A unit commitment method considering transient frequency constraint and optimal load shedding of new energy

Abstract

The application discloses a unit commitment method considering transient frequency constraint and optimal load shedding of new energy, and belongs to the field of safe and economic dispatching of power systems; the method comprises the following steps: 1, predicting the load output and the maximum output of new energy on a certain day, setting the maximum load shedding percentage of the new energy, and obtaining the operation state of each unit and the output condition of the new energy unit according to the first layer optimization; 2, checking whether the minimum point frequency meets the transient frequency constraint according to the operation state of the unit and the output condition of the new energy, if the condition is met, not carrying out the second layer optimization, and outputting the result; 3, if the frequency requirement is not met, carrying out the second layer optimization to obtain the optimal load shedding state of the unit, and checking whether the minimum point frequency meets the transient frequency constraint; if the condition is met, carrying out step 1, and if the requirement is not met, increasing the maximum load shedding percentage; and 4, repeating steps 1 to 3. Compared with the traditional unit commitment strategy, the application is more economical, and power generation is safer and more reliable.

Description

Technical Field

[0001] This invention belongs to the field of power system operation and dispatching, and in particular relates to a unit combination method that considers transient frequency constraints and optimal load reduction of new energy sources. Background Technology

[0002] Unit combination refers to the arrangement of unit start-up and shutdown status and output on the day-ahead dispatch timescale. Currently, most traditional unit combination-based power generation plans only meet the system's conventional safety constraints. With the integration of a high proportion of renewable energy, the transient stability of the power system should also receive greater attention. Therefore, unit combination should also meet transient frequency constraints to enhance the ability of the new power system with a high proportion of renewable energy to resist large-scale power disturbances. The scarcity of traditional frequency regulation resources has incentivized more diversified frequency regulation resources to participate in the power system's frequency regulation process. The prerequisite for renewable energy plants / stations to participate in frequency regulation is load shedding and reserving sufficient frequency regulation reserves. However, the determination of load shedding under frequency constraints and its coordination with traditional generating units have not been effectively resolved.

[0003] In summary, there is an urgent need among existing scheduling methods for a unit combination optimization strategy that considers transient frequency constraints and optimal load reduction of new energy sources. Summary of the Invention

[0004] The purpose of this invention is to provide a unit combination method that considers transient frequency constraints and optimal load reduction of new energy sources, so as to solve the technical problems mentioned in the background art.

[0005] To achieve the above objectives, the specific technical solution of the unit combination method of the present invention, which considers transient frequency constraints and optimal load reduction of new energy sources, is as follows:

[0006] The unit combination optimization strategy of this invention, which considers transient frequency constraints and optimal load reduction of new energy sources, is as follows:

[0007] like Figures 1-7 As shown, this invention constructs a unit combination optimization model considering dynamic frequency constraints based on the traditional unit combination model. When solving the optimization model, an atomic search algorithm is introduced to collaboratively consider the optimal load shedding optimization of new energy sources supported by frequency and the unit combination optimization, establishing a two-layer optimization strategy. Compared with the traditional unit combination strategy, this novel unit combination strategy provides a more economical unit combination method and safer and more reliable power generation.

[0008] The unit combination strategy of this invention simultaneously considers transient frequency constraints and optimal load shedding of renewable energy sources. The specific steps of the unit combination method considering transient frequency constraints and optimal load shedding of renewable energy sources are as follows:

[0009] Step 1: Predict the load output and maximum output of new energy sources on a certain day, set the maximum load reduction percentage of new energy sources (initially set to 0), and obtain the operating status of each unit and the output of new energy units based on the first layer of optimization.

[0010] Step one, based on the first-level optimization to obtain the operating status of each unit and the output of the new energy units, includes the following steps:

[0011] First, you need to input the load forecast for a certain day, the maximum output of new energy sources, and the maximum load reduction percentage obtained from the initial settings or the second-level optimization.

[0012] The objective function for the first layer of optimization is the power generation cost of thermal power units.

[0013] (1)

[0014] In the formula This represents the on / off status of the thermal power unit at time t, where 1 indicates on and 0 indicates off. , , Let be the power generation cost coefficient of thermal power unit i. , For the start-up and shutdown costs of thermal power unit i, Let represent the power generation of the thermal power unit at time t (i = 1). To ensure maximum power generation from renewable energy plants / stations, the cost of wind and solar power generation is set to 0.

[0015] The first layer of optimization constraints includes power balance constraints, output constraints of thermal power units and new energy power plants / stations, ramping constraints, start-up and shutdown time constraints, and reserve constraints.

[0016] (2)

[0017] (3)

[0018] (4)

[0019] (5)

[0020] (6)

[0021] (7)

[0022] Equation (2) is the power balance constraint, Equations (3) and (4) are the output constraints of thermal power units and new energy power plants / stations, respectively, Equation (5) is the ramp constraint, Equation (6) is the start-stop time constraint, and Equation (7) is the standby constraint.

[0023] in, , , This represents the output of wind turbines, photovoltaic power plants, and wind farms at time t. This represents the system load at time t. , This represents the lower and upper limits of the output of thermal power unit i. , This represents the uphill and downhill power of thermal power unit i. , This indicates the continuous start-up and shutdown time of a thermal power unit. Indicates the reserve factor. , This represents the percentage of load reduction for both photovoltaic power plants and wind farms at time t. , The power generation units of wind farms and photovoltaic power plants are respectively tracking the power generation of their operating states at maximum power. , These represent the actual power generation of wind farms and photovoltaic power plants, respectively, with m, n, and k indicating the number of wind power, photovoltaic, and thermal power plants.

[0024] The optimization process is solved using the cplex solver in MATLAB, and the on / off status of the unit is obtained by inputting the known quantities.

[0025] Then proceed to the second step, step two: check whether the minimum point frequency meets the transient frequency constraint based on the unit's operating status and the output of new energy sources. If the condition is met, no second-level optimization is needed, and the result is output.

[0026] Specifically, the unit's start-up and shutdown states and the known load reduction are substituted into the expression for the transient frequency constraint:

[0027] (8)

[0028] (9)

[0029] (10)

[0030] (11)

[0031] in, The lowest frequency; The system reference frequency; The time to reach the lowest frequency point. This is the system's power deficit; This is the equivalent inertial response time constant; , Intermediate variables for ease of calculation; It is the natural vibration frequency; The damping ratio; This is the load damping coefficient; For the first k The inertial time constant of a thermal power unit; For thermal power units i time t The unit's on / off status; For the first k The capacity of the thermal power unit; The capacity of the entire system; The time constant of the speed controller; , and For parameters , and Substitute t Constantly monitor the on / off status of thermal power units And the values ​​after the load reduction of new energy power plants / stations; and The first i Photovoltaic power stations and wind farms at any time t The adjustment coefficient; For the first i Mechanical power gain factor of thermal power units in Taiwan; For the first i The fraction of the total power generated by the high-pressure steam turbine; The first i The droop coefficient of a thermal power unit; This indicates the on / off status of all thermal power units. This indicates the load reduction status of new energy power plants / stations.

[0032] If formula (11) is satisfied, it means that the combination of the obtained start-up and shutdown states and the load reduction satisfies the transient frequency constraint, and the result can be output.

[0033] The process for obtaining the transient frequency constraint in step two is as follows:

[0034] A multi-machine frequency response model of a complex power system is established by combining conventional thermal power units with new energy units. Taking a power system that simultaneously includes thermal power units, wind farms, and photovoltaic power plants as an example, the multi-machine frequency response model is as follows: Figure 1 .

[0035] This model can reflect the inertial support of thermal power units and the primary frequency regulation response process of all units. Figure 1 The transfer function is expressed as equation (16):

[0036] (16)

[0037] in Let D be the system's inertial time constant, and D be the load damping coefficient. Let be the system frequency offset at time t. The power deficit of the system at time t; and The first i Adjustment coefficients for individual photovoltaic power plants and wind farms; , and The first i Time constants of speed governors for thermal power plants, photovoltaic power plants, and wind power plants; m, n, k This indicates the number of wind turbine units, photovoltaic power plants, and thermal power units; For the first i Mechanical power gain factor of thermal power units in Taiwan; For the first i The fraction of the total power generated by the high-pressure steam turbine; The first i The droop coefficient of the thermal power unit.

[0038] Within permissible limits, the time constants of different speed controllers have little impact on the lowest point of the system frequency. To simplify calculations, the time constants of all unit speed controllers in equation (8) can be determined by the same constant value T. R Substitution. Simplifying equation (8) yields Frequency domain expression:

[0039] (17)

[0040] (18)

[0041] (19)

[0042] In the formula: Let D be the system's inertial time constant, and D be the load damping coefficient. Let be the system frequency offset at time t. The power deficit of the system at time t; and The first i Adjustment coefficients for individual photovoltaic power plants and wind farms; , and The first i Time constants of speed governors for thermal power plants, photovoltaic power plants, and wind power plants; m、 n, k This indicates the number of wind turbine units, photovoltaic power plants, and thermal power units; For the first i Mechanical power gain factor of thermal power units in Taiwan; For the first iThe fraction of the total power generated by the high-pressure steam turbine; The first i The droop coefficient of a thermal power unit; It is the natural vibration frequency; is the damping ratio.

[0043] Performing an inverse Laplace transform on the frequency domain expression yields information about... The time-domain expression (12) is used to obtain the lowest frequency in the frequency response process. The derivative of the expression is then set to... The lowest frequency is obtained under the conditions of power deficit and fixed unit operating status:

[0044] (8)

[0045] (9)

[0046] The on / off status of thermal power units And the unloaded state of new energy power plants / stations is introduced into equations (14) and (15) to obtain , and ;

[0047] (10)

[0048] in, The lowest frequency; The system reference frequency; The time to reach the lowest frequency point. This is the system's power deficit; This is the equivalent inertial response time constant; , Intermediate variables for ease of calculation; It is the natural vibration frequency; The damping ratio; This is the load damping coefficient; For the first k The inertial time constant of a thermal power unit; For thermal power units i time t The unit's on / off status; For the first k The capacity of the thermal power unit; The capacity of the entire system; The time constant of the speed controller; , and For parameters , and Substitute tConstantly monitor the on / off status of thermal power units And the values ​​after the load reduction of new energy power plants / stations; and The first i Photovoltaic power stations and wind farms at any time t The adjustment coefficient; For the first i Mechanical power gain factor of thermal power units in Taiwan; For the first i The fraction of the total power generated by the high-pressure steam turbine; The first i The droop coefficient of a thermal power unit; This indicates the on / off status of all thermal power units. This indicates the load reduction status of new energy power plants / stations.

[0049] Substituting equation (10) into equation (11), the transient frequency constraint is expressed as:

[0050] (11).

[0051] If formula (11) is not satisfied, then proceed to the second layer of optimization in step three;

[0052] Step 3: If the frequency requirement is not met, perform a second-level optimization to obtain the unit's optimal load reduction state and check whether the minimum frequency meets the transient frequency constraint. If the condition is met, proceed to Step 1; if the requirement is not met, increase the maximum load reduction percentage and then run Step 1.

[0053] The specific principle of the second-level optimization is as follows: based on the unit start-up and shutdown status obtained from the first-level optimization as the input, the objective function is to minimize the load reduction of the new energy plant / station, and the transient frequency is used as the constraint. Within a certain load reduction space, the optimal load reduction is found. The objective function of the optimal load reduction process of new energy is Equation (12), the transient frequency constraint is Equation (11), and the load reduction space is Equation (13).

[0054] (12)

[0055] In the formula , This represents the percentage of load reduction for both photovoltaic power plants and wind farms at time t. , The power generation units of wind farms and photovoltaic power plants are respectively tracking the power generation of their operating states at maximum power. , This is the load reduction cost coefficient for wind farms and photovoltaic power plants;

[0056] (13)

[0057] In the formula , They are respectively t The percentage of load reduction for wind farms and solar power plants at any given time; , This refers to the maximum allowable load reduction for wind farms and photovoltaic power plants.

[0058] The specific relationship between the load reduction percentage d and the sag coefficient R is as follows:

[0059] In economic dispatching, unit combination, and planning, new energy units typically adopt a simplified model simulating the governor of a synchronous unit, namely:

[0060] (20)

[0061] Among them, R n T is the adjustment coefficient for new energy power plants / stations. n The controller time constant, For the power adjustment amount of new energy power plants / stations, This is the frequency offset.

[0062] Similar to the governor of a conventional synchronous generator set, droop control supplies power to the grid in a manner that is proportional to the reserve capacity and frequency deviation.

[0063] (twenty one)

[0064] (twenty two)

[0065] In the formula: For the load reduction of new energy power generation units, This refers to the power output of the new energy power generation unit under maximum operating conditions. This represents the percentage of load reduction. The power output for droop control. This is a frequency deviation signal. This is the adjustment coefficient.

[0066] Based on the ASO algorithm, a penalty function method is introduced to handle inequalities and equality constraints in the model, so as to solve the second-level optimization model.

[0067] The load reduction optimization takes the minimum wind and solar load reduction, i.e., equation (12), as the objective function. To ensure transient frequency stability, inequality constraints (11) are added, and new second-level optimization objective functions (14) and (15) are formed through the penalty factor M:

[0068] (11)

[0069] (14)

[0070] (15)

[0071] In the formula Indicates the lowest permissible frequency of the system; Indicates the lowest frequency point; penalty factor M When frequency constraints are not met m The value of is guaranteed to be much greater than ; This indicates the cost of load reduction; a maximum load reduction amount is set to meet the constraints.

[0072] To avoid excessive load reduction and waste of new energy resources, a maximum load reduction should be set to meet the following constraints:

[0073] (13)

[0074] In the formula , They are respectively t The percentage of load reduction for wind farms and solar power plants at any given time; , This refers to the maximum allowable load reduction for wind farms and photovoltaic power plants.

[0075] Determining the optimal load reduction for new energy sources is a nonlinear optimization problem that includes transient frequency constraints. This invention uses the Atom Search Algorithm (ASO) to solve the second-level optimization, determine the optimal load reduction, and obtain the optimal load reduction percentage for new energy farms / stations within a finite load reduction interval.

[0076] Furthermore, based on the atomic search optimization algorithm, a penalty function method is introduced to handle inequalities and equality constraints in the model, so as to solve the lower-level optimization model.

[0077] The Atomic Dynamics Algorithm (ASO) is a metaheuristic global optimization method based on atomic dynamics. This method outperforms other intelligent algorithms in parameter estimation problems. By introducing a penalty function method to handle inequalities and equality constraints in the model, a solution to the model can be achieved.

[0078] The specific steps of the ASO algorithm for solving the optimal load reduction problem using a penalty function are as follows:

[0079] S1. Determine the variables and objective function. The variable is the load reduction ratio of the new energy units, the load reduction space is Equation (13), and the dimension of the intelligent algorithm is the number of new energy units. The objective function is Equation (14).

[0080] S2. Configure the parameters. In addition to the common parameters for all optimization algorithms, ASO only requires setting two parameters: depth weights and Lagrange multiplier weights.

[0081] S3. After determining the number of atoms N, assign each atom an initial position (i.e., an initial solution to the constraint problem) and velocity.

[0082] S4. Input the position information of the atoms into the fitness function (i.e., the objective function), sort them according to the value of the function from best to worst, and take the first K atoms as the optimal subset.

[0083] S5. Calculate the mass of each atom using the atom search optimization algorithm.

[0084] S6. Calculate the acceleration of each atom in each dimension according to the algorithm.

[0085] S7. After determining the acceleration of the atom, redetermine the velocity and position of the atom.

[0086] S8. Continue iterating according to 4-7 until T iterations, then select the optimal solution.

[0087] The second level of optimization can obtain the optimal load reduction. Optimal load reduction Substitute into Equation 11 to see if the transient frequency constraint is met. If the transient frequency constraint is met, return directly to step one for the first layer of optimization. If the transient frequency constraint is not met, the load reduction space needs to be increased, and then return to step one for the first layer of optimization.

[0088] The unit combination method of this invention, which considers transient frequency constraints and optimal load shedding of renewable energy sources, has the following advantages: Based on the traditional unit combination model, this invention constructs a unit combination optimization model that considers dynamic frequency constraints; when solving the optimization model, an atomic search algorithm is introduced to collaboratively consider the optimal load shedding of renewable energy sources supported by frequency and the unit combination optimization, establishing a two-layer optimization strategy. Compared with traditional unit combination strategies, this novel unit combination strategy provides a more economical unit combination method and more safe and reliable power generation.

[0089] The proposed dual-layer optimization strategy divides the linear constraints in traditional unit combination and the nonlinear constraints in finding the optimal load reduction of new energy sources into two layers of optimization, which facilitates calculation. Neither of these two optimizations can guarantee that the transient frequency will not exceed the limit. Therefore, the method of layered optimization plus constraint verification can take into account both optimizations and ensure that the transient frequency does not exceed the limit. Attached Figure Description

[0090] Figure 1 This is a multi-machine frequency response model for a power system that simultaneously includes thermal power units, wind farms, and photovoltaic power plants.

[0091] Figure 2 This is a flowchart of a unit combination method that considers transient frequency constraints and optimal load reduction of new energy sources.

[0092] Figure 3 The load curve and the daily power generation forecast of the new energy power station are used as examples in the implementation of this invention.

[0093] Figures 4(a)-4(c) show three different unit combination schemes in the implementation examples of this invention. (Figure 4(a) shows the unit combination under only conventional constraints; Figure 4(b) shows the unit combination under conventional constraints and transient frequency constraints, where the new energy source is operating at full load and cannot regulate the frequency; Figure 4(c) shows the unit combination under conventional constraints and transient frequency constraints, where the new energy source reduces its load and regulates the frequency.)

[0094] Figures 5(a)-5(c) show the minimum frequency distribution of three different unit combination schemes in the implementation examples of the present invention (Figure 5(a) shows the minimum frequency distribution of scheme a, Figure 5(b) shows the minimum frequency distribution of scheme b, and Figure 5(c) shows the minimum frequency distribution of scheme c).

[0095] Figure 6 Examples b and c of the present invention are new energy power generation scenarios.

[0096] Figure 7 The number of thermal power units started in the implementation examples a, b, and c of this invention. Detailed Implementation

[0097] To better understand the purpose, structure, and function of this invention, the following detailed description, in conjunction with the accompanying drawings, provides a method for combining generating units that considers transient frequency constraints and optimal load reduction of new energy sources.

[0098] like Figure 2 As shown, this invention provides a unit combination method that considers transient frequency constraints and optimal load shedding from new energy sources. It simultaneously considers the long-term stability and short-term frequency stability of the power system, incorporating both linear and nonlinear constraints. Specifically, the method consists of two optimization layers. The first layer optimizes the unit combination based on conventional safety constraints, obtaining the optimal on / off state of the thermal power units and then verifying the transient frequency constraints. The second layer optimizes the optimal load shedding amount using the known unit combination and the ASO optimization algorithm, while also applying transient frequency constraints. The transient frequency constraints introduced in this patent make this unit combination more comprehensive than traditional unit combinations, and considering optimal load shedding makes power generation more economical.

[0099] The present invention will be further described in detail below with reference to embodiments:

[0100] The feasibility of this method was verified using a ten-unit system, which includes eight thermal power units, one wind farm, and one photovoltaic power station. The thermal power units have a capacity of 2800MW, while the wind farm and photovoltaic power station each have an installed capacity of 1000MW. The installed capacity of new energy sources accounts for 41.6% of the system capacity, classifying it as a high-proportion new energy system.

[0101] The unit combination selection has a 24-hour dispatch space. The fundamental frequency for the example is 50Hz, the minimum frequency requirement is 49.2Hz, and each load disturbance is set to 5% of the total load. The load curve for a typical day and the daily power generation of the renewable energy power station are shown below. Figure 3 .

[0102] The following three unit combination schemes were adopted and the results were analyzed:

[0103] Option a only considers the unit combination under conventional constraints.

[0104] Option B considers unit combination and transient frequency constraints under conventional constraints, where frequency regulation is not possible when renewable energy is operating at full load.

[0105] Scheme C considers unit combination and transient frequency constraints under conventional constraints, and frequency regulation by reducing load on new energy sources.

[0106] Figure 5(a) shows the lowest frequency of the power system frequency response under scheme a when encountering load disturbances at different times. The results indicate that this scheme only considers static constraints and cannot meet transient frequency constraints. Evaluating the power system under these circumstances is overly optimistic.

[0107] The optimal load reduction for scheme c, the new energy power plant / station, is shown in Table 1. The lowest frequency during the frequency response process for both schemes is shown in Figure 5(b) and Figure 5(c). Both schemes can meet the requirement of not exceeding the frequency limit.

[0108] Table 1 shows the optimal load reduction for new energy power plants / stations under scheme c.

[0109] time wind farm Photovoltaic power station time wind farm Photovoltaic power station 0:00 5.00% 0.00% 12:00 4.83% 4.66% 1:00 4.77% 0.00% 13:00 4.63% 4.79% 2:00 0.16% 0.00% 14:00 4.90% 4.99% 3:00 4.95% 0.00% 15:00 4.57% 4.91% 4:00 4.95% 0.00% 16:00 4.82% 4.80% 5:00 4.98% 0.00% 17:00 4.98% 4.81% 6:00 4.91% 4.96% 18:00 4.96% 0.81% 7:00 0.08% 0.39% 19:00 4.95% 0.00% 8:00 4.82% 4.80% 20:00 4.95% 0.00% 9:00 0.58% 0.26% 21:00 4.98% 0.00% 10:00 0.94% 4.96% 22:00 4.64% 0.00% 11:00 4.80% 4.85% 23:00 0.40% 0.00%

[0110] pass Figure 6 It can be seen that at multiple times within the dispatch interval, the power generation of new energy plants / stations under scheme c is greater than that under scheme b. That is, under the condition of new energy load reduction, the power generation of new energy is greater than that under the condition of relying solely on thermal power frequency regulation at some times. This indicates that scheme c satisfies the dynamic constraints of system frequency while also improving the system's new energy absorption capacity.

[0111] pass Figure 7At multiple times within the dispatch interval, the number of thermal power units started in scheme c is less than that in scheme b. The thermal power units in scheme b start and stop more frequently. It can be seen that after considering the frequency issue, more units will be started to provide frequency support for the system. Comparing schemes b and c, when renewable energy reduces load to provide frequency support, the frequency regulation pressure of thermal power units can be reduced, making the unit combination more reasonable.

[0112] The reference indicators for schemes b and c are shown in Table 2. The renewable energy curtailment rate (α) is the ratio of renewable energy curtailment to renewable energy generation, and the renewable energy penetration rate (β) is the percentage of renewable energy generation to the total system load.

[0113] (twenty three)

[0114] (twenty four)

[0115] In the formula: For the first i New energy power plants / stations t The amount of new energy power generated at any given moment. For the first i New energy power plants / stations t The proportion of new energy vehicles reducing load at any given time. For the first i New energy power plants / stations t The maximum power generation of new energy sources at any given time. This represents the system load at time t.

[0116] As shown in Table 2, Scheme C has a higher penetration rate and a lower curtailment rate than Scheme B. The participation of renewable energy in frequency regulation can reduce the frequency regulation pressure on thermal power plants, lower the curtailment rate of renewable energy, and reduce power generation costs. In this system, the curtailment rate of renewable energy is less than 5%, indicating that the installed capacity is relatively reasonable.

[0117] Table 2 Comparison of Acceptance Capacity of Schemes b and c

[0118] index Option b Option C Penetration 25.29% 27.94% curtailment rate 12.18% 2.39% Electricity generation cost <![CDATA[6.61×10 7 ]]> <![CDATA[6.56×10 7 ]]>

[0119] It is understood that the present invention has been described through some embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.

Claims

1. A unit combination method considering transient frequency constraints and optimal load reduction of new energy sources, characterized in that, The steps are as follows, and they are performed in sequence: Step 1: Predict the load output and maximum output of new energy sources on a certain day, set the maximum load reduction percentage of new energy sources, and initially set it to 0. Based on the first layer of optimization, obtain the operating status of each unit and the output of new energy units. Step 2: Based on the unit's operating status and the output of new energy sources, check whether the minimum frequency meets the transient frequency constraint. If it meets the condition, no second-level optimization is needed, and output the result. The expression for the transient frequency constraint in step two is as follows: (8) (9) (10) (11) in, The lowest frequency; The system reference frequency; The time to reach the lowest frequency point. This is the system's power deficit; This is the equivalent inertial response time constant; , Intermediate variables for ease of calculation; It is the natural vibration frequency; The damping ratio; This is the load damping coefficient; For the first k The inertial time constant of a thermal power unit; For thermal power units i time t The unit's on / off status; For the first k The capacity of the thermal power unit; The capacity of the entire system; The time constant of the speed controller; , and For parameters , and Substitute t Constantly monitor the on / off status of thermal power units And the values ​​after the load reduction of new energy power plants / stations; and The first i A photovoltaic power station, a wind farm at any time t The adjustment coefficient; For the first i Mechanical power gain factor of thermal power units in Taiwan; For the first i The fraction of the total power generated by the high-pressure steam turbine; The first i The droop coefficient of a thermal power unit; This indicates the on / off status of all thermal power units. This indicates the load reduction status of new energy power plants / stations; If formula (11) is satisfied, then the result is obtained; If formula (11) is not satisfied, then proceed to the second layer of optimization in step three; Step 3: If the frequency requirement is not met, perform the second-level optimization to obtain the optimal load reduction state of the unit and check whether the minimum point frequency meets the transient frequency constraint; if the condition is met, proceed to Step 1; if the requirement is not met, increase the maximum load reduction percentage. The second layer of optimization in step three is based on the unit start-up and shutdown status obtained from the first layer of optimization as the input quantity, with the minimum load reduction of the new energy plant / station as the objective function, and the transient frequency as the constraint, to find the optimal load reduction within a certain load reduction space. The objective function of the optimal load reduction process of new energy is Equation (12), the transient frequency constraint is Equation (11), and the load reduction space is Equation (13). (12) In the formula , This represents the percentage of load reduction for both photovoltaic power plants and wind farms at time t. , The power generation units of wind farms and photovoltaic power plants are respectively tracking the power generation of their operating states at maximum power. , This is the load reduction cost coefficient for wind farms and photovoltaic power plants; The load reduction optimization takes the minimum wind and solar load reduction, i.e., equation (12), as the objective function, and adds inequality constraints (11). Through the penalty factor M, a new second-level optimization objective function is formed, namely (14) and (15): (14) (15) In the formula Indicates the lowest allowed frequency of the system; Indicates the lowest frequency point; penalty factor M When frequency constraints are not met m The value of is guaranteed to be much greater than ; Indicate the cost of load reduction; set the maximum load reduction amount to meet the constraints: (13) In the formula , They are respectively t The percentage of load reduction for wind farms and solar power plants at any given time; , This refers to the maximum allowable load reduction for wind farms and photovoltaic power plants. The second-level optimization is solved by using the atomic search optimization algorithm, and the optimal load reduction is determined as the second-level optimization, so as to obtain the optimal load reduction percentage of new energy farms / stations within the finite load reduction range; The atomic search optimization algorithm, with its penalty function, solves the optimal load reduction problem by following these steps, which are performed sequentially: S1. Determine the variables and objective function. The variable is the load reduction ratio of the new energy unit, the load reduction space is Equation (13), the dimension of the intelligent algorithm is the number of new energy units, and the objective function is Equation (14). S2. Configure the parameters. In addition to the general parameters for all optimization algorithms, only two parameters need to be set for atomic search optimization: depth weight and Lagrange multiplier weight. S3. After determining the number of atoms N, assign an initial position and velocity to each atom; S4. Input the position information of the atoms into the fitness function, sort them from best to worst according to the value of the function, and take the first K atoms as the optimal subset; S5. Calculate the mass of each atom using the atom search optimization algorithm; S6. Calculate the acceleration of each atom in each dimension according to the algorithm; S7. After determining the acceleration of the atom, redetermine the velocity and position of the atom; S8. Continue iterating according to 4-7 until T times, then select the optimal solution; The second-level optimization yields the optimal load reduction. Optimal load reduction Substitute into equation (11) to see if the transient frequency constraint is met. If the transient frequency constraint is met, return directly to step one for the first layer of optimization. If the transient frequency constraint is not met, the load reduction space needs to be increased, and then return to step one for the first layer of optimization. Step 4: Repeat steps 1 through 3.

2. The unit combination method considering transient frequency constraints and optimal load reduction of new energy sources according to claim 1, characterized in that, The first layer of optimization in step one is a traditional unit combination optimization, with the objective function being the power generation cost of the thermal power unit: (1) In the formula For thermal power units i time t The generator set's on / off status is indicated by 1 for on and 0 for off. , , For thermal power units i The power generation cost coefficient, , For thermal power units i Start-up and shutdown costs For thermal power units i time t The amount of electricity generated.

3. The unit combination method considering transient frequency constraints and optimal load reduction of new energy sources according to claim 2, characterized in that, The constraints used in the first layer of optimization in step one include power balance constraints, output constraints of thermal power units and new energy power plants / stations, ramping constraints, start-up and shutdown time constraints, and reserve constraints. (2) (3) (4) (5) (6) (7) Equation (2) is the power balance constraint, Equations (3) and (4) are the output constraints of thermal power units and new energy power plants / stations, respectively, Equation (5) is the ramp constraint, Equation (6) is the start-stop time constraint, and Equation (7) is the standby constraint. in, , , This represents the output of wind turbines, photovoltaic power plants, and wind farms at time t. This represents the system load at time t. , This represents the lower and upper limits of the output of thermal power unit i. , This represents the uphill and downhill power of thermal power unit i. , This indicates the continuous start-up and shutdown time of a thermal power unit. Indicates the reserve factor. , This represents the percentage of load reduction for both photovoltaic power plants and wind farms at time t. , The power generation units of wind farms and photovoltaic power plants are respectively tracking the power generation of their operating states at maximum power. , These represent the actual power generation of wind farms and photovoltaic power plants, respectively, with m, n, and k indicating the number of wind power, photovoltaic, and thermal power plants.

4. The unit combination method considering transient frequency constraints and optimal load reduction of new energy sources according to claim 1, characterized in that, The first layer of optimization in step one adopts hybrid linear integer optimization, which is solved using the cplex solver built into MATLAB to obtain the on / off state u of the thermal power unit.

Images