A unit commitment optimization method considering load side inertia uncertainty

Abstract

A unit commitment optimization method considering load side inertia uncertainty, comprising: step 1: obtaining technical parameters, economic parameters, load prediction data and wind power prediction value of each synchronous generator unit in the power system, and dynamic frequency constraint of the power system RoCoF Protection setting value and frequency extreme value allowable range; step 2: deriving the dynamic frequency constraint of the power system according to the dynamic frequency stability range; step 3: considering the uncertainty of high proportion of wind power energy, obtaining the typical scene wind power output curve through scene generation and scene reduction; step 4: considering the uncertainty of load side inertia, converting the dynamic frequency constraint into a deterministic constraint through opportunity constraint, and building a unit commitment optimization model considering typical wind power scene and dynamic frequency constraint. The method comprehensively considers the influence of power supply side and load side inertia, and considers the uncertainty of load side inertia and wind power uncertainty, so that the unit commitment model is more frequency safe and economical.

Description

Technical Field

[0001] This invention relates to the field of power system optimization operation technology, and specifically to a unit combination optimization method that considers the uncertainty of load-side inertia. Background Technology

[0002] Currently, power systems dominated by conventional synchronous power sources are gradually evolving into those dominated by asynchronous power sources such as wind and solar. However, new energy sources like wind and solar power are generally connected to the grid via power electronic converters, which cannot actively provide inertia support under active power disturbances and thus cannot mitigate system frequency changes. Therefore, with the increasing penetration of new energy sources and the significant replacement of conventional power sources by asynchronous power sources, the inertia level of the power system is continuously decreasing, and the frequency stability problem is becoming increasingly prominent. Most studies exploring potential inertia resources to enhance the system's inertia level only consider the inertia on the power source side and do not investigate the impact of load-side inertia in the unit combination model.

[0003] However, research on load-side inertia is still in its early stages. Current studies assess the overall system inertia, suggesting it originates from synchronous generators, virtual inertia, and load inertia. Some researchers have found a positive impact of load-side asynchronous motor inertia on dynamic frequency, improving the system's frequency change rate and minimum frequency. Others have quantitatively assessed the effective inertia of asynchronous motors during the inertia response phase. International researchers have also studied frequency safety incidents in GB systems to determine the approximate proportion of the equivalent load-side inertia in the overall system inertia.

[0004] The above studies have demonstrated the positive effect of load-side inertia on frequency, but no studies have yet incorporated the equivalent load-side inertia into the frequency response system, considered its uncertainty, and included it in the unit combination model to analyze its related impact.

[0005] Against this backdrop, incorporating the equivalent inertia of the load side into the frequency response system, improving the overall system inertia system, and considering the uncertainty of load side inertia prediction are of great significance to the frequency security and stability of day-ahead scheduling. Summary of the Invention

[0006] This invention provides a unit combination optimization method that considers the uncertainty of load-side inertia. This method comprehensively considers the influence of power supply-side and load-side inertia, and takes into account the uncertainty of load-side inertia and wind power uncertainty, making the unit combination model more frequency-safe and economical.

[0007] The technical solution adopted in this invention is as follows:

[0008] A unit combination optimization method considering load-side inertia uncertainty includes the following steps:

[0009] Step 1: Obtain the technical parameters, economic parameters, load forecast data, and wind farm power forecast values ​​for each synchronous generator unit in the power system, as well as the RoCoF protection setting values ​​and frequency extreme value allowable values ​​in the power system;

[0010] Step 2: Based on the influence of load-side inertia during the frequency response process, derive the dynamic frequency constraints of the power system according to the dynamic frequency stability range.

[0011] Step 3: Taking into account the uncertainty of high proportion of wind power energy, wind power output curves for multiple typical scenarios are obtained through scenario generation and scenario reduction;

[0012] Step 4: Taking into account the uncertainty of load-side inertia, the dynamic frequency constraint is transformed into a deterministic constraint through chance constraints, and a unit combination optimization model considering typical wind power scenarios and dynamic frequency constraints is built.

[0013] In step 1, the technical parameters mainly include the inertial time constant, unit regulating power, maximum output, minimum output, uphill gradient, downhill gradient, minimum continuous operating time, minimum continuous downtime, and initial state of the synchronous generator set. The relevant parameters are shown in Table 1 below:

[0014] Table 1 Technical parameters of synchronous generator sets

[0015]

[0016] The economic parameters mainly include the operating cost coefficient, hot start cost, cold start cost, and cold start time of the synchronous generator set. Specific parameters are shown in Table 2 below.

[0017] Table 2 Economic parameters of synchronous generator sets

[0018]

[0019]

[0020] The load forecast data consists of predicted load power for each time period within a 24-hour period, while the wind farm power forecast data consists of predicted wind power for each time period within a 24-hour period. The predicted values ​​for both are as follows: Figure 2 As shown.

[0021] The RoCoF protection setting value is |RoCoF set,max That is, the maximum allowable rate of frequency change of the system is 0.5 Hz / s; the allowable extreme value of the frequency is f. set,min That is, the lowest frequency allowed by the system is 49Hz.

[0022] In step 2, the influence of load-side inertia is considered. Specifically, this avoids the conservatism of only considering the power source-side inertia in the original power system. The frequency response equation comprehensively considers both the power source-side inertia and the load-side inertia.

[0023] In step 2, the dynamic frequency stability range mainly refers to the fact that the maximum rate of frequency change obtained by the system should be below the maximum allowable rate of frequency change of 0.5 Hz / s; and the minimum frequency obtained by the system should be above the minimum allowable frequency of 49 Hz.

[0024] In step 2, the dynamic frequency constraint of the system is derived as follows:

[0025]

[0026]

[0027] Among them, |RoCoF max |To obtain the maximum rate of change of frequency;|RoCoF set,max | represents the maximum permissible rate of frequency change for the system, in Hz / s; ΔP L E represents the active power imbalance caused by the disturbance, measured in MW. SYS f represents the total inertia of the system, expressed in MW·s. min f is the lowest point of the system frequency. n f is the system's rated frequency, f0 is the system's original operating frequency, f set,min R is the lowest permissible frequency of the system, expressed in Hz. SSi U represents the primary frequency regulation response rate of each synchronous generator, in MW / s; i represents the synchronous generator set; N represents the total number of generator sets i; U it This represents the start-up and shutdown status of a conventional unit i during time period t.

[0028] In step 3, the wind power uncertainty model is as follows:

[0029]

[0030]

[0031] in, Indicates the power output of wind power; ΔP represents the predicted wind power output. Wt This indicates the error in wind power prediction.

[0032] In step 3

[0033] Scene generation is specifically based on Figure 2 The wind power prediction data and the known wind power prediction error distribution patterns were used to generate M scenarios through Monte Carlo sampling.

[0034] Scene reduction specifically involves sampling and generating M scenes, and then using a scene distance reduction method to obtain several typical wind power output scene curves.

[0035] The invention specifically reduces the wind power output curves to five typical wind power output scenarios, such as... Figure 3 As shown.

[0036] In step 4, the established load-side inertia uncertainty model is as follows:

[0037]

[0038] ΔE L ~N(0,σ) (6);

[0039] Among them, E L For load-side inertia, To set the predicted value of load-side inertia, ΔE L The prediction error difference is expressed in MW·s; N(0,σ) represents the Gaussian distribution model; and σ represents the standard deviation of the Gaussian distribution model.

[0040] In step 4, the opportunity constraint model is as follows:

[0041]

[0042] Where, minf(x) is the minimum value of the system objective function; Pr[g j [(x,ξ)≤0] represents the probability that the constraint condition holds; g j (x,ξ) are constraint functions; x is the optimization variable; ξ is a random variable; α j Let α be the threshold at which the decision result causes the j-th constraint to fail, where 0 ≤ α j ≤1; j represents the number of constraints; p represents the total number of constraints.

[0043] Dynamic frequency opportunity constraints:

[0044]

[0045]

[0046] in, E represents the inertia of a synchronous turbine in a wind power scenario s, expressed in MW·s. L The load-side inertia is expressed in MW·s; |RoCoF set,max | represents the maximum permissible rate of frequency change for the system, in Hz / s; f n f is the system's rated frequency. set,minR is the lowest permissible frequency of the system, in Hz; α is a given confidence level for meeting the maximum rate of change requirement; SSi The primary frequency regulation response rate for each synchronous generator is expressed in MW / s. β represents the start-up and shutdown status of conventional unit i in wind power scenario s during time period t; β is a given confidence level that satisfies the minimum frequency requirement.

[0047] Dynamic frequency deterministic constraints:

[0048]

[0049]

[0050] in, This is the predicted load-side inertia value, in MW·s; Φ -1 (1-α) is the inverse function of the 1-α probability distribution function; f n This is the system's rated frequency.

[0051] The unit combination model takes minimizing system scheduling cost and wind curtailment cost as its objective function, and is expressed as follows:

[0052]

[0053] Among them, F s Let be the minimum expected total cost of the turbine combination in wind power scenario s; In wind power scenario s, the start-up and shutdown status of conventional unit i during time period t-1 is given; t represents each time period within the cycle; N represents the number of conventional units; T represents the number of hours within the study period; E s This represents the expected total cost of the wind turbine in wind power scenario s. The active power output of conventional unit i in time period t under wind scenario s;

[0054] The operating cost of conventional turbine i in wind power scenario i. Where: a i b i c i These are the coefficients of the cost function;

[0055] This refers to the start-up and shutdown status of a conventional wind turbine unit i during time period t in a wind power scenario s. Indicates that the program is running. Indicates that the machine is stopped; is the startup cost of unit i in time period t under wind power scenario s; w is the penalty parameter for wind curtailment; The wind power output power under wind power scenario s; This represents the actual power output value under wind power scenario s.

[0056] This invention provides a unit combination optimization method considering the uncertainty of load-side inertia, with the following technical advantages:

[0057] 1) The frequency opportunity constraint model proposed in this invention, which considers the uncertainty of the inertia on the rate load side, makes the frequency constraint function characteristics in the unit combination model more complete and shows better results in terms of frequency stability and overall unit cost.

[0058] 2) This invention takes into account the uncertainties of new energy wind power, making the model more adaptable to current power systems with a high proportion of wind power;

[0059] 3) The model proposed in this invention demonstrates its superiority when the proportion of wind farm capacity increases, significantly improving economic efficiency while ensuring frequency stability. In the context of a high proportion of new energy sources, this is beneficial to the stable and economical operation of the power system. Attached Figure Description

[0060] Figure 1 This is a flowchart of the solution process for the unit combination optimization model of the present invention.

[0061] Figure 2 This is a graph showing wind power and load forecasts.

[0062] Figure 3 A diagram illustrating the power generation capabilities of five typical wind power scenarios.

[0063] Figure 4 This is a comparison chart of the lowest frequency points for each scheme in scenario 4 at each time period.

[0064] Figure 5 This is a comparison chart of the maximum frequency change rate for each scheme in Scenario 4 over a given time period.

[0065] Figure 6 This is a cost comparison chart of the turbine units under five typical wind power scenarios in Scheme 3.

[0066] Figure 7 This is a comparison chart of the lowest frequency points under different confidence levels in scenario 4.

[0067] Figure 8 This is a comparison chart of frequency change rates under different confidence levels for scenario 4.

[0068] Figure 9 A comparison chart of predicted wind power output for different wind power capacity percentages.

[0069] Figure 10 The diagram shows five typical wind power scenarios with different wind power capacity ratios.

[0070] Figure 11 This is a comparison chart showing the lowest frequency points for each scheme with a wind power capacity share of 10%.

[0071] Figure 12This is a comparison chart of the maximum frequency variation rate for various schemes with a wind power capacity share of 10%.

[0072] Figure 13 This is a comparison chart showing the lowest frequency points for each scheme with a 50% wind power capacity ratio.

[0073] Figure 14 This is a comparison chart of the maximum frequency variation rate for each scheme with a wind power capacity share of 50%. Detailed Implementation

[0074] A unit combination optimization method based on load-side inertia uncertainty includes: Step 1, acquiring power system generator parameters, load and wind power forecast data, and the allowable range of frequency indicators; Step 2, considering the influence of load-side inertia during the frequency response process, deriving the system's dynamic frequency constraints based on the dynamic frequency stability range; Step 3, considering the uncertainty of high-proportion wind power energy in the unit combination model, obtaining wind power output curves for multiple typical scenarios through scenario generation and scenario reduction; Step 4, considering the uncertainty of load-side inertia in the unit combination model, transforming the dynamic frequency constraints into deterministic constraints through chance constraints, and constructing a unit combination optimization model considering typical wind power scenarios and dynamic frequency constraints. This invention enables the unit combination model to possess both economic efficiency and frequency security, and its significant role is even more evident in systems with a high proportion of new energy sources.

[0075] Preferred solutions include Figures 1 to 14 As shown:

[0076] A unit combination optimization method considering inertia uncertainty and dynamic frequency constraints is proposed. The specific unit combination model is as follows: The objective function of the established model is:

[0077]

[0078] In equation (12): F s Let be the minimum expected total cost of the turbine combination in wind power scenario s; In wind power scenario s, the start-up and shutdown status of conventional unit i during time period t-1 is given; t represents each time period within the cycle; N represents the number of conventional units; T represents the number of hours within the study period; E s This represents the expected total cost of the wind turbine in wind power scenario s. The active power output of conventional unit i in time period t under wind scenario s; The operating cost of conventional turbine i in wind power scenario i; a i b i c i These are the coefficients of the cost function; This refers to the start-up and shutdown status of a conventional wind turbine unit i during time period t in a wind power scenario s. Indicates that the program is running. Indicates that the machine is stopped; Let be the startup cost of unit i during time period t in wind power scenario s; w is the penalty parameter for wind curtailment. The wind power output power in wind power scenario s. This represents the actual power output value under wind power scenario s.

[0079] The constraints include:

[0080] 1) Dynamic frequency constraint:

[0081]

[0082]

[0083] 2) Wind power output constraints:

[0084]

[0085] 3) System power balance constraints:

[0086]

[0087] In the formula: P Lt Let t be the load demand during time period t.

[0088] 4) Unit output power constraints:

[0089]

[0090] In the formula: P i max P i min These are the upper and lower limits of the output of the i-th generator, respectively.

[0091] 5) Unit ramp-up rate constraint:

[0092]

[0093] In the formula: L i D i These represent the upper and lower limits of the ramp-up for unit i, respectively.

[0094] 6) Minimum start-up and shutdown time constraints:

[0095]

[0096] In the formula: Let T be the continuous start-up time and shutdown time of the i-th generator during the first t-1 time periods, respectively; i on T i down...

[0097] 7) System spin-off standby constraint:

[0098] In the model, ηP Lt δP Lt As the prediction error for the power of the spinning reserve balancing load, η and δ are taken as 5%.

[0099]

[0100]

[0101] Example:

[0102] The implementation example provided by this invention is an improved ten-unit system, comprising ten synchronous generator sets with a total installed capacity of 1662MW. Relevant parameters are shown in Tables 1 and 2 above. Based on testing, a wind farm with an installed capacity of 900MW is connected, accounting for 35% of the total installed capacity of the system. Considering the day-ahead dispatch timescale, the planned dispatch cycle is 24 hours. The 24-hour wind power and load forecast curves are shown below. Figure 2 As shown, the RoCoF protection setting value of the example system is RoCoF. set The system's rated frequency f is 0.5 Hz / s. n The frequency limit is 50Hz, f. set The frequency is 49Hz, and the power of the disturbance is 15%. Based on Monte Carlo sampling and scene reduction, five typical scenarios are derived as follows: Figure 3 As shown, the probabilities for each typical wind power scenario are listed in Table 3. The standard deviation of the load-side inertia prediction error is taken as...

[0103] Table 3. Probabilities of 5 Typical Wind Power Scenarios

[0104]

[0105] (1) Scheduling results of different schemes:

[0106] To demonstrate the effectiveness and superiority of the model proposed in this invention, three schemes were set up for comparison (all considering the uncertainties of wind power). Scheme 1: Traditional unit combination model, without considering frequency-related constraints; Scheme 2: Unit combination model that only considers the dynamic frequency constraints of the power source side inertia; Scheme 3: The method proposed in this invention—a unit combination model that considers the frequency constraints of the source-load side inertia response.

[0107] Table 4 shows the expected costs of the unit combinations for the three schemes mentioned above. Scheme 2 has the highest costs across all categories, followed by Scheme 3, with Scheme 1 having the lowest. Scheme 3, compared to Scheme 2, reduces the unit combination operating cost by 4.9%, the wind curtailment cost by 64.4%, and the expected total cost by 6.4%. Selecting a typical scenario with the highest probability of wind power output (Scenario 4), we compare the specific system frequency changes of the three schemes. Figure 4 and Figure 5 For each scheme in wind power output scenario 4, the system frequency minimum point and maximum frequency change are analyzed. Scheme 1's frequency minimum point falls below the 49Hz frequency limit during periods 1-3 and 23-24, and the maximum frequency change rate exceeds the RoCoF protection setting value of 0.5Hz / s for multiple periods, failing to meet frequency stability requirements. Schemes 2 and 3 have both frequency minimum points and maximum frequency change rates within safe thresholds for each period. Therefore, while Scheme 1 offers good economic efficiency, it lacks frequency stability. Scheme 3, while meeting frequency safety requirements, is also more economical than Scheme 2 and significantly reduces wind curtailment costs. In summary, the scheme of this invention promotes wind power consumption and possesses good frequency stability and economic efficiency.

[0108] Table 4 Comparison of Expected Costs of Unit Combinations under Different Schemes

[0109]

[0110] (2) Scheduling results at different confidence levels

[0111] Table 5 presents the optimized cost results of the system under Scheme 3 of this invention at different confidence levels. As shown in Table 5, with the continuous increase in confidence level, the total expected cost of power generation obtained after weighted summation of the decision schemes in the wind power scenario increases accordingly. The trend chart of the total unit combination cost of Scheme 3 under five typical wind power scenarios is shown below. Figure 6 As shown, it is clear that under the same wind power output scenario, the higher the confidence level, the higher the combined cost of the turbines, and this pattern holds true for each scenario. However, the degree of influence of confidence level on the combined cost of the turbines varies across different wind power scenarios. In wind power scenario 1, the combined cost of the turbines changes significantly under different confidence levels, while in wind power scenario 5, the combined cost of the turbines does not change significantly under different confidence levels. Therefore, the degree of influence of confidence level on the combined cost of the turbines is affected by the wind power output scenario.

[0112] Selecting a typical wind power output scenario 4, we analyze the changes in system frequency under different confidence levels in Scheme 3. Figure 7 and Figure 8The table shows the minimum frequency and frequency change rate for each scheme. It can be seen that in most periods, such as 1-2 and 4-20, the minimum frequency decreases to some extent and the maximum frequency change rate increases with increasing confidence. Combined with the unit start-up and shutdown results in Table 6, it can be seen that the unit start-up and shutdown scheme remains unchanged during the current period. However, in a few periods, such as 3 and 21-23, as the confidence increases, the system adds units to maintain frequency safety, changing the unit start-up and shutdown scheme, resulting in abrupt changes in both frequency evaluation indicators. Therefore, increasing the confidence level will affect frequency indicator changes during periods where the system's inertial response capability is insufficient, altering the unit start-up and shutdown scheme, thus causing the changes in unit combination costs mentioned earlier, increasing costs while ensuring frequency safety. In actual operation, different confidence levels can be selected for UC decisions based on the system's trade-off between frequency safety risks and costs.

[0113] Table 5. Unit Combination Costs under Scheme 3 at Different Confidence Levels

[0114]

[0115] Table 6. Start-up and shutdown results of different confidence level unit combinations in this invention.

[0116]

[0117] (3) Dispatch results for different wind power capacity proportions:

[0118] This paper analyzes the impact of different wind power capacity proportions on the cost and frequency security of three schemes. By changing the capacity of the connected wind farms, the simulation results are compared and analyzed for different wind farm capacities and proportions, with corresponding predicted wind power values ​​as follows: Figure 9 As shown, five typical wind power scenarios with different wind power capacity ratios are as follows: Figure 10 As shown.

[0119] Table 8 shows the unit combination costs under different wind power capacity ratios. As can be seen from Table 8, when the wind power capacity ratio is 10%, the curtailment cost for each scheme is 0, and the total operating cost of the three schemes is the same. When the wind power capacity ratio is 35%, which is the original wind power capacity ratio of this invention, Scheme 3 reduces the unit combination operating cost by 4.9% and the curtailment cost by 64.4% compared to Scheme 2, with an expected total cost reduction of 6.4%. When the wind power capacity ratio is 50%, the cost differences among the three schemes are significant. Compared to Scheme 2, Scheme 3 reduces its operating cost by 23.4%, its curtailment cost by 41.6%, and its total cost by 37%. This demonstrates that in power systems with a small proportion of wind power capacity, where all wind power is absorbed and synchronous generators account for a high percentage, the system has sufficient inertia response capability. Considering frequency constraints on the power source or load side inertia has little impact on cost changes. However, in power systems with a large proportion of wind power capacity, where the system's inertia response capability is insufficient, the cost changes of each scheme's units become significant after considering frequency constraints. When the proportion of wind power increases from 35% to 50%, the total cost of Scheme 3 compared to Scheme 2 increases from a 6.4% reduction to a 34% reduction. This indicates that increasing load-side inertia can play a better role in systems with a high proportion of renewable energy.

[0120] Figure 11 and Figure 12 The table shows the minimum frequency and frequency change rate for each scheme with a wind power capacity share of 10%. It can be seen that Scheme 1 and Scheme 2 have the same minimum frequency and frequency change rate. Scheme 3 has a slightly higher minimum frequency and a slightly lower frequency change rate compared to the first two schemes. All three schemes have frequency indicators within limits and meet frequency safety requirements. The table shows the minimum frequency and frequency change rate for each scheme with a wind power capacity share of 50%. Figure 13 and Figure 14 As shown, both frequency indicators of Scheme 1 are clearly outside the safe range, while the frequency indicators of Scheme 2 and Scheme 3 meet the requirements. This indicates that frequency safety needs to be considered more in new energy wind power systems to ensure the stable operation of the system.

[0121] Table 7. Probabilities of Typical Scenarios with Different Wind Power Capacity Ratios

[0122]

[0123]

[0124] Table 8 Unit Combination Costs under Different Wind Power Capacity Ratios

[0125]

[0126] (4) Dispatch results for different wind power capacity proportions:

[0127] The error model for the uncertainty factor in the system inertia is H(0,σ), where the value of σ mentioned earlier is 0.2H. L In a normal distribution, the smaller the value of σ, the narrower the curve, the more concentrated the distribution of inertia error values, and the closer it is to the actual value under the same probability. When the value is larger, the normal distribution curve is flatter and wider, the distribution of inertia error values ​​is more dispersed, and the further it is from the actual value under the same probability.

[0128] When standard deviation Then, the impact of prediction errors on costs was analyzed. Table 9 shows that, at the same confidence level, the smaller the value of σ, the lower the total cost. This indicates that the smaller the value of σ, the closer the confidence band of the load-side inertia prediction model is to the true value, and the lower the unit combination cost. This illustrates the importance of load-side inertia prediction and the importance of the accuracy of establishing the load-side inertia uncertainty model.

[0129] Table 9. Unit Costs Based on Standard Deviation of Inertia Error on Different Load Sides in this Invention

[0130]

[0131] This invention addresses the uncertainties in renewable energy output and load-side inertia in unit combination. It derives a frequency constraint equation incorporating load-side inertia and establishes a stochastic unit combination model considering the dual uncertainties of wind power output and load-side inertia. This model improves the characteristics of the frequency constraint function in the unit combination model and exhibits superior results in terms of frequency stability and overall unit cost. The proposed model demonstrates its superiority as the proportion of wind farm capacity increases, contributing to the stable and economical operation of the power system in the context of a high proportion of renewable energy.

Claims

1. A unit combination optimization method considering load-side inertia uncertainty, characterized in that... Includes the following steps: Step 1: Obtain the technical parameters, economic parameters, load forecast data, and wind farm power forecast values ​​for each synchronous generator unit in the power system, as well as the power system's... RoCoF Protection settings and frequency extreme value allowable values; Step 2: Consider the influence of load-side inertia during the frequency response process, and derive the dynamic frequency constraints of the power system based on the dynamic frequency stability range; Step 3: Taking into account the uncertainty of high proportion of wind power energy, wind power output curves for multiple typical scenarios are obtained through scenario generation and scenario reduction; Step 4: Taking into account the uncertainty of load-side inertia, the dynamic frequency constraint is transformed into a deterministic constraint through chance constraints, and a unit combination optimization model considering typical wind power scenarios and dynamic frequency constraints is built. In step 4, the established load-side inertia uncertainty model is as follows: （5）； （6）； in, For load-side inertia, To set the predicted value of load-side inertia, The difference in prediction error is expressed in MW. s; It is a Gaussian distribution model; The standard deviation of the Gaussian distribution model; In step 4, the opportunity constraint model is as follows: （7）； in, This represents the minimum value of the system's objective function; The probability that the constraint condition is true; For constraint functions; To optimize variables; It is a random variable; The decision result caused the first The threshold where the constraint condition is not met, where ; The table shows the number of constraints. This represents the total number of constraints. Dynamic frequency opportunity constraints: （8）； （9）； in, For wind power scenarios Inertia of the lower synchronous generator unit, in MW s; Load-side inertia, in MW s; This represents the maximum permissible rate of frequency change for the system, expressed in Hz / s. The system's rated frequency, The lowest allowed frequency by the system, in Hz; For a given confidence level that satisfies the maximum rate of change requirement; The primary frequency regulation response rate for each synchronous generator is expressed in MW / s. For wind power scenarios conventional units The start / stop status during time period t; For a given confidence level that satisfies the minimum frequency requirement; The active power imbalance caused by the disturbance; Dynamic frequency deterministic constraints: （10）； （11）； in, This is the predicted load-side inertia value, in MW. s; for The inverse function of the probability distribution function; This is the system's rated frequency.

2. The unit combination optimization method considering load-side inertia uncertainty according to claim 1, characterized in that: In step 2, the dynamic frequency constraint of the system is derived as follows: （1）； （2）； in, To obtain the maximum rate of change of frequency; This represents the maximum permissible rate of frequency change for the system, expressed in Hz / s. The active power imbalance caused by the disturbance, expressed in MW; The total inertia of the system, in MW. s; This is the lowest point of the system frequency. The system's rated frequency, This is the system's original operating frequency. The lowest allowed frequency by the system, in Hz; The primary frequency regulation response rate for each synchronous generator is expressed in MW / s. It is a synchronous generator set; For generator sets The total number; For conventional units The start / stop status during time period t.

3. The unit combination optimization method considering load-side inertia uncertainty according to claim 1, characterized in that: In step 3, the wind power uncertainty model is as follows: （3）； （4）； in, Indicates the power output of wind power; This represents the predicted wind power output. This indicates the error in wind power prediction.

4. The unit combination optimization method considering load-side inertia uncertainty according to claim 1, characterized in that: The unit combination model takes minimizing system scheduling cost and wind curtailment cost as its objective function, and is expressed as follows: （12）； in, For wind power scenarios Minimum expected total cost of current unit combination; For wind power scenarios conventional units During the period The start / stop status; For each time period within the cycle; This refers to the number of conventional generating units; The number of hours within the research period; Representative wind power scenario Current expected total cost of generating units; For wind scene conventional units The output active power during time period t; For wind power scenarios conventional units Operating costs ,in: , , These are the coefficients of the cost function; For wind power scenarios conventional units During the start / stop state of time period t Indicates that the program is running. Indicates that the machine is stopped; For wind power scenarios Lower unit Startup cost during time period t; This is the penalty parameter for wind curtailment; For wind power scenarios The wind power output power below; For wind power scenarios The actual output value.

Images