An active power optimal dispatching method based on unit comprehensive evaluation index

Abstract

This invention discloses an active power optimization scheduling method based on comprehensive unit evaluation indicators. The method includes defining unit evaluation indicators, including accumulated unit fatigue damage, unit regulation margin (including power up-adjustment margin and power down-adjustment margin), and unit power fluctuation. The constraints to be considered in wind farm power scheduling include: overall farm output constraints, unit output constraints, and wind farm power change constraints. The method calculates unit indicator values ​​based on the unit evaluation indicators, optimizes scheduling based on the comprehensive unit indicator values, and classifies the units according to their current output and status. Finally, by using a rolling optimization method based on real-time wind turbine status evaluation, the method can effectively reduce the number of unit adjustments and improve the accuracy of wind farm power demand tracking.

Description

Technical Field

[0001] This invention belongs to the field of wind farm site optimization scheduling, and particularly relates to an active power optimization scheduling method based on the comprehensive evaluation index of the generating units. Background Technology

[0002] In recent years, large-scale centralized wind power grid connection has become one of the main characteristics of wind power generation. However, with the high penetration rate of wind power into the power grid, its volatility, randomness, and uncertainty have significantly impacted the operation and control of the power system. Therefore, considering wind power forecasting information, how to formulate an active power control strategy that balances transmission security and coordinates the output power of multiple wind farms, and improve the grid's capacity to accommodate wind power clusters, has become one of the urgent problems to be solved.

[0003] Grid connection of wind power clusters requires consideration of the distribution of wind power among various wind farms. Analyzing the actual situation of local power grids with wind farm clusters, Wen Zhiwei et al. proposed a method for economic grid dispatch to save power generation costs, reduce wind curtailment, and enhance the safe and stable operation of the power grid. Xue Feng et al., using the Jiuquan wind power base in Gansu as an example, proposed an active power control approach for wind power clusters, introducing the control system configuration scheme, system functions, and control strategy design principles. Rohring et al. designed an overall framework for hierarchical and zoned control of intermittent power clusters. Tang Yi et al., based on different optimization objectives, proposed active power allocation strategies for wind farm output within wind power clusters. To avoid the conflict between wind power consumption and economic system operation, Wen Jing introduced multi-objective optimization theory into the source-load coordination problem, establishing a multi-objective optimization scheduling model for source-load coordination in wind power cluster access. To fully explore the coordination effect of active power control among wind farms within a wind power cluster, Liu Yang et al. proposed a rapid coordinated allocation method for wind curtailment power in wind power clusters under output limitation. To improve the ability of wind power clusters to track system dispatch commands, reduce the power output fluctuations of wind farms within the cluster, and achieve power allocation among wind farms within the cluster, Wang Jing et al. proposed a multi-objective, two-layer model predictive control (MPC) active power control method for wind power clusters based on model predictive control theory. To address the intraday dispatch and real-time control problems of large-scale wind power grid connection, Ye Lin et al. proposed an optimal dispatch method for wind power clusters based on the time correlation of power fluctuations and stochastic model predictive control theory (FTC-SMPC).

[0004] Currently, the power grid requires wind farms to have a certain active power control capability. When the power grid issues power commands to the wind farm or the wind farm operates under power limitation, the wind farm needs to optimize and control the wind farm while meeting the power grid dispatch commands. Summary of the Invention

[0005] Technical Problem: In view of the problems and shortcomings of the existing technology, the purpose of this invention is to provide an active power optimization scheduling method based on the comprehensive evaluation index of the unit. Based on the real-time status evaluation and ranking of wind turbines and adopting a rolling optimization method, the method effectively reduces the number of unit control operations and improves the power command tracking accuracy of wind farms.

[0006] Technical Solution: To achieve the above-mentioned objectives, this invention proposes an active power optimization scheduling method based on comprehensive unit evaluation indicators. This method includes the following steps:

[0007] 1) Define unit evaluation indicators, including the cumulative amount of unit fatigue damage, unit regulation margin, including power up-adjustment margin and power down-adjustment margin, and unit power fluctuation;

[0008] 2) Construct the constraints that need to be considered for wind farm power dispatch, including overall farm output constraints, unit output constraints, and wind farm power variation constraints;

[0009] 3) Calculate the unit index values ​​based on the unit evaluation index definition, and combine them with constraints to optimize the scheduling of wind turbines in the wind farm while meeting the power demand of the power grid.

[0010] Furthermore, the unit evaluation indicators specifically defined in step 1) include:

[0011] A. Cumulative fatigue damage of the unit

[0012]

[0013] In the formula, a1 to a9 represent the coefficients of the fitting function, where a1 = 0.38, a2 = -0.0084, a3 = -0.059, a4 = 0.00535, a5 = -0.00014, a6 = 0.099, a7 = -0.184, a8 = 0.0126, and a9 = -0.00028; δ represents the power limiting degree; v represents the wind speed; f(δ,v) represents the fatigue damage per unit time of the blades operating under power limiting conditions at different wind speeds without considering turbulence, where the unit time is 1 minute.

[0014] The effects of turbulence are simplified to a turbulence fatigue damage factor K1. The fatigue damage considering turbulence is as follows:

[0015] f(I,δ,v)=K1·f(δ,v)

[0016] In the formula, f(I,δ,v) represents fatigue damage considering turbulence; K1 represents the turbulence fatigue damage factor, the value of which is obtained from interpolation in the table. When the turbulence intensity is 0, K1 is 0; when the turbulence intensity is 0.12, K1 is 1.1318; when the turbulence intensity is 0.14, K1 is 1.1794; when the turbulence intensity is 0.16, K1 is 1.2464.

[0017] f i (t)=f i (t -1 )+f i (I,δ,v)

[0018] In the formula, f i (t) represents the cumulative fatigue damage of unit i at present, f i (t -1 f represents the cumulative fatigue damage of unit i at the previous moment. i (I,δ,v) represents the fatigue damage of unit i in the current cycle;

[0019] B. Unit adjustment margin

[0020] Increase power margin:

[0021] ΔP 1i (t)=P i max (t+T)-P i (t)

[0022] Reduce power margin:

[0023]

[0024] In the formula, P i (t) represents the current power of the unit, P i max (t+T) represents the predicted power of the unit, where t is the current time, T is the control cycle length, and P... i min (t) represents the minimum output limit of the unit;

[0025] C. Unit power fluctuation:

[0026]

[0027]

[0028] In the formula, g i (t) represents the power standard deviation of unit i over the most recent 5 cycles, used to measure the power fluctuation of the unit, n is the number of cycles for which the power fluctuation of the computer group is obtained, t is the current time, and T is the control cycle length; This represents the average power output of unit i over the most recent 5 cycles.

[0029] Furthermore, the constraints specifically defined in step 2) include:

[0030] A. Overall power output constraints:

[0031]

[0032] B. Unit output constraints:

[0033]

[0034] C. Constraints on wind farm power variation:

[0035] |P(t+T)-P(t)|<ΔP m (t)

[0036] In the formula, P r (t) represents the total power demand of the wind farm at time t, P i min (t) represents the minimum output of the generator unit, P i max (t) represents the maximum output of the unit, P i (t) represents the unit output at time t, ΔP m (t) represents the maximum allowable power variation within the wind farm optimization period.

[0037] Furthermore, the specific method for step 3) is as follows:

[0038] Based on the unit evaluation index scheme, the comprehensive index value of the unit is calculated as follows:

[0039]

[0040]

[0041] In the formula, ΔP(t) is the difference between the current power of the wind farm and the power demand in the next cycle, P r (t+T) represents the power demand of the wind farm in the next cycle, k1, k2, and k3 are weighting coefficients with values ​​in the interval [0, 1], and f i (t) represents the cumulative fatigue damage of unit i at present, ΔP 1i (t) represents the power margin adjustment, ΔP 2i (t) represents the power margin reduction, g i (t) represents the power standard deviation of unit i over the most recent 5 cycles, h i (t) represents the comprehensive index value of the unit;

[0042] During operation, first compare the current wind farm output power with the grid demand P. commandDetermine whether the wind farm needs to increase or decrease its power generation (DP) based on its current power generation capacity:

[0043]

[0044] Among them, P Ireal Let n be the current power generation of each unit, and n be the number of units.

[0045] (1) If the wind farm needs to increase its power generation, i.e. DP<0, compare the current output power of each unit with the predicted power to determine the power increase or power decrease dP for each unit. i If unit i stops operating, i.e., S i =0, then the unit is ignored. 0 indicates the unit is shut down, and 1 indicates the unit is started. If the unit is operating normally, its power increase or decrease is:

[0046] dP i =P Ireal -P ifor

[0047] Among them, P Ireal P represents the current power generation of each generating unit. ifor Predicted power output for the generating units;

[0048] If the predicted power of unit i is less than the minimum output limit of the unit, i.e., P ifor <P imin If this happens, the unit will stop operating. At this time:

[0049] dP i =P ireal

[0050] Assume that the predicted power output of m generating units is less than the current power generation, i.e., dP j >=0, dP j If the power increase of unit j is the amount of increase, then the power increase required for the other nm units is:

[0051]

[0052] Wherein dP T The amount of power that needs to be increased for the additional nm units;

[0053] Evaluation is based on the power boosting response speed of the wind turbine, k = 1, ..., n, with priority scheduling h. i (t) Units whose value is less than the preset threshold;

[0054] (a) If the generator set k=1 can continue to increase its power generation, dP k <0, dP k This refers to the amount of power increase or decrease for unit k.

[0055] i) If dP k >=dP T Then, by increasing the power of unit k=1, the power requirement can be met. At this time:

[0056] P kref =P kreal +dP T

[0057] Among them, P kref P is the unit power setpoint. kreal The actual power output of unit k is given, while the current power generation of other units remains unchanged.

[0058] ii) If dP k <dP T Then, unit k will increase its power generation to the predicted value, that is:

[0059] P kref =P kfor

[0060] Among them, P Kfor The predicted power value for unit k;

[0061] The required power increase for the subsequent nm-1 units is as follows:

[0062]

[0063] (b) If unit k cannot continue to increase power generation, i.e., dP k If the value is greater than or equal to 0, then unit k generates electricity according to the predicted value, and similarly:

[0064] P kref =P kfor

[0065] If the predicted power of unit k is less than the minimum output limit of the unit, i.e., P kfor <P kmin If this happens, the unit will stop operating, that is:

[0066] P kref =0

[0067] The above operation is repeated for the next nm-1 units in sequence until the wind farm's power generation reaches the grid's demand, or all units operate at maximum output, i.e., the predicted power P of the units. ifor run;

[0068] (2) If the wind farm needs to reduce its power generation, i.e., DP>=0, compare the current output power of each unit with the lower limit of output P. imin Determine the power reduction amount for each unit. If unit i is shut down, i.e., S i=0, then ignore the unit. If the unit is operating normally, its power increase or decrease dP i for:

[0069] dP i =P ireal -P imin

[0070] If the predicted power of unit i is less than the minimum output limit of the unit, i.e., P ifor <P imin If this happens, the unit will stop operating. At this time:

[0071] dP i =P Ireal

[0072] Assuming that the predicted power output of m generating units is less than the minimum output limit, then the power reduction required for the remaining nm generating units is:

[0073]

[0074] (a) If dP T <0 indicates that another nm unit needs to increase its power |dP T |In order to meet the grid demand, the nm generator set is upgraded according to the method shown in (1) power upgrade step;

[0075] (b) If dP T >0, based on the wind turbine's power reduction response speed, prioritize scheduling h. i (t) Units whose t is less than the preset threshold,

[0076] If the predicted power of the unit (k=1) is greater than the minimum output limit of the unit, i.e., P kfor >P kmin Perform the following steps: i) ii) Reduce power consumption:

[0077] i) If dP T <= dP k If the unit k continues to reduce its power dPT, it can meet the grid demand, that is:

[0078] P kref =P kreal -dP T

[0079] The active power output of the other nm-1 units remains unchanged;

[0080] ii) If dP T >dP k Then unit k operates at its minimum output limit;

[0081] Pkref =P kmin

[0082] The power reduction required for the subsequent nm-1 units is as follows:

[0083]

[0084] The power reduction steps in step (2) are performed sequentially on the subsequent nm-1 units until the power demand is met, or all nm units are operating at the minimum output limit.

[0085] (c) If dP T =0, then the output of all nm units remains unchanged.

[0086] Beneficial effects: Compared with the prior art, the technical solution of the present invention has the following beneficial technical effects:

[0087] This invention studies the active power optimization scheduling and control problem of wind farms under power-limited conditions based on unit evaluation indicators and ranking control strategies. By ranking wind turbines based on real-time status evaluation and employing a rolling optimization method, the number of unit adjustments is effectively reduced, and the accuracy of wind farm power command tracking is improved. Attached Figure Description

[0088] Figure 1 The graph shows the power curve of the wind farm turbines. The cut-in wind speed is 3 m / s, corresponding to a power of 19.86 kW; the cut-out wind speed is 25 m / s.

[0089] Figure 2 A comparison chart of the final cumulative fatigue of each unit under different schemes;

[0090] Figure 3 The standard deviation of unit output during 144 optimization cycles under different schemes;

[0091] Figure 4 , 5 A comparison chart of the output of two randomly selected generating units under different operating conditions;

[0092] Figure 6 This is a flowchart of the method of the present invention. Detailed Implementation

[0093] The present invention will be further illustrated below with reference to the accompanying drawings and specific examples. It should be understood that these examples are for illustrative purposes only and are not intended to limit the scope of the invention. After reading this invention, any modifications of the invention in various equivalent forms by those skilled in the art will fall within the scope defined by the appended claims.

[0094] This invention proposes an active power optimization scheduling method based on the comprehensive evaluation index of generating units. The method includes the following steps:

[0095] 1) Define unit evaluation indicators, including the cumulative amount of unit fatigue damage, unit regulation margin, including power up-adjustment margin and power down-adjustment margin, and unit power fluctuation;

[0096] 2) Construct the constraints that need to be considered for wind farm power dispatch, including overall farm output constraints, unit output constraints, and wind farm power variation constraints;

[0097] 3) Calculate the unit index values ​​based on the unit evaluation index definition, and combine them with constraints to optimize the scheduling of wind turbines in the wind farm while meeting the power demand of the power grid.

[0098] Furthermore, the unit evaluation indicators specifically defined in step 1) include:

[0099] A. Cumulative fatigue damage of the unit

[0100]

[0101] In the formula, a1 to a9 represent the coefficients of the fitting function, where a1 = 0.38, a2 = -0.0084, a3 = -0.059, a4 = 0.00535, a5 = -0.00014, a6 = 0.099, a7 = -0.184, a8 = 0.0126, and a9 = -0.00028; δ represents the power limiting degree; v represents the wind speed; f(δ,v) represents the fatigue damage per unit time of the blades operating under power limiting conditions at different wind speeds without considering turbulence, where the unit time is 1 minute.

[0102] The effects of turbulence are simplified to a turbulence fatigue damage factor K1. The fatigue damage considering turbulence is as follows:

[0103] f(I,δ,ν)=K1·f(δ,ν)

[0104] In the formula, f(I,δ,ν) represents fatigue damage considering turbulence; K1 represents the turbulence fatigue damage factor, the value of which is obtained from interpolation in the table. When the turbulence intensity is 0, K1 is 0; when the turbulence intensity is 0.12, K1 is 1.1318; when the turbulence intensity is 0.14, K1 is 1.1794; when the turbulence intensity is 0.16, K1 is 1.2464.

[0105] f i (t)=f i (t -1 )+f i (I,δ,v)

[0106] In the formula, fi (t) represents the cumulative fatigue damage of unit i at present, f i (t -1 f represents the cumulative fatigue damage of unit i at the previous moment. i (I,δ,v) represents the fatigue damage of unit i in the current cycle;

[0107] B. Unit adjustment margin

[0108] Increase power margin:

[0109] ΔP 1i (t)=P i max (t+T)-P i (t)

[0110] Reduce power margin:

[0111]

[0112] In the formula, P i (t) represents the current power of the unit, P i max (t+T) represents the predicted power of the unit, where t is the current time, T is the control cycle length, and P... i min (t) represents the minimum output limit of the unit;

[0113] C. Unit power fluctuation:

[0114]

[0115]

[0116] In the formula, g i (t) represents the power standard deviation of unit i over the most recent 5 cycles, used to measure the power fluctuation of the unit, n is the number of cycles for which the power fluctuation of the computer group is obtained, t is the current time, and T is the control cycle length; This represents the average power output of unit i over the most recent 5 cycles.

[0117] Furthermore, the constraints specifically defined in step 2) include:

[0118] A. Overall power output constraints:

[0119]

[0120] B. Unit output constraints:

[0121]

[0122] C. Constraints on wind farm power variation:

[0123] |P(t+T)-P(t)|<ΔP m (t)

[0124] In the formula, P r (t) represents the total power demand of the wind farm at time t, P i min (t) represents the minimum output of the generator unit, P i max (t) represents the maximum output of the unit, P i (t) represents the unit output at time t, ΔP m (t) represents the maximum allowable power variation within the wind farm optimization period.

[0125] Furthermore, the specific method for step 3) is as follows:

[0126] Based on the unit evaluation index scheme, the comprehensive index value of the unit is calculated as follows:

[0127]

[0128]

[0129] In the formula, ΔP(t) is the difference between the current power of the wind farm and the power demand in the next cycle, P r (t+T) represents the power demand of the wind farm in the next cycle, k1, k2, and k3 are weighting coefficients with values ​​in the interval [0, 1], and f i (t) represents the cumulative fatigue damage of unit i at present, ΔP 1i (t) represents the power margin adjustment, ΔP 2i (t) represents the power margin reduction, g i (t) represents the power standard deviation of unit i over the most recent 5 cycles, h i (t) represents the comprehensive index value of the unit;

[0130] During operation, first compare the current wind farm output power with the grid demand P. command Determine whether the wind farm needs to increase or decrease its power generation (DP) based on its current power generation capacity:

[0131]

[0132] Among them, P Ireal Let n be the current power generation of each unit, and n be the number of units.

[0133] (1) If the wind farm needs to increase its power generation, i.e. DP<0, compare the current output power of each unit with the predicted power to determine the power increase or power decrease dP for each unit. i If unit i stops operating, i.e., S i=0, then the unit is ignored. 0 indicates the unit is shut down, and 1 indicates the unit is started. If the unit is operating normally, its power increase or decrease is:

[0134] dP i =P Ireal -P ifor

[0135] Among them, P Ireal P represents the current power generation of each generating unit. ifor Predicted power output for the generating units;

[0136] If the predicted power of unit i is less than the minimum output limit of the unit, i.e., P ifor <P imin If this happens, the unit will stop operating. At this time:

[0137] dP i =P ireal

[0138] Assume that the predicted power output of m generating units is less than the current power generation, i.e., dP j >=0, dP j If the power increase of unit j is the amount of increase, then the power increase required for the other nm units is:

[0139]

[0140] Wherein dP T The amount of power that needs to be increased for the additional nm units;

[0141] Evaluation is based on the power boosting response speed of the wind turbine, k = 1, ..., n, with priority scheduling h. i (t) Units whose value is less than the preset threshold;

[0142] (a) If the generator set k=1 can continue to increase its power generation, dP k <0, dP k This refers to the amount of power increase or decrease for unit k.

[0143] i) If dP k >=dP T Then, by increasing the power of unit k=1, the power requirement can be met. At this time:

[0144] P kref =P kreal +dP T

[0145] Among them, P kref P is the unit power setpoint. kreal The actual power output of unit k is given, while the current power generation of other units remains unchanged.

[0146] ii) If dP k <dP T Then, unit k will increase its power generation to the predicted value, that is:

[0147] P kref =P kfor

[0148] Among them, P Kfor The predicted power value for unit k;

[0149] The required power increase for the subsequent nm-1 units is as follows:

[0150]

[0151] (b) If unit k cannot continue to increase power generation, i.e., dP k If the value is greater than or equal to 0, then unit k generates electricity according to the predicted value, and similarly:

[0152] P kref =P kfor

[0153] If the predicted power of unit k is less than the minimum output limit of the unit, i.e., P kfor <P kmin If this happens, the unit will stop operating, that is:

[0154] P kref =0

[0155] The above operation is repeated for the next nm-1 units in sequence until the wind farm's power generation reaches the grid's demand, or all units operate at maximum output, i.e., the predicted power P of the units. ifor run;

[0156] (2) If the wind farm needs to reduce its power generation, i.e., DP>=0, compare the current output power of each unit with the lower limit of output P. imin Determine the power reduction amount for each unit. If unit i is shut down, i.e., S i =0, then ignore the unit. If the unit is operating normally, its power increase or decrease dP i for:

[0157] dP i =P ireal -P imin

[0158] If the predicted power of unit i is less than the minimum output limit of the unit, i.e., P ifor <P imin If this happens, the unit will stop operating. At this time:

[0159] dP i =P Ireal

[0160] Assuming that the predicted power output of m generating units is less than the minimum output limit, then the power reduction required for the remaining nm generating units is:

[0161]

[0162] (a) If dP T <0 indicates that another nm unit needs to increase its power |dP T |In order to meet the grid demand, the nm generator set is upgraded according to the method shown in (1) power upgrade step;

[0163] (b) If dP T >0, based on the wind turbine's power reduction response speed, prioritize scheduling h. i (t) Units whose t is less than the preset threshold,

[0164] If the predicted power of the unit (k=1) is greater than the minimum output limit of the unit, i.e., P kfor >P kmin Perform the following steps: i) ii) Reduce power consumption:

[0165] i) If dP T <= dP k If the unit k continues to reduce its power dPT, it can meet the grid demand, that is:

[0166] P kref =P kreal -dP T

[0167] The active power output of the other nm-1 units remains unchanged;

[0168] ii) If dP T >dP k Then unit k operates at its minimum output limit;

[0169] P kref =P kmin

[0170] The power reduction required for the subsequent nm-1 units is as follows:

[0171]

[0172] The power reduction steps in step (2) are performed sequentially on the subsequent nm-1 units until the power demand is met, or all nm units are operating at the minimum output limit.

[0173] (c) If dP T =0, then the output of all nm units remains unchanged.

[0174] Example

[0175] To verify the effectiveness of the proposed wind farm power optimization control method based on turbine classification and ranking, wind speed data from SCADA records of a wind farm in Northwest my country, collected from 00:00 to 23:50 on February 10, 2016, were analyzed. The data acquisition period was 10 minutes. This wind farm has 25 2MW turbines with a total installed capacity of 50MW. The turbine power curves are shown below. Figure 1 As shown, its cut-in wind speed is 3 m / s, corresponding to a power of 19.86 KW; the cut-out wind speed is 25 m / s.

[0176] Based on different unit evaluation indicators, four sets of schemes were set up for comparison:

[0177] Option 1: Perform fatigue balance optimization only. This option prioritizes units based on their accumulated fatigue damage. When it is necessary to increase wind farm power generation, priority is given to increasing the output of units with lower fatigue levels; when it is necessary to reduce wind farm power generation, priority is given to reducing the output of units with higher fatigue levels.

[0178] Option 2: Reduce only the difference in unit output fluctuation. This option prioritizes units with smaller output fluctuations based on the magnitude of their output fluctuations, and when wind farm power needs to be adjusted, it prioritizes units with smaller output fluctuations.

[0179] Option 3: Considering only the margin. This option prioritizes units based on whether their output margin can be increased or decreased. When it is necessary to increase the wind farm's power generation, priority is given to increasing the output of units with a larger output margin; when it is necessary to decrease the wind farm's power generation, priority is given to decreasing the output of units with a larger output margin.

[0180] Option 4: Determine the unit adjustment sequence based on a comprehensive ranking of the first three indicators.

[0181] The actual power and power demand of the entire field will be compared under four control schemes and proportional allocation, and the relative errors will be compared.

[0182] Overall, the total power output under different schemes is relatively close and basically reaches the total power command. A comparison of the total power output and relative error analysis for each scheme shows that the relative error distribution of the four schemes is almost identical. In contrast, the traditional proportional allocation method has a larger error between the power allocation result and the power command, and the error exceeds the upper and lower limits more widely. During the power regulation process of a wind farm, the ability of the turbines to increase power is more limited than their ability to decrease power due to the influence of wind speed. Therefore, the phenomenon of relative error exceeding the lower limit is more common in wind farm power control. Especially in the low wind speed range, the turbine power regulation capability is poor and many turbines are shut down, resulting in a large deviation between the actual output of the wind farm and the total power command during this period. During periods of good wind conditions, the actual power output of the wind farm can track the power command. The results show that the wind farm power control method proposed in this patent can effectively reduce the power tracking control error of the wind farm.

[0183] Figure 2 This chart compares the final cumulative fatigue values ​​of each unit under different schemes. As can be seen from the chart, Scheme 1, considering only fatigue balance, shows the smallest difference in the final cumulative fatigue value curves among the wind farm units. Compared to the other four schemes, its final cumulative fatigue value distribution is the most balanced. This scheme prioritizes increasing the output of units with lower fatigue values ​​when increasing wind farm power generation, and prioritizes reducing the output of units with higher fatigue values ​​when reducing wind farm power generation. Scheme 4, a comprehensive optimization, is the next best, followed by Scheme 2. Scheme 3, considering only margin, shows the largest difference in fatigue value distribution.

[0184] Figure 3 The figure represents the standard deviation of unit output over 144 optimization cycles under different schemes. The greater the fluctuation in unit output during operation, the larger the final standard deviation. As shown in the figure, in Scheme 2, which only considers reducing unit output fluctuation, the overall output fluctuation of each unit is relatively smaller compared to other schemes. This scheme prioritizes units with smaller output fluctuations based on their output fluctuation magnitude, and when adjusting wind farm power, it prioritizes units with smaller output fluctuations. Scheme 4, under comprehensive optimization, has the second largest overall fluctuation.

[0185] Figure 4 , 5 The figures show a comparison of the power output of two randomly selected units under different operating schemes. It can be seen that the power curve of the unit fluctuates the least under Scheme 2. This scheme has a significant effect on reducing the power output fluctuation of the unit, and the optimization effect is mainly reflected in the first half of the period. Since the wind speed is high during this period, there is almost no shutdown, and there is a large optimization space.

Claims

1. An active power optimization scheduling method based on unit comprehensive evaluation indicators, characterized in that, The method includes the following steps: 1) Define unit evaluation indicators, including the cumulative amount of unit fatigue damage, unit regulation margin, including power up-adjustment margin and power down-adjustment margin, and unit power fluctuation; 2) Construct the constraints that need to be considered for wind farm power dispatch, including overall farm output constraints, unit output constraints, and wind farm power variation constraints; 3) Calculate the unit index values ​​based on the unit evaluation index definition, and combine the constraints to optimize the scheduling of wind turbines in the wind farm while meeting the power demand of the grid. The unit evaluation indicators specifically defined in step 1) include: A. Cumulative fatigue damage of the unit ； In the formula, a1~a9 represent the coefficients of the fitting function, where a1=0.38, a2=-0.0084, a3=-0.059, a4=0.00535, a5=-0.00014, a6=0.099, a7=-0.184, a8=0.0126, a9=-0.00028; δ represents the power limiting degree; v represents the wind speed; This represents the amount of fatigue damage per unit time for blades operating at different wind speeds under power-limited conditions, without considering turbulence; where the unit time is 1 minute. The effects of turbulence are simplified into a turbulence fatigue damage factor. Considering the fatigue damage caused by turbulence: ； In the formula, This indicates fatigue damage considering turbulence; K1 represents the turbulence fatigue damage factor, with values ​​obtained from interpolation in the table. When the turbulence intensity is 0, K1 is 0; when the turbulence intensity is 0.12, K1 is 1.1318; when the turbulence intensity is 0.14, K1 is 1.1794; and when the turbulence intensity is 0.16, K1 is 1.2464. ； In the formula, f i (t) represents the cumulative fatigue damage of unit i at present, f i (t -1 f represents the cumulative fatigue damage of unit i at the previous moment. i (I,δ,v) represents the fatigue damage of unit i in the current cycle; B. Unit adjustment margin Increase power margin: ； Reduce power margin: ； In the formula, P i (t) represents the current power of the unit, P i max (t+T) represents the predicted power of the unit, where t is the current time, T is the control cycle length, and P... i min (t) represents the minimum output limit of the unit; C. Unit power fluctuation: ; ； In the formula, g i (t) represents the power standard deviation of unit i over the most recent 5 cycles, used to measure the power fluctuation of the unit, n is the number of cycles for which the power fluctuation of the computer group is obtained, t is the current time, and T is the control cycle length; This represents the average power output of unit i over the most recent 5 cycles. The constraints specifically defined in step 2) include: A. Overall power output constraints: ； B. Unit output constraints: ； C. Constraints on wind farm power variation: ； In the formula, P r (t) represents the total power demand of the wind farm at time t, P i min (t) represents the minimum output of the generator unit, P i max (t) represents the maximum output of the unit, P i (t) represents the unit output at time t, ∆P m (t) represents the maximum allowable power variation within the wind farm optimization period.

2. The active power optimization scheduling method based on the comprehensive evaluation index of generating units according to claim 1, characterized in that, The specific method for step 3) is as follows: Based on the unit evaluation index scheme, the comprehensive index value of the unit is calculated as follows: ； ； In the formula, ΔP(t) is the difference between the current power of the wind farm and the power demand in the next cycle, P r (t+T) represents the power demand of the wind farm in the next cycle, k1, k2, and k3 are weighting coefficients with values ​​in the interval [0, 1], and f i (t) represents the cumulative fatigue damage of unit i at present, ΔP 1i (t) represents the power margin for adjustment, ΔP 2i (t) represents the power margin reduction, g i (t) represents the power standard deviation of unit i over the most recent 5 cycles, h i (t) represents the unit's comprehensive performance index value; At runtime, first compare the current wind farm output power and grid demand P command Determine whether the wind farm needs to increase or decrease power generation DP based on the current power generation ； where P Ireal is the current power generation of each unit, and n is the number of units. (1) If the wind farm needs to increase its power generation, i.e. DP<0, compare the current output power of each unit with the predicted power to determine the power increase or power decrease dP of each unit. i If unit i stops operating, i.e., S i =0, then the unit is ignored. 0 indicates the unit is shut down, and 1 indicates the unit is started. If the unit is operating normally, its power increase or decrease is: ； P Ireal is the current power generation of each unit, P ifor is the predicted power of the unit; If the predicted power of the unit i is less than the minimum limit of the unit output, i.e. P ifor <P imin , the unit will stop running at this time: ； Assume that the predicted power output of m generating units is less than the current power generation, i.e., dP j >=0, dP j If the power increase of unit j is the amount of increase, then the power increase required for the other nm units is: ； wherein dP T is the power size needed to be raised for the other n-m units; According to the response speed of the wind turbine to improve power, k = 1, …, n, the priority scheduling h i (t) less than the preset threshold unit (a) if the unit k = 1 can continue to increase power generation, dP k <0, dP k is the power increase or power decrease amount of the unit k; i) if dP k >= dP T , then the power demand is met by increasing the power of the hoisting group k = 1, in which case: ； P kref is the power set value of the unit, P kreal is the actual power of the k unit, and the other units remain unchanged. ii) If dP k <dP T Then, unit k will increase its power generation to the predicted value, that is: ； Among them, P Kfor The predicted power value for unit k; The required power increase for the subsequent nm-1 units is as follows: ； (b) If unit k cannot continue to increase power generation, i.e., dP k If the value is greater than or equal to 0, then unit k generates electricity according to the predicted value, and similarly: ； If the predicted power of unit k is less than the minimum output limit of the unit, i.e., P kfor <P kmin If this happens, the unit will stop operating, that is: ； The above operation is repeated for the next nm-1 units in sequence until the wind farm's power generation reaches the grid's demand, or all units operate at maximum output, i.e., the predicted power P of the units. ifor run; (2) If the wind farm needs to reduce power generation, i.e. DP>=0, compare the current output power of each unit with the lower limit of output P. imin Determine the power reduction amount for each unit. If unit i is shut down, i.e., S i If the value is 0, then the unit is ignored. If the unit is operating normally, its power increase or decrease dP is... i for: ； If the predicted power of unit i is less than the minimum output limit of the unit, i.e., P ifor <P imin If this happens, the unit will stop operating. At this time: ； Assuming that the predicted power output of m generating units is less than the minimum output limit, then the power reduction required for the remaining nm generating units is: ； (a) If dP T <0 indicates that another nm unit needs to increase its power |dP T |In order to meet the grid demand, the nm generator set is upgraded according to the method shown in (1) power upgrade step; (b) If dP T >0, based on the wind turbine's power reduction response speed, prioritize scheduling h. i (t) Units whose t is less than the preset threshold, If the predicted power of the unit (k=1) is greater than the minimum output limit of the unit, i.e., P kfor >P kmin Perform the following steps: i) ii) Reduce power consumption: i) If dP T <=dP k If the unit k continues to reduce its power dPT, it can meet the grid demand, that is: ； The active power output of the other nm-1 units remains unchanged; ii) If dP T >dP k Then unit k operates at its minimum output limit; ； The power reduction required for the subsequent nm-1 units is as follows: ； The power reduction steps in step (2) are performed sequentially on the subsequent nm-1 units until the power demand is met, or all nm units are operating at the minimum output limit. (c) If dP T If =0, then the output of all nm units remains unchanged.

Images