimization represents random variables in the form of a set, requiring that the requirements be met in the worst case, making it too conservative and possibly lacking the best solution. Common stochastic optimization methods include scenario-based stochastic optimization and chance-constrained programming-based stochastic optimization. However, scenario-based stochastic optimization only considers limited scenarios and it is difficult to guarantee robustness. Stochastic optimization based on chance-constrained programming is established under a certain confidence level of random constraints, which can guarantee a certain degree of robustness and make the results more reliable.

[0004] Due to environmental degradation, a large number of renewable energy sources have been connected to the power grid, which has led to a shortage of reserve capacity. Due to the increasing demand for electricity, the reserve margin of modern power systems is low. In this case, the maintenance of generator sets plays a vital role in the operation and planning of power systems. A reasonable maintenance plan should minimize the operation and maintenance costs of the power generation system or improve the reliability of the system. Most studies focus on single-objective optimization, and there are fewer studies on multi-objective optimization. In addition, due to the large-scale connection of renewable energy to the grid, the maintenance of the units will have an impact on the consumption of wind power and solar power generation, so the risk of power abandonment also needs to be considered in the maintenance plan.

[0005] In summary, with the large-scale grid connection of renewable energy such as wind power and solar power, their characteristics have put forward higher requirements for the optimization of power systems. Due to the inherent uncertainty of renewable energy, maintenance and dispatching will face more challenges. Summary of the invention

[0006] In order to solve the above technical problems, the present invention provides a multi-objective maintenance optimization scheduling method for generator sets taking into account the access of renewable energy, which can provide a basis for the grid dispatching department to formulate a generator set maintenance plan and improve the economy and reliability of system operation. The method can be applied to the grid to arrange the maintenance scheduling of generator sets, improve the economic benefits, reliability and the consumption of renewable energy.

[0007] The technical solution adopted by the present invention is:

[0008] The multi-objective maintenance optimization scheduling method of power generation units considering the access of renewable energy includes the following steps:

[0009] Step 1: Use fuzzy attributes to describe the output of wind power and photovoltaic power in the medium and long term, and establish an uncertain model of renewable energy output and load;

[0010] Step 2: Considering the economic goal, reliability goal and renewable energy consumption goal, the minimization of total cost, minimization of the sum of squares of reserve capacity and minimization of wind and solar power abandonment are selected for comprehensive evaluation to build a multi-objective unit maintenance optimization model;

[0011] Step 3: Considering the system power balance constraint, spare capacity constraint, unit output constraint and unit maintenance-related constraint, fuzzy chance constrained programming is used to transform the system power balance constraint and spare capacity constraint into deterministic constraints with clear equivalence;

[0012] Step 4: Based on fuzzy theory, the descending semi-linear shape is selected as the membership function of each objective, and the multi-objective unit maintenance optimization model is transformed into a single-objective unit maintenance optimization model;

[0013] Step 5: Use the improved ant colony algorithm to solve the single-objective unit maintenance optimization model transformed from fuzzy theory. When the maximum satisfaction is obtained, the maintenance and operation status of each unit is obtained at the same time.

[0014] In step 1, the membership function is a useful tool for quantifying fuzzy variables. The size of the membership indicates the degree to which each element belongs to the fuzzy set. The closer the membership is to 1, the higher the possibility that the element belongs to the fuzzy set; the closer the membership is to 0, the lower the possibility that the element belongs to the fuzzy set. Among them, the triangular membership function has the advantages of conservative, possible and optimistic fuzzy properties, and is suitable for simulating wind power and photovoltaic power generation output.

[0015] Fuzzy variables are introduced to describe the uncertainty of renewable energy output and load. The wind power fuzzy parameters at time t are expressed as The photovoltaic fuzzy parameters at time t are expressed as The load fuzzy parameter at time t is expressed as The expressions of these fuzzy parameters are based on the triangular fuzzy parameter model and are expressed as follows:

[0016]

[0017] In formula (1): P W1,t It is expressed as the most conservative membership parameter of wind power at time t; P W2,t It is expressed as the most likely membership parameter of wind power at time t; P W3,tIt is expressed as the most optimistic membership parameter of wind power at time t. V1,t It is expressed as the most conservative membership parameter of photovoltaic at time t; P V2,t It is expressed as the most likely membership parameter of photovoltaic at time t; P V3,t It is expressed as the most optimistic membership parameter of photovoltaic at time t. L1,t It is expressed as the most conservative membership parameter of the load at time t; P L2,t It is expressed as the most likely membership parameter of the load at time t; P L3,t It is expressed as the most optimistic membership parameter of the load at time t.

[0018] The expression of the expected value of the triangle fuzzy parameter is as follows:

[0019]

[0020] In formula (2): The expected value of the triangular fuzzy parameter representing wind power; The expected value of the triangle fuzzy parameter representing photovoltaic; represents the expected value of the triangular fuzzy parameter of the load; α is the confidence level; P Wc,t , P Vc,t and P Lc,t are the predicted values ​​of wind power, photovoltaic power and load at time t respectively.

[0021] In step 2, a multi-objective unit maintenance optimization model is established that takes into account the economy, reliability and renewable energy consumption of the power system, and the objective function includes:

[0022] Economic goal: Minimize the total cost, which is the maintenance cost F M and operating cost F S composition;

[0023] Reliability objective: Minimize the sum of squares of reserve capacity R so that the reserve capacity remains balanced during maintenance;

[0024] Renewable energy consumption target: Minimize the amount of wind and solar power abandoned, which is determined by the abandoned wind volume C W and the amount of abandoned light C P The constraints of the multi-objective unit maintenance optimization model include: system power balance constraints, spare capacity constraints, thermal power unit output constraints, thermal power unit ramp constraints, thermal power unit continuous start and stop constraints, wind power and photovoltaic output constraints, unit maintenance status constraints, unit maintenance time constraints, unit maintenance continuity constraints, and unit maintenance resource constraints.

[0025] In step 2, the multi-objective unit maintenance optimization model is constructed, starting from multiple perspectives such as economic benefits, reliability, and renewable energy consumption, and selecting minimization of total cost, minimization of the sum of squares of reserve capacity, and minimization of wind and solar power abandonment for comprehensive evaluation;

[0026] 1) Minimize total cost:

[0027] The total cost is determined by the maintenance cost F M and operating cost F S composition:

[0028] minF=F M +F S (3);

[0029]

[0030] Where: T is the total number of dispatch periods; N is the number of thermal power units; M i is the unit maintenance cost of thermal power unit i; P i max is the maximum active power output of thermal power unit i; x i,t is the maintenance status of thermal power unit i in period t, if it is in maintenance status, it is 1, otherwise it is 0; △t is the step length of the scheduling calculation period. i , b i and c i It is expressed as the fuel cost coefficient of thermal power unit i; P i,t is the active power output of thermal power unit i in period t; u i,t is the operating status of thermal power unit i in period t, 1 indicates operation and 0 indicates shutdown.

[0031] 2) Minimize the sum of squares of spare capacity:

[0032] System reliability can be represented by minimizing the sum of squares of reserve capacity, so that the reserve capacity remains balanced during maintenance periods;

[0033]

[0034] Where: is the maximum output power of unit i in period t; D t is the load demand power during period t.

[0035] 3) Minimize the amount of wind and solar power curtailment:

[0036] The amount of wind and solar power abandoned is determined by the amount of abandoned wind C W and the amount of abandoned light C P composition:

[0037] minC=C W +C P (7);

[0038]

[0039] Where: C W and C P are the total amount of wind power abandoned and solar power abandoned in the system respectively; P WF,t and P W,t are the available active power and actual active power of the wind farm in period t respectively; P VF,t and P V,t are the available active power and actual active power of the PV power plant in period t respectively.

[0040] The constraints include:

[0041] 1) System power balance constraints:

[0042]

[0043] Where: represents the fuzzy parameters of wind power at time t; represents the fuzzy parameters of photovoltaic at time t; represents the fuzzy parameter of the load at time t; α represents the confidence level; Pr{} represents the possibility of the event.

[0044] 2) Reserve capacity constraints:

[0045]

[0046] Where: P i max represents the maximum output power of unit i; α represents the confidence level.

[0047] 3) Unit output constraints:

[0048] ① Output constraints of thermal power units:

[0049] u i,t P i min ≤P i,t ≤u i,t P i max (12);

[0050] Where: P i min is the minimum active output of thermal power unit i.

[0051] ② Thermal power unit climbing constraints:

[0052] -r i d ≤P i,t -P i,t-1 ≤ri u (13);

[0053] Where: P i,t-1 is the active power output of thermal power unit i in period t-1; r i d is the lower limit of the climbing slope of thermal power unit i; r i u is the upper limit of the climbing rate of thermal power unit i.

[0054] ③ Constraints on continuous start and stop of thermal power units:

[0055]

[0056] Where: and are the continuous start-up time and continuous shutdown time of thermal power unit i in period t-1 respectively; T i on and T i off is the minimum continuous start-up time and the minimum continuous shutdown time of thermal power unit i.

[0057] u i,t represents the operating status of thermal power unit i in period t; u i,t-1 Represents the operating status of thermal power unit i in period t-1.

[0058] ④ Wind power and photovoltaic output constraints:

[0059]

[0060] Where: is the expected value of the triangular fuzzy parameter of wind power; is the expected value of the triangular fuzzy parameter of optoelectronics. 4) Constraints related to unit maintenance:

[0061] Unit maintenance status constraints:

[0062]

[0063] x i,t =0,t <z i or t>d i (19);

[0064] y i,t =0,t <z i or t>d i +l i -1(20);

[0065] Where: y i,t is the state variable of unit i in period t, which is 1 if it is in maintenance state, otherwise it is 0; zi is the earliest time that maintenance of unit i is allowed to start; d i is the latest time allowed for the start of maintenance for unit i; l i is the maintenance duration of unit i.

[0066] Unit maintenance time constraints:

[0067]

[0068] Unit maintenance continuity constraints:

[0069] y i,t -y i,t-1 ≤x i,t (twenty two);

[0070] Where: y i,t-1 Represents the state variable of unit i in period t-1.

[0071] y i,1 ≤x i,1 (twenty three);

[0072] Where: y i,1 represents the state variable of unit i at the initial time; x i,1 Represents the maintenance status of unit i at the initial moment.

[0073] Unit maintenance resource constraints:

[0074]

[0075] Where: B t It is the upper limit of maintenance items to be carried out at the same time; N represents the number of thermal power units.

[0076] In step 3, the fuzzy parameters are transformed into clear equivalence classes:

[0077] The system power balance constraint and the reserve capacity constraint involve fuzzy variables and chance constraints. Fuzzy chance constraint programming can be transformed into deterministic constraints with explicit equivalence. The explicit equivalent expression of the system power balance constraint is:

[0078]

[0079] In the formula: α represents the confidence level.

[0080] A clear equivalent expression for the spare capacity constraint is:

[0081]

[0082] In step 3, for the multi-objective unit maintenance optimization model, the fuzzy theory is used to establish the membership function of the objective function, and the multi-objective unit maintenance optimization problem is converted into a single-objective unit maintenance optimization problem. It is expected that the total cost, the sum of squares of the reserve capacity, and the amount of wind and solar power abandoned will be minimized as much as possible under all constraints. The lower the total cost of the system, the smaller the sum of squares of the reserve capacity, and the lower the amount of wind and solar power abandoned, the higher the satisfaction of the decision maker, and the closer its membership function is to 1.

[0083] The present invention selects the descending half-line as the membership function of each objective. The membership function corresponding to the three objective functions is:

[0084]

[0085] Where: μ(F) represents the membership function of the total cost; F represents the total cost objective function; F1 represents the ideal value of the total system cost; δ1 represents the elastic value of the total cost F that the decision maker can accept; μ(R) represents the membership function of the sum of squares of the reserve capacity; R represents the objective function of the sum of squares of the reserve capacity; R1 represents the ideal value of the sum of squares of the reserve capacity; δ e is the expansion value of the sum of squares of reserve capacity R that is acceptable to decision makers; μ(C) represents the membership function of the amount of wind and solar power abandoned; C represents the objective function of the amount of wind and solar power abandoned; C1 is the ideal value of the amount of wind and solar power abandoned; δ s It is the flexible value of the amount of wind and solar power curtailment C that decision makers can accept.

[0086] According to the fuzzification principle, the satisfaction degree of μ(F), μ(R) and μ(C) is defined as follows:

[0087] μ=min{μ(F),μ(R),μ(C)} (30);

[0088] Where: min{μ(F),μ(R),μ(C)} represents the minimum value of all membership functions.

[0089] Therefore, the multi-objective unit maintenance optimization problem is transformed into a single-objective unit maintenance optimization problem with the maximum satisfaction index μ:

[0090]

[0091] Its constraints also include equations (10) to (24).

[0092] The step 4 comprises the following steps:

[0093] S4.1: Input the original data and use the mixed integer programming algorithm to solve the single-objective unit maintenance model F with the minimum system total cost as the objective function, and obtain the total cost F1, the sum of squares of the reserve capacity R′ and the amount of wind and solar power abandoned C′.

[0094] Mixed Integer Programming Algorithm:

[0095]

[0096] Where: h i (x,y) is an equality constraint; g j (x,y) is an inequality constraint; x is a real number variable; y is an integer variable; m is the number of equality constraints; l is the number of inequality constraints.

[0097] S4.2: Using the same method as above, solve the single-objective unit maintenance model R with the minimum sum of squares of system reserve capacity as the objective function, and obtain the sum of squares of reserve capacity R1, total cost F′ and amount of wind and solar power abandoned C″.

[0098] S4.3: Using the same method as above, solve the single-objective unit maintenance model C with the objective function of minimizing the amount of wind and solar power abandoned in the system, and obtain the amount of wind and solar power abandoned C1, the total cost F″ and the sum of squares of the reserve capacity R″.

[0099] S4.4: Based on S4.1 to S4.3, each single target is scaled to a certain extent to determine δ1, δ e and δ s , its value range is:

[0100] 0≤δ1≤max(F′-F1,F″-F1);

[0101] 0≤δ e ≤max(R′-R1,R″-R1);

[0102] 0≤δ s ≤max(C′-C1,C″-C1).

[0103] According to the above constraints, and combined with the actual situation, a reasonable expansion value is determined. Theoretically, δ1, δ e and δ s The smaller the value of , the better, but it will bring certain difficulties to the solution.

[0104] S4.5: δ1, δ e and δ s Substituting into equations (27), (28) and (29), we get the membership function, thus converting the multi-objective into a single-objective function. Since the mathematical model shown in equation (31) is a mixed integer linear programming model, the improved ant colony algorithm is used to solve it.

[0105] The pheromone update strategy in the traditional ant colony algorithm will cause the algorithm to stagnate, converge slowly and other shortcomings. To this end, the present invention proposes an improved ant colony algorithm, which includes a new pheromone hybrid update strategy, that is, the optimal solution of this iteration is used for update in the first 15 cycles of the search, and after 15 fixed cycles, the global optimal solution is used for update. After adopting the hybrid pheromone update strategy, the ant colony algorithm will first find more feasible solutions, and can also maintain the ability to continue searching for the optimal solution, while maintaining a faster convergence speed. The improved pheromone update rules are as follows:

[0106] τ(t+1)=(1-ρ)τ(t)+△τ s (33);

[0107]

[0108] Where: τ(t) is the concentration of pheromone at time t; ρ is the pheromone volatility factor; △τ s is the change in pheromone concentration; f s is the iterative optimal solution.

[0109] The method for determining the minimum and maximum values ​​of pheromone concentration is as follows:

[0110] When pheromones are not updated:

[0111]

[0112] Where: G(s) is the length of the optimal path found; τ min (t) is the minimum value of pheromone concentration; τ max (t) is the maximum value of pheromone concentration.

[0113] After the pheromone update:

[0114]

[0115] When the maximum satisfaction is obtained, the maintenance and operation status of each unit is obtained at the same time. When the load is at a low point, the load demand is low, and fewer generators are required, so more thermal power units can be planned for maintenance. And because the output of wind and solar power is large during these periods, it can make up for the power lost due to maintenance of thermal power units, reduce the generation of wind and solar power abandonment, and solve the problem of renewable energy consumption. When the load is at a peak point, the load demand is high, and for the sake of operational safety, thermal power units are not planned for maintenance.

[0116] The present invention provides a multi-objective maintenance optimization scheduling method for generator sets taking into account renewable energy access, and the technical effects are as follows: 1) The present invention constructs a random optimization scheduling method based on fuzzy chance constraint programming to solve the optimization scheduling problem caused by supply and demand uncertainty, and uses three fuzzy attributes of the most conservative, most likely, and most optimistic to describe the output scale of wind power and photovoltaic power in the medium and long term. Compared with the traditional robust optimization scheduling whose processing results are too conservative, it is established under random constraints with a certain confidence level, which can ensure a certain robustness and make the results more reliable.

[0117] 2) The present invention establishes a multi-objective maintenance optimization model for generator sets taking into account the access of renewable energy. The model takes into account the economy, reliability and absorption of renewable energy of the power system, and constructs a multi-objective unit maintenance optimization model to minimize the total cost of system operation, the sum of squares of spare capacity and the absorption amount. Compared with the traditional model that only considers one objective, it obtains better economy, higher reliability and less wind and solar power abandonment.

[0118] 3) The present invention introduces fuzzy theory and uses the maximum satisfaction index method to transform multi-objective optimization into single-objective optimization, and uses the improved ant colony algorithm to solve the single-objective model. The improved ant colony algorithm can avoid the problems of stagnation and slow convergence of the traditional ant colony algorithm as much as possible, and provides a reference for solving single-objective problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0119] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0120] Figure 1 This is a comparison chart of the total number of generator sets overhauled according to an embodiment of the present invention.

[0121] Figure 2 This is a comparison chart of the total capacity of the generator sets under maintenance according to an embodiment of the present invention.

[0122] Figure 3 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0123] A multi-objective maintenance optimization scheduling method for generator sets considering the access of renewable energy is proposed. A multi-objective unit maintenance optimization model considering the access of renewable energy is established, which takes into account the economy, reliability and absorption of renewable energy in the power system. Firstly, the output and load of renewable energy are represented by fuzzy parameters, and fuzzy opportunity constraints are used to deal with uncertainty to reduce the unsafe operation caused by the output deviation of renewable energy, and it is converted into a clear equivalence class for easy solution; secondly, a multi-objective unit maintenance optimization model is established to minimize the total cost of system operation, the sum of squares of spare capacity and the absorption amount. Fuzzy theory is introduced, and the maximum satisfaction index method is used to convert multi-objective optimization into single-objective optimization. The improved ant colony algorithm is used to solve the single-objective model.

[0124] The multi-objective maintenance optimization scheduling method of power generation units considering the access of renewable energy includes the following steps:

[0125] Step 1: Propose an uncertain model of renewable energy output and load based on membership function:

[0126] 1.1: Fuzzy variables are introduced to describe the uncertainty of renewable energy output and load, and the expression of fuzzy parameters is described as a triangular fuzzy parameter model.

[0127]

[0128] 1.2: Construct the expression of the expected value of the triangle fuzzy parameter:

[0129]

[0130] Step 2: From multiple perspectives such as economic benefits, reliability, and renewable energy consumption, a multi-objective unit maintenance optimization model taking into account renewable energy access is constructed, and a comprehensive evaluation is performed by minimizing total cost, minimizing the sum of squares of reserve capacity, and minimizing the amount of wind and solar power abandoned.

[0131] 2.1: Minimize the total cost:

[0132] minF=F M +F S (3);

[0133]

[0134] 2.2: Minimize the sum of squares of spare capacity:

[0135]

[0136] 2.3: Minimize the amount of wind and solar power curtailment:

[0137] minC=C W +C P(7);

[0138]

[0139] 2.4: Construct system power balance constraints, spare capacity constraints, unit output constraints and unit maintenance related constraints.

[0140] System power balance constraints:

[0141]

[0142] Reserve capacity constraints:

[0143]

[0144] Unit output constraints:

[0145]

[0146] Relevant constraints for unit maintenance:

[0147]

[0148] 2.5: Convert fuzzy chance constraints into crisp equivalence classes.

[0149] The explicit equivalent expression of the system power balance constraint is:

[0150]

[0151] A clear equivalent expression for the spare capacity constraint is:

[0152]

[0153] Step 3: Solving the multi-objective unit maintenance optimization model:

[0154] 3.1: For the multi-objective unit maintenance optimization model, the fuzzy theory is used to establish the membership function of the objective function, and the multi-objective unit maintenance optimization problem is transformed into a single-objective unit maintenance optimization problem. The membership functions corresponding to the three objective functions are:

[0155]

[0156] 3.2: Based on the fuzzy principle, define the degree of satisfaction.

[0157] μ=min{μ(F),μ(R),μ(C)} (30);

[0158]

[0159] 3.3: Obtain the membership function, transform the multi-objective into a single-objective function, and use the CPLEX solver to solve it. When the maximum satisfaction is obtained, the maintenance and operation status of each unit is also obtained.

[0160] Verification example:

[0161] Table 1 Extreme values ​​of each single target

[0162]

[0163] Table 2 Maximum satisfaction index and optimization results

[0164]

[0165] As shown in Table 2, after fuzzy processing, the satisfaction of the multi-objective unit maintenance optimization model at the global optimum is 0.811. At this time, the total cost of the system is 149.104 million yuan, which is 523,000 yuan more than the cost when the total cost is minimized as the single objective optimization, that is, an increase of 0.352%, but it is 1.764 million yuan and 201,000 yuan less than the cost when the reserve capacity or the amount of wind and solar power abandoned is minimized as the single objective optimization, that is, 1.183% and 0.135% respectively. The sum of squares of the reserve capacity at the global optimum is 80617099MW 2 , an increase of 1967432MW compared to the single target of minimizing the sum of squares of reserve capacity 2 , which is an increase of 2.502%. However, compared with the single-objective optimization with the minimum total cost or the minimum amount of wind and solar power curtailment, the sum of squares of reserve capacity decreased by 5708335MW. 2 and 3537482MW 2 , i.e. 7.081% and 4.388%. At the global optimum, the amount of wind and solar power abandoned is 2022.77MWh, which is 11.34MWh more than when the single objective optimization is to minimize the amount of wind and solar power abandoned, i.e. an increase of 0.564%, but it is 14.51MWh and 22.87MWh less than when the single objective optimization is to minimize the total cost or the reserve capacity, i.e. 0.717% and 1.131% respectively.

[0166] As can be seen from Table 2, with the increase of satisfaction index, the corresponding total system cost and the amount of wind and solar power abandoned are increasing, while the sum of squares of spare capacity is greatly reduced. In order to significantly reduce the sum of squares of spare capacity, the total system cost must be increased, which shows that the multi-objective unit maintenance optimization model in this paper is necessary and effective for optimizing reliability. By choosing a compromise scheduling scheme between the total system cost, the amount of wind and solar power abandoned, and the sum of squares of spare capacity, not only the economic efficiency of the system is improved, but also the safe operation of the system and the absorption of renewable energy are guaranteed.

[0167] Table 3 Optimal maintenance scheduling of generator sets

[0168]

[0169]

Claims

1. A multi-objective maintenance optimization scheduling method for power generation units taking into account renewable energy access, characterized in that The following steps are involved: Step 1: Use fuzzy attributes to describe the output of wind power and photovoltaic power in the medium and long term, and establish an uncertain model of renewable energy output and load; Step 2: Considering the economic goal, reliability goal and renewable energy consumption goal, the minimization of total cost, minimization of the sum of squares of reserve capacity and minimization of wind and solar power abandonment are selected for comprehensive evaluation to build a multi-objective unit maintenance optimization model; Step 3: Considering the system power balance constraint, spare capacity constraint, unit output constraint and unit maintenance-related constraint, fuzzy chance constrained programming is used to transform the system power balance constraint and spare capacity constraint into deterministic constraints with clear equivalence; Step 4: Based on fuzzy theory, the membership function of each target is selected to transform the multi-target unit maintenance optimization model into a single-target unit maintenance optimization model; Step 5: Use the improved ant colony algorithm to solve the single-objective unit maintenance optimization model transformed from fuzzy theory. When the maximum satisfaction is obtained, the maintenance and operation status of each unit is obtained at the same time.

2. The multi-objective maintenance optimization scheduling method for power generation units taking into account renewable energy access according to claim 1 is characterized in that: In step 1, fuzzy variables are introduced to describe the uncertainty of renewable energy output and load. The wind power fuzzy parameters at time t are expressed as The photovoltaic fuzzy parameters at time t are expressed as The load fuzzy parameter at time t is expressed as The expressions of these fuzzy parameters are based on the triangular fuzzy parameter model and are expressed as follows: In formula (1): P W1,t It is expressed as the most conservative membership parameter of wind power at time t; P W2,t It is expressed as the most likely membership parameter of wind power at time t; P W3,t It is expressed as the most optimistic membership parameter of wind power at time t; P V1,t It is expressed as the most conservative membership parameter of photovoltaic at time t; P V2,t It is expressed as the most likely membership parameter of photovoltaic at time t; P V3,t It is expressed as the most optimistic membership parameter of photovoltaic at time t; P L1,t It is expressed as the most conservative membership parameter of the load at time t; P L2,t It is expressed as the most likely membership parameter of the load at time t; P L3,t It is expressed as the most optimistic membership parameter of the load at time t; The expression of the expected value of the triangle fuzzy parameter is as follows: In formula (2): The expected value of the triangular fuzzy parameter representing wind power; The expected value of the triangle fuzzy parameter representing photovoltaic; represents the expected value of the triangular fuzzy parameter of the load; α is the confidence level; P Wc,t , P Vc,t and P Lc,t are the predicted values ​​of wind power, photovoltaic power and load at time t respectively.

3. The multi-objective maintenance optimization scheduling method for power generation units taking into account renewable energy access according to claim 1 is characterized in that: In step 2, a multi-objective unit maintenance optimization model is established that takes into account the economy, reliability and renewable energy consumption of the power system, and the objective function includes: Economic goal: Minimize the total cost, which is the maintenance cost F M and operating cost F S composition; Reliability objective: Minimize the sum of squares of reserve capacity R so that the reserve capacity remains balanced during maintenance; Renewable energy consumption target: Minimize the amount of wind and solar power abandoned, which is determined by the abandoned wind volume C W and the amount of abandoned light C P composition; The constraints of the multi-objective unit maintenance optimization model include: system power balance constraints, spare capacity constraints, thermal power unit output constraints, thermal power unit ramping constraints, thermal power unit continuous start and stop constraints, wind power and photovoltaic output constraints, unit maintenance status constraints, unit maintenance time constraints, unit maintenance continuity constraints, and unit maintenance resource constraints.

4. The multi-objective maintenance optimization scheduling method for power generation units taking into account renewable energy access according to claim 3 is characterized in that: The multi-objective unit maintenance optimization model constructed takes into account economic benefits, reliability and renewable energy consumption, and selects minimization of total cost, minimization of the sum of squares of reserve capacity and minimization of wind and solar power abandonment for comprehensive evaluation; 1) Minimize total cost: The total cost is determined by the maintenance cost F M and operating cost F S composition: minF=F M +F S (3); Where: T is the total number of dispatching periods; N is the number of thermal power units; M i is the unit maintenance cost of thermal power unit i; P i max is the maximum active output of thermal power unit i; x i,t is the maintenance status of thermal power unit i in period t, if it is in maintenance status, it is 1, otherwise it is 0; △t is the step length of the scheduling calculation period; a i , b i and c i It is expressed as the fuel cost coefficient of thermal power unit i; P i,t is the active power output of thermal power unit i in period t; u i,t is the operating status of thermal power unit i in period t, 1 indicates operation and 0 indicates shutdown; 2) Minimize the sum of squares of spare capacity: System reliability is represented by minimizing the sum of squares of reserve capacity, so that the reserve capacity remains balanced during maintenance periods; Where: is the maximum output power of unit i in period t; D t is the load demand power in period t; 3) Minimize the amount of wind and solar power curtailment: The amount of wind and solar power abandoned is determined by the amount of abandoned wind C W and the amount of abandoned light C P composition: minC=C W +C P (7); Where: C W and C P are the total amount of wind power abandoned and solar power abandoned in the system respectively; P WF,t and P W,t are the available active power and actual active power of the wind farm in period t respectively; P VF,t and P V,t are the available active power and actual active power of the PV power plant in period t respectively.

5. The multi-objective maintenance optimization scheduling method for power generation units taking into account renewable energy access according to claim 4 is characterized in that: The constraints of the multi-objective unit maintenance optimization model include: 1) System power balance constraints: Where: represents the fuzzy parameters of wind power at time t; represents the fuzzy parameters of photovoltaic at time t; represents the fuzzy parameter of the load at time t; α represents the confidence level; Pr{} represents the probability of the event; 2) Reserve capacity constraints: Where: P i max represents the maximum output power of unit i; α represents the confidence level; 3) Unit output constraints: ① Output constraints of thermal power units: u i,t Q i min ≤P i,t ≤u i,t Q i max (12); Where: P i min is the minimum active output of thermal power unit i; ② Thermal power unit climbing constraints: -r i d ≤P i,t -P i,t-1 ≤r i u (13); Where: P i,t-1 is the active power output of thermal power unit i in period t-1; r i d is the lower limit of the climbing slope of thermal power unit i; r i u is the upper limit of the climbing rate of thermal power unit i; ③ Constraints on continuous start and stop of thermal power units: Where: and are the continuous start-up time and continuous shutdown time of thermal power unit i in period t-1 respectively; T i on and T i off is the minimum continuous start time and minimum continuous shutdown time of thermal power unit i; u i,t represents the operating status of thermal power unit i in period t; u i,t-1 Indicates the operating status of thermal power unit i in period t-1; ④ Wind power and photovoltaic output constraints: Where: is the expected value of the triangular fuzzy parameter of wind power; is the expected value of the triangular fuzzy parameter of optoelectronics; 4) Constraints related to unit maintenance: Unit maintenance status constraints: x i,t =0,t <z i or t>d i (19); y i,t = 0, t < z i or t > d i + l i -1(20); Where: y i,t is the state variable of unit i in period t, which is 1 if it is in maintenance state, otherwise it is 0; z i is the earliest time that maintenance of unit i is allowed to start; d i is the latest time allowed for the start of maintenance for unit i; l i is the maintenance duration of unit i; Unit maintenance time constraints: Unit maintenance continuity constraints: and i,t -and i,t-1 ≤x i,t (22); Where: y i,t-1 represents the state variable of unit i in period t-1; and i,1 ≤x i,1 (23); Where: y i,1 represents the state variable of unit i at the initial time; x i,1 represents the maintenance status of unit i at the initial time; Unit maintenance resource constraints: Where: B t It is the upper limit of maintenance items to be carried out at the same time; N represents the number of thermal power units.

6. The multi-objective maintenance optimization scheduling method for power generation units taking into account renewable energy access according to claim 1 is characterized in that: In step 3, the fuzzy parameters are transformed into clear equivalence classes: The system power balance constraint and reserve capacity constraint involve fuzzy variables and chance constraints; the fuzzy chance constraint programming is transformed into a deterministic constraint with explicit equivalence; the explicit equivalent expression of the system power balance constraint is: Where: α represents the confidence level; A clear equivalent expression for the spare capacity constraint is:

7. The multi-objective maintenance optimization scheduling method for power generation units taking into account renewable energy access according to claim 6 is characterized in that: In the step 3, for the multi-objective unit maintenance optimization model, the membership function of the objective function is established by using fuzzy theory, and the multi-objective unit maintenance optimization problem is converted into a single-objective unit maintenance optimization problem; It is expected that the total cost, the sum of squares of reserve capacity and the amount of wind and solar power abandoned will be minimized as much as possible under all constraints; the lower the total cost of the system, the smaller the sum of squares of reserve capacity and the lower the amount of wind and solar power abandoned, the higher the decision maker's satisfaction will be, and the closer its membership function will be to 1; the descending half-line is selected as the membership function of each objective; the membership functions corresponding to the three objective functions are: Where: μ(F) represents the membership function of the total cost; F represents the total cost objective function; F1 represents the ideal value of the total system cost; δ1 represents the elastic value of the total cost F that the decision maker can accept; μ(R) represents the membership function of the sum of squares of the reserve capacity; R represents the objective function of the sum of squares of the reserve capacity; R1 represents the ideal value of the sum of squares of the reserve capacity; δ e is the expansion value of the sum of squares of reserve capacity R that is acceptable to decision makers; μ(C) represents the membership function of the amount of wind and solar power abandoned; C represents the objective function of the amount of wind and solar power abandoned; C1 is the ideal value of the amount of wind and solar power abandoned; δ s is the flexible value of the amount of wind and solar power curtailment C that decision makers can accept; According to the fuzzification principle, the satisfaction degree of μ(F), μ(R) and μ(C) is defined as follows: μ=min{μ(F),μ(R),μ(C)} (30); Where: min{μ(F),μ(R),μ(C)} represents the minimum value of all membership functions; Therefore, the multi-objective unit maintenance optimization problem is transformed into a single-objective unit maintenance optimization problem with the maximum satisfaction index μ:

8. The multi-objective maintenance optimization scheduling method for power generation units taking into account renewable energy access according to claim 1 is characterized in that: The step 4 comprises the following steps: S4.1: Input the original data and use the mixed integer programming algorithm to solve the single-objective unit maintenance model F with the minimum system total cost as the objective function, and obtain the total cost F1, the sum of squares of the reserve capacity R′ and the amount of wind and solar power abandoned C′; Mixed Integer Programming Algorithm: Where: h i (x,y) is an equality constraint; g j (x,y) is an inequality constraint; x is a real variable; y is an integer variable; m is the number of equality constraints; l is the number of inequality constraints; S4.2: Solve the single-objective unit maintenance model R with the minimum sum of squares of system reserve capacity as the objective function, and obtain the sum of squares of reserve capacity R1, total cost F′ and the amount of wind and solar power abandoned C″; S4.3: Solve the single-objective unit maintenance model C with the minimum amount of wind and solar power abandoned in the system as the objective function, and obtain the amount of wind and solar power abandoned C1, the total cost F″ and the sum of squares of spare capacity R″; S4.4: Based on S4.1 to S4.3, each single target is scaled to a certain extent to determine δ1, δ e and δ s , its value range is: 0≤δ1≤max(F′-F1,F″-F1); 0≤δ e ≤max(R′-R1,R″-R1); 0≤δ s ≤max(C′-C1,C″-C1); S4.5: δ1, δ e and δ s Substituting into equations (27), (28) and (29), we get the membership function, thereby converting the multi-objective into a single-objective function; the mathematical model shown in equation (31) is a mixed integer linear programming model.

9. The multi-objective maintenance optimization scheduling method for power generation units taking into account renewable energy access according to claim 8 is characterized in that: The improved ant colony algorithm is used to solve the mixed integer linear programming model, as follows: The improved ant colony algorithm adopts a mixed pheromone update strategy. The pheromone update rules are as follows: τ(t+1)=(1-ρ)τ(t)+△τ s (33); Where: τ(t) is the concentration of pheromone at time t; ρ is the pheromone volatility factor; △τ s is the change in pheromone concentration; f s is the iterative optimal solution; The method for determining the minimum and maximum values ​​of pheromone concentration is as follows: When pheromones are not updated: Where: G(s) is the length of the optimal path found; τ min (t) is the minimum value of pheromone concentration; τ max (t) is the maximum value of pheromone concentration; After the pheromone update: When maximum satisfaction is achieved, the maintenance and operating status of each unit can be obtained at the same time.

10. The multi-objective maintenance optimization scheduling method for power generation units taking into account renewable energy access according to claim 9 is characterized in that: The maintenance and operating conditions of each unit include: when the load is at a low point, the load demand is low and fewer generating units are required, so more thermal power units are planned to be maintained; when the load is at a peak point, the load demand is high and for operational safety reasons, no thermal power units are planned to be maintained.

Images