Wind-fire-storage optimal economic dispatching method based on giant trevally optimizer and taking into account electricity price demand response

Abstract

The present invention relates to the technical field of economic dispatching of power systems, and specifically relates to a wind-fire-storage optimal economic dispatching method based on a giant trevally optimizer (GTO) and taking into account an electricity price demand response. The method comprises: establishing a wind-fire-storage joint economic dispatching model, and taking into consideration actual constraint conditions; taking into consideration the lifespan factor of a storage battery, and using a weighted throughput method to predict the lifespan of the storage battery; establishing an electricity price response model which is based on consumer psychology principles, and integrating a price-based demand response into economic dispatching; and using a GTO to solve the problem of wind-fire-storage optimal economic dispatching which takes into account an electricity price demand response. In the design of the present invention, an advanced GTO is innovatively combined with multi-objective optimal economic dispatching of power grids, so that a fast convergence speed and high accuracy are achieved, showing good performance with regard to handling the problem of wind-fire-storage optimal economic dispatching; in addition, peak-valley time-of-use electricity prices are used to guide changes in electricity usage structures and electricity usage patterns of users, and a weighted throughput method is used to predict the lifespans of storage batteries.

Description

An optimal economic dispatch method for wind-thermal-storage based on the giant scad optimizer and considering electricity price demand response Technical Field

[0001] The present invention relates to the technical field of economic dispatch of power systems, and in particular to a wind-thermal-storage optimal economic dispatch method based on a giant scad optimizer and taking into account electricity price demand response. Background Art

[0002] Power system economic dispatch is a method of dispatching that maximizes resource and equipment utilization and ensures reliable power supply to users at the lowest cost. This precondition requires operational safety and high-quality electricity. Power system economic operation is crucial in grid operation. It is characterized by multiple constraints, nonlinearity, nonconvexity, and multidimensional hybridity. This is particularly complex with the large-scale grid integration of renewable energy generators, such as wind turbines.

[0003] Reasonable economic dispatch of power systems can bring higher economic benefits, but current economic dispatch is mostly designed for thermal or hydropower generation systems, and most of its methods are not suitable for more complex situations such as large-scale grid access of renewable energy generators such as wind power. The randomness and intermittent nature of wind power place higher demands on economic dispatch schemes for large-scale grid-connected operations. During the economic dispatch process of power systems, the throughput method can be used to predict the life of batteries. The traditional throughput method is simple and easy to implement, meets many calculation requirements, and has experimental basis. It is a good method for models that only require rough calculations. However, in actual operation, many factors will affect the life of batteries, resulting in reduced prediction accuracy. During the economic dispatch process of power systems, it is also necessary to establish a demand response model. When establishing this model, there are two main modeling methods: one based on elasticity matrix and the other based on statistical principles. Modeling methods based on the demand elasticity matrix require analyzing a large amount of market operating data to determine the self-elasticity coefficient and the mutual elasticity coefficient. These two elasticity coefficients are significantly correlated with load fluctuations, while demand load is significantly affected by seasons, weather, and holidays. Therefore, the model stability is poor, and the demand response parameters need to be regularly revised. The key to demand response modeling methods based on statistical principles is to determine the fitting curve model. The rationality of the curve model directly affects the accuracy of the user response calculation results. The selection of the curve model requires continuous trial and error, which is computationally intensive and may not necessarily result in a reasonable curve model. In addition, the Giant Trawler Optimizer (GTO) uses an advanced metaheuristic algorithm, but to date, this algorithm has not been applied to the field of power system economic dispatch.

[0004] In view of this, we propose a wind-thermal-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account the electricity price demand response, in order to provide a reference for the optimal dispatch and economic combination of renewable energy power generation systems.

[0005] Summary of the Invention

[0006] The purpose of the present invention is to provide a wind-thermal-storage optimal economic dispatch method based on the giant scad optimizer and taking into account the electricity price demand response, so as to solve the problems raised in the above background technology.

[0007] To solve the above technical problems, one of the objectives of the present invention is to provide a wind-thermal-storage optimal economic dispatch method based on the Giant Trevally optimizer and taking into account electricity price demand response, comprising the following steps:

[0008] S1. Starting from the goal of minimizing the total power generation cost, a wind-thermal-storage combined economic dispatch model is established, taking into account various constraints in actual operation;

[0009] S2. Consider the battery life factor in the economic dispatch of the microgrid and apply the weighted throughput method to predict the battery life;

[0010] S3. Establish an electricity price response model based on consumer psychology principles and integrate price-based demand response into economic scheduling;

[0011] S4. Extend the Giant Trevally Optimizer (GTO) to the field of multi-objective optimization, and use GTO to solve the wind-thermal-storage optimal economic dispatch problem taking into account electricity price demand response, obtain the optimal solution, and thus achieve optimal economic dispatch.

[0012] As a further improvement of the present technical solution, in step S1, the specific method for establishing the wind-thermal-storage combined economic dispatch model includes the following steps:

[0013] S1.1. Wind power modeling based on Weibull distribution: The probability distribution parameter of wind speed is the most representative indicator of the statistical characteristics of local wind resources and is also an important parameter in wind power research. There are many models that can be used to approximate the probability distribution of wind speed, including Weibull distribution, Rayleigh distribution, lognormal distribution, and normal distribution. Since the two-parameter Weibull model can adjust parameters to adapt to the periodic changes in wind speed, it is widely used in wind speed probability modeling. The probability density function that obeys the Weibull distribution is defined as follows:

[0014] In formula (1), the subscript t represents the time interval per hour, v is the wind speed, and v t is the wind speed at time t, k t is the shape parameter of the wind speed probability distribution function at time t, which determines the overall shape of the wind speed probability distribution density curve, c tis the scale parameter at time t, reflecting the average wind speed of the wind field, which plays the role of amplifying or shrinking the curve, but does not affect the shape of the distribution density curve; the corresponding cumulative distribution function is:

[0015] Assuming that the output characteristic of the wind turbine is a simplified piecewise linear function, its output power for a given wind speed v is defined as follows:

[0016] In formula (3), P Wt is the output power of a single wind turbine; is the rated power of the fan; v in is the cut-in wind speed of the wind turbine; v r is the rated wind speed of the wind turbine; v out is the cut-out wind speed of the wind turbine;

[0017] When the output power of the wind turbine is within a continuously changing range, its Weibull probability density function is as follows:

[0018] In formula (4),

[0019] S1.2. Battery Model: The prediction of battery state of charge (SOC) is an important part of power generation planning and economic balance. The formula for calculating SOC using the ampere-hour method is:

[0020] In formula (5), SOC0 is the initial state of charge; C A is the rated capacity of the battery; η is the charge and discharge coefficient of the battery; I is the charge and discharge current of the battery; t0 is the initial time of calculation; t is the end time of calculation;

[0021] In economic dispatch, the change of SOC is determined by the charge and discharge current; let the rated voltage of the system be U n , the load power requirement is P LD , the system remaining power is P e , then:

[0022] Discharge current I d :

[0023] Charging current I c :

[0024] Among them, U n is the system rated voltage, P LD is the load power demand, P e is the remaining power of the system;

[0025] S1.3. Wind-heat-storage joint economic dispatch model: This model aims to minimize the economic cost of the system. Its objective function is as follows: Min C total =C Wt,l (P l )+C DG,i (P i )+C OM,k (P k )+C DP,k (P k ) (8)

[0026] In formula (8), C Wt,l (Pl) is the cost of wind power generation, l=1,2,..,N l ; C DG,i (P i ) is the cost of thermal power generation, i=1,2,...,N i ; C OM,k (P k ) and C DP,k (P k ) are the maintenance cost and depreciation cost of the kth battery pack, k=1,2,...,N k ;

[0027] Typically, wind farms are privately owned. Grid operators are required to purchase electricity from wind power producers under power purchase agreements and pay a fee. The fee consists of three parts. The first part is the direct cost of purchasing wind power: d , Wt (P j )=d Wt,l P Wt,l ,l=1,2,…,N l (9)

[0028] In formula (9), d Wt,l is the direct cost coefficient of wind power;

[0029] The second part is the penalty cost caused by underestimating wind power output. Due to the large fluctuation of wind power generation, wind power forecasts usually have some errors. When the planned wind power generation is lower than the available wind power generation, the excess wind energy will be wasted:

[0030] In formula (10), K ue,Wt,l is the underestimated coefficient of wind power; P Wt,rate,l is the rated power of the first wind turbine; P Wt,l is the planned output power of the first wind turbine;

[0031] The third component is the cost of ancillary services due to overestimation of wind power output; if the planned wind power generation is higher than the actual available wind power generation, then the load demand should be met by obtaining backup power from elsewhere:

[0032] In formula (11), K oe,Wt,l is the overestimation factor of wind power;

[0033] The total cost of wind power generation is: C Wt,l (P l )=C d,Wt (P j )+C ue,Wt,l (P l )+C oe,Wt,l (P l ),l=1,2,…,N l (12)

[0034] The total cost of a thermal power plant mainly consists of the economic cost of coal consumption and the operating cost of environmental protection equipment. The economic cost of coal consumption of a single thermal power plant depends on its power generation capacity and is generally approximated by quadratic function fitting. The formula for calculating the economic cost of coal consumption of all thermal power plants is:

[0035] In formula (13), P T,i (t) is the power generation capacity of thermal power plant i during period t; T is the number of dispatch periods; N T is the number of thermal power plants; a i 、b i 、c i is the operating cost coefficient of thermal power plant i;

[0036] The main pollutants emitted by thermal power plants are NO x and SO2; to reduce pollutant emissions, coal-fired boilers are retrofitted with ultra-low emissions and denitrification and desulfurization equipment is installed; the environmental cost per unit of power generated by the equipment is used to describe the operating costs of denitrification and desulfurization. To simplify the model, the denitrification and desulfurization cost database is referenced, and M0 is selected as the desulfurization cost per unit of power generation, and M1 is the denitrification cost per unit of power generation. The formula for calculating the operating cost of environmental protection equipment is:

[0037] The total cost of thermal power generation is: C DG,i (P i )=C CS,DG,i (P i )+C EC,DG,i (P i ) (15)

[0038] The maintenance cost and depreciation cost function of the energy storage device is: C OM,k (P k)=K OM,k |P k |Δt (16) C DP,k (P k )=(E rated,k C E,k +P rated,k C P,k )L loss,k (17)

[0039] In formula (16) and formula (17), K OM,k E is the unit power operation and maintenance cost coefficient of the kth battery group; rated,k 、P rated,k are the rated capacity and rated power of the kth battery pack respectively; C E,k 、C P,k are the present values ​​of the unit capacity and unit power installation costs of the kth battery pack respectively; L loss,k is the life loss coefficient of the kth battery pack;

[0040] S1.4 Constraints, including:

[0041] The power balance constraint is:

[0042] In formula (18), P d P is the power demanded by the electric load; loss For network loss;

[0043] The operating constraints of wind power generation are: 0≤P Wt,l ≤P Wt,rate,l ,l=1,2,…,N l (19)

[0044] The maximum and minimum output constraints of thermal power plants are: P T,i,min ≤P T,i (t)≤P T,i,max (20)

[0045] Where, P T,i,max and P T,i,min are the upper and lower limits of the output of thermal power plant i, respectively;

[0046] The ramp constraint of the thermal power plant is: |P T,i (t)-P T,i (t-1)|≤P T,i speed (twenty one)

[0047] Where, P T,i speed is the upward or downward ramp rate limit of thermal power plant i;

[0048] The battery's cyclic charge and discharge and discharge depth will affect its life, so its operating state must be constrained, including SOC, charging current I c and discharge current I d , the battery operation constraint is: SOC min ≤SOC≤SOC max (22) I c ≤I c,max (23) I d ≤I d,max (twenty four)

[0049] In formula (22)-(24), SOC min SOC is the lower limit of the battery state of charge; max is the upper limit of the battery state of charge; I c,max The upper limit of the battery charging current; I d,max It is the upper limit of battery discharge current.

[0050] As a further improvement of the present technical solution, in step S2, the specific method for predicting the battery life includes the following steps:

[0051] S2.1. Battery life prediction model based on improved throughput method:

[0052] The life of a battery is generally expressed by the number of cycles at different depths of discharge given by the manufacturer. The function of the total number of cycles of the battery is as follows:

[0053] In formula (25), h DOD is the depth of discharge of the battery; a1-a5 are parameters provided by the manufacturer;

[0054] However, batteries are usually charged and discharged frequently, and the depth of discharge cannot be guaranteed to be the same each time, which makes it difficult to calculate the number of cycles. The throughput method is a simple and easy life prediction method proposed to address the difficulty in counting the number of cycles in battery operation. The life calculation formula is as follows:

[0055] In formula (26), E throughput is the total energy throughput during the battery life; E rated is the rated capacity of the battery; n test is the number of tests on different depths of discharge of the battery; h DODi is the discharge depth of the battery in the i-th test; N i is the total number of cycles of the battery in the i-th test; the number of cycles at different depths of discharge and the corresponding total energy throughput are provided by the manufacturer;

[0056] Assuming that the total energy throughput of the battery is a certain value, the life loss coefficient formula is as follows:

[0057] In formula (27), E loss To calculate the energy throughput of the battery during the cycle, the following formula is shown:

[0058] In formula (28), T is the calculation time period; P(t) is the charge and discharge power of the battery at time t;

[0059] The simple throughput method can roughly calculate the battery life, but the actual battery life is affected by several factors. The most important factor in economic scheduling is the battery's SOC value. When a battery charges or discharges the same amount of energy, the loss of battery life is greater when the state of charge is lower, and less when the state of charge is higher. The battery equivalent energy throughput after considering the battery life loss weight is:

[0060] In formula (29), f(SOC(t)) is the battery life loss weight related to the state of charge;

[0061] S2.2. Battery life loss treatment:

[0062] In order to calculate the battery life more accurately, the battery state of charge factor is taken into account, but this introduces high-order variables, see formula (29), which makes formula (17) a nonlinear constraint; in order to simplify the model, formula (29) is reduced in order and piecewise linearized. This scheme introduces the cumulative function f of the battery life loss weight a (SOC), the formula is as follows:

[0063] The linear approximation function of formula (30) can be written as:

[0064] In formula (31) and formula (32), z k A variable introduced during the process of reducing the order and piecewise linearization, z k ≥0;

[0065] Formula (29) can be rewritten as a linear constraint form:

[0066] In formula (33), E ratedis the rated capacity of the battery; SOC(t) and SOC(t-1) represent the battery charge capacity at time t and time t-1 respectively; SOC(0) is the initial remaining capacity of the battery before economic dispatch.

[0067] As a further improvement of the present technical solution, in step S3, a specific method for establishing an electricity price response model based on the principles of consumer psychology and integrating price-based demand response into economic scheduling includes the following steps:

[0068] S3.1. Electricity Price Response Model Based on Consumer Psychology Principles:

[0069] Electricity is a special commodity, and electricity users are consumers. The psychological changes of users when using electricity are in line with the principles of consumer psychology. According to the principles of consumer psychology, users' electricity consumption behavior has three response areas to fluctuations in electricity prices, namely the response saturation area, the response linear area and the response dead area. This scheme adopts a demand response model based on the principles of consumer psychology to participate in economic scheduling.

[0070] A piecewise linear function can be used to describe how users' electricity consumption responds to changes in electricity prices. This function contains three important parameters: the dead zone threshold, the saturation zone threshold, and the linear zone slope. When electricity prices are set higher during peak load periods, users will reduce their electricity consumption to reduce their bills. This means that users shift their electricity consumption from peak periods to normal and off-peak periods. The amount of this shift is described by the load shift rate:

[0071] In formula (34), a pvj is the dead zone threshold; b pvj is the saturation threshold; k pvj is the slope of the linear region; ΔP pvj represents the difference between peak and valley electricity prices; j represents the jth type of user (such as enterprises, schools, residents, etc.); Indicates the maximum value of load transfer rate;

[0072] After obtaining the load transfer rate, the system power consumption in each period after implementing demand response can be calculated as follows:

[0073] In formula (35), the subscript 0 indicates that the demand response project is not implemented; the superscript t indicates the tth period; represents the electricity consumption after the implementation of the demand response project; Indicates the electricity consumption when the demand response project is not implemented; L pav Indicates the average electricity consumption during peak hours when the demand response project is not implemented; L fav represents the average electricity consumption during normal hours when the demand response project is not implemented; T P 、Tf 、T v Represent peak, flat and valley time sets respectively; λ pvj is the load transfer rate from peak period to valley period; fvj is the load transfer rate from normal period to off-peak period; pfj is the load transfer rate from peak period to normal period; T is the scheduling period, which is generally 24 hours;

[0074] S3.2. Use peak and valley time-of-use electricity prices (TOU) to guide and change users’ electricity consumption structure and usage patterns:

[0075] In order to avoid users' resistance to the implementation of demand response projects due to unreasonable peak and valley time-of-use prices (TOU), which hinders the implementation of demand response projects, the concept of user electricity satisfaction is introduced into power generation planning. User electricity satisfaction describes the user's electricity consumption experience during the implementation of demand response, including electricity consumption mode satisfaction ω and electricity bill expenditure satisfaction μ, as shown in the following formula: TOU =v1ω+v2μ (36)

[0076] In formula (36), v1 is the weight coefficient of satisfaction with electricity consumption mode; v2 is the weight coefficient of satisfaction with electricity expenditure;

[0077] The implementation of TOU (Time-of-Use) electricity pricing will change users' electricity usage habits. Users will adjust their electricity usage structure appropriately according to the difference between peak and off-peak electricity prices to reduce electricity bills. Before TOU was implemented, users would reasonably arrange their electricity usage based on their own schedules and electricity needs. At this time, their electricity usage satisfaction was the highest, with a maximum value of 1. However, after TOU was implemented, users were forced to transfer electricity from high-price periods to normal and off-peak periods, or from normal periods to off-peak periods. The expression for electricity usage satisfaction is as follows:

[0078] Through the leverage of TOU, users will change their most comfortable electricity consumption structure, which will correspondingly reduce their electricity bills. The electricity bill satisfaction μ describes the changes in users' electricity bills before and after TOU implementation, as shown below:

[0079] In formula (38), p0, p t Respectively represent the electricity prices before and after TOU implementation; μ = 1 means that the electricity expenditure before and after TOU implementation remains unchanged, μ> 1 means that the electricity expenditure before and after TOU implementation has decreased, and μ< 1 means that the electricity expenditure before and after TOU implementation has increased; thus: p t,min ≤p t ≤p t,max (39)

[0080] In formula (39), p t,min and p t,max Respectively represent the maximum and minimum values ​​after TOU implementation.

[0081] As a further improvement of the present technical solution, in step S3.1, in the electricity price response model based on the principles of consumer psychology, the three response areas of the user's electricity consumption behavior to electricity price fluctuations, namely, the response saturation area, the response linear area, and the response dead area, specifically include:

[0082] The response saturation zone means that when the electricity price increases to a certain threshold, users will not respond to the change in electricity price, that is, they will no longer reduce their electricity consumption due to the increase in electricity price;

[0083] The linear response region means that users are relatively sensitive to electricity price fluctuations. As electricity prices increase, electricity consumption will decrease. Conversely, as electricity prices decrease, electricity consumption will increase. The relationship between electricity consumption and electricity prices is linear.

[0084] The response dead zone means that when the electricity price drops to a certain threshold, users will no longer respond to changes in electricity prices on the basis of meeting their daily needs, that is, they will no longer increase their electricity consumption due to the reduction in electricity prices.

[0085] As a further improvement of the present technical solution, in step S4, the Giant Trevally Optimizer (GTO) is used to solve the wind-thermal-storage optimal economic dispatch problem taking into account electricity price demand response. The specific method for obtaining the optimal solution includes the following steps:

[0086] S4.1. Population initialization:

[0087] In GTO, each giant trevally is a feasible solution or candidate solution to the optimization problem; from a mathematical point of view, each member of the population is a vector, which constitutes the algorithm's population matrix:

[0088] In formula (40), X is the candidate solution of GTO; X i is the i-th candidate solution of GTO; Dim is the number of decision variables of the given problem; N is the number of GTO population; x i,j is the jth variable value of the i-th candidate solution;

[0089] In the solution space of the problem, positions must be randomly assigned to each giant trevally before they can begin to play a role; this random assignment must cover all feasible areas in the N×Dim dimensional search space, as shown in the following formula: i,j =Minimum j +(Maximum j -Minimum j )×R (41)

[0090] Where, i=1,2,...,N;j=1,2,...,Dim;R is a random number in [0,1],Maximum j and Minimum j is the upper and lower bounds of the population of the j-th dimension problem;

[0091] The objective function can be evaluated using each candidate solution:

[0092] In formula (42), F i represents the function value of the i-th member, and F represents the fitness function;

[0093] S4.2. Mathematical model of GTO:

[0094] GTO mimics the behavior of giant trevally hunting seabirds; therefore, GTO can be roughly divided into three steps: extensive search using Levy flight, defining the hunting area, and pursuing and attacking prey by jumping out of the water. The first step of GTO is the exploration phase, while the second and third steps are the development phase.

[0095] As a further improvement of the present technical solution, in step S4.2, the specific method of performing calculations on the mathematical model of GTO in three steps includes the following steps:

[0096] Step 1: Search widely; use the following formula to imitate the feeding movement of giant trevally: X(t+1)=Best p ×R+((Maximum-Minimum)×R+Minimum)×Levy(Dim) (43)

[0097] In formula (43), X(t+1) is the position of the giant trevally in the next iteration, Best p is the best position from the previous search, and Levy(Dim) is a Levy flight, a special type of non-Gaussian random process whose step size is determined by the Levy distribution. The algorithm's ability to perform global searches is due to its large random step size. Furthermore, the main advantage of using Levy flights is avoiding local optima and accelerating convergence.

[0098] Levy flight is calculated using the following formula:

[0099] In formula (44), step is the step size, which is fixed at 0.01; β is the distribution function exponent of the Levy flight, which ranges from 0 to 2 and is set to 1.5 in this scheme; u and v are random numbers with a normal distribution in the range of (0, 1); σ is calculated by the following formula:

[0100] In formula (45), Γ() is the gamma function;

[0101] Step 2: Select hunting area: Giant trevally determine and select the best area in the selected search space based on the amount of food (seabirds) where they can hunt prey; the following formula simulates this behavior: X(t+1)=Best p ×A×R+Mean_Info-X i (t)×R (46)

[0102] In formula (46), Best p is the current optimal solution, A is the position change control parameter, ranging from 0.3 to 0.4; Mean_Info is the mean information, indicating that these giant trevally have used up all the previously available information, calculated by the following formula:

[0103] Among them, X i (t) is the current position of the i-th giant trevally;

[0104] Step 3: Chasing and attacking prey: This step represents the development phase of the GTO. To simulate the behavior of the giant trevally when chasing and attacking prey, the GTO assumes that the giant trevally will be affected by light refraction and experience visual distortion. Snell's law is used for calculations; Snell's law is: η1sinθ1=η2sinθ2 (48)

[0105] Where η1 = 1.00029, η2 = 1.33; θ1 and θ2 represent the angle of incidence and the angle of refraction, respectively;

[0106] Simulate the behavior of giant trevally when chasing and jumping out of the water: X(t+1)=ψ+Γ+Η (49)

[0107] In formula (49), X(t+1) is the acceleration solution generated by the attack; ψ is the launch velocity; Γ represents the visual distortion; and H is the jump slope function, which enables the algorithm to adaptively perform the appropriate transition from the exploration phase to the exploitation phase. It can be calculated using the following formula:

[0108] In formula (50), t and T are the current number of iterations and the maximum number of iterations, respectively;

[0109] The optimal solution obtained after the algorithm iterative convergence is what we are looking for.

[0110] The second object of the present invention is to provide an optimal economic dispatch platform device, including a processor, a memory, and a computer program stored in the memory and running on the processor, wherein the processor is used to implement the steps of the above-mentioned wind-fire-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account electricity price demand response when executing the computer program.

[0111] The third object of the present invention is to provide a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the above-mentioned wind-fire-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account the electricity price demand response.

[0112] Compared with the prior art, the present invention has the following beneficial effects:

[0113] 1. This wind-thermal-storage optimal economic dispatch method, based on the Giant Scad Optimizer and taking into account electricity price demand response, innovatively combines advanced GTO with multi-objective optimal economic dispatch of power grids, demonstrating excellent performance in handling the wind-thermal-storage optimal economic dispatch problem taking into account electricity price demand response. It also uses peak and valley time-of-use electricity prices to guide changes in users' electricity consumption structure and patterns, and applies a weighted throughput method to predict battery life.

[0114] 2. This wind-thermal-storage optimal economic dispatch method, based on the Giant Trevally optimizer and taking into account electricity price demand response, uses a weighted throughput method to predict battery life. This improved weighted throughput method, which takes factors affecting battery life into account in a weighted manner, improves the accuracy of the model while maintaining the simplicity of the throughput method.

[0115] 3. This wind-thermal-storage optimal economic dispatch method, based on the Giant Trevally optimizer and taking into account electricity price demand response, employs a demand response model based on the principles of consumer psychology. Compared to the two existing demand response modeling methods, this one, based on the principles of consumer psychology, is more consistent with consumer behavior and is less affected by uncertainties such as season and weather.

[0116] 4. In this wind-thermal-storage optimal economic dispatch method based on the giant scad optimizer and taking into account electricity price demand response, GTO is used to solve the wind-thermal-storage optimal economic dispatch problem taking into account electricity price demand response; this scheme innovatively combines GTO with the multi-objective optimal economic dispatch of the power grid. It can conduct a global search to avoid falling into local optimality, while having a fast convergence speed and high accuracy. It can provide scientific and reasonable dispatching plans for dispatching and operating personnel, avoiding the blindness of dispatching and operating personnel's decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0117] FIG1 is a schematic diagram of an exemplary overall method flow chart of the present invention;

[0118] FIG2 is a schematic diagram of an exemplary GTO process of the present invention;

[0119] FIG3 is a schematic diagram of the light refraction principle of an exemplary GTO in the present invention;

[0120] FIG4 is a schematic diagram of an exemplary GTO visual distortion in the present invention;

[0121] FIG5 is a structural diagram of an exemplary electronic computer platform device in the present invention. DETAILED DESCRIPTION

[0122] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0123] Example 1

[0124] As shown in Figures 1-4, this embodiment provides a wind-thermal-storage optimal economic dispatch method based on the Giant Trevally Optimizer and taking into account electricity price demand response. The design philosophy of this method is as follows: First, starting from the goal of minimizing total power generation cost, a wind-thermal-storage joint economic dispatch model is established, taking into account various constraints in actual operation; second, the battery life factor is considered in the economic dispatch of the microgrid, and the weighted throughput method is applied to predict the battery life; furthermore, an electricity price response model based on consumer psychology principles is established to integrate price-based demand response into economic dispatch; finally, GTO is extended to the field of multi-objective optimization and used to solve the wind-thermal-storage optimal economic dispatch problem taking into account electricity price demand response. As shown in Figure 1, the method specifically includes the following steps.

[0125] S1. Starting from the goal of minimizing the total power generation cost, a wind-thermal-storage combined economic dispatch model is established, taking into account various constraints in actual operation;

[0126] In this step, the specific method of establishing the wind-thermal-storage joint economic dispatch model includes the following steps:

[0127] S1.1. Wind power modeling based on Weibull distribution:

[0128] The probability distribution parameter of wind speed is the most representative indicator of the statistical characteristics of local wind resources and is also an important parameter in wind power research. Many models can be used to approximate the probability distribution of wind speed, including Weibull distribution, Rayleigh distribution, lognormal distribution, and normal distribution. Because the two-parameter Weibull model can adjust parameters to adapt to the periodic changes in wind speed, it is widely used in wind speed probability modeling. The probability density function that obeys the Weibull distribution is defined as follows:

[0129] In formula (1), the subscript t represents the time interval per hour, v is the wind speed, and v t is the wind speed at time t, k t is the shape parameter of the wind speed probability distribution function at time t, which determines the overall shape of the wind speed probability distribution density curve, c t is the scale parameter at time t, reflecting the average wind speed of the wind field, which plays the role of amplifying or shrinking the curve, but does not affect the shape of the distribution density curve; the corresponding cumulative distribution function is:

[0130] Assuming that the output characteristic of the wind turbine is a simplified piecewise linear function, its output power for a given wind speed v is defined as follows:

[0131] In formula (3), P Wt is the output power of a single wind turbine; is the rated power of the fan; v in is the cut-in wind speed of the wind turbine; v r is the rated wind speed of the wind turbine; v out is the cut-out wind speed of the wind turbine;

[0132] When the output power of the wind turbine is within a continuously changing range, its Weibull probability density function is as follows:

[0133] In formula (4),

[0134] S1.2, Battery Model:

[0135] The prediction of battery state of charge (SOC) is an important part of power generation planning and economic balance. The formula for calculating SOC using the ampere-hour method is:

[0136] In formula (5), SOC0 is the initial state of charge; C A is the rated capacity of the battery; η is the charge and discharge coefficient of the battery; I is the charge and discharge current of the battery; t0 is the initial time of calculation; t is the end time of calculation;

[0137] In economic dispatch, the change of SOC is determined by the charge and discharge current; let the rated voltage of the system be U n , the load power requirement is P LD , the system remaining power is P e , then:

[0138] Discharge current I d :

[0139] Charging current I c :

[0140] Among them, U n is the system rated voltage, P LD is the load power demand, P e is the remaining power of the system;

[0141] S1.3. Wind-heat-storage joint economic dispatch model: This model aims to minimize the economic cost of the system. Its objective function is as follows: Min C total =C Wt,l (P l )+C DG,i (P i )+C OM,k (P k )+C DP,k (P k ) (8)

[0142] In formula (8), C Wt,l (P l ) is the cost of wind power generation, l=1,2,..,N l ; C DG,i (P i ) is the cost of thermal power generation, i=1,2,...,N i ; C OM,k (P k ) and C DP,k (P k ) are the maintenance cost and depreciation cost of the kth battery pack, k=1,2,...,N k ;

[0143] Typically, wind farms are privately owned. Grid operators are required to purchase electricity from wind power producers under power purchase agreements and pay a fee. The fee consists of three parts. The first part is the direct cost of purchasing wind power: d , Wt (P j )=d Wt,l P Wt,l,l=1,2,…,N l (9)

[0144] In formula (9), d Wt,l is the direct cost coefficient of wind power;

[0145] The second part is the penalty cost caused by underestimating wind power output. Due to the large fluctuation of wind power generation, wind power forecasts usually have some errors. When the planned wind power generation is lower than the available wind power generation, the excess wind energy will be wasted:

[0146] In formula (10), K ue,Wt,l is the underestimated coefficient of wind power; P Wt,rate,l is the rated power of the first wind turbine; P Wt,l is the planned output power of the first wind turbine;

[0147] The third component is the cost of ancillary services due to overestimation of wind power output; if the planned wind power generation is higher than the actual available wind power generation, then the load demand should be met by obtaining backup power from elsewhere:

[0148] In formula (11), K oe,Wt,l is the overestimation factor of wind power;

[0149] The total cost of wind power generation is: C Wt,l (P l )=C d,Wt (P j )+C ue,Wt,l (P l )+C oe,Wt,l (P l ),l=1,2,…,N l (12)

[0150] The total cost of a thermal power plant mainly consists of the economic cost of coal consumption and the operating cost of environmental protection equipment. The economic cost of coal consumption of a single thermal power plant depends on its power generation capacity and is generally approximated by quadratic function fitting. The formula for calculating the economic cost of coal consumption of all thermal power plants is:

[0151] In formula (13), P T,i (t) is the power generation capacity of thermal power plant i during period t; T is the number of dispatch periods; N T is the number of thermal power plants; a i 、b i 、c i is the operating cost coefficient of thermal power plant i;

[0152] The main pollutants emitted by thermal power plants are NO xand SO2; to reduce pollutant emissions, coal-fired boilers are retrofitted with ultra-low emissions and equipped with denitrification and desulfurization equipment; this embodiment uses the environmental cost per unit of power generated by the equipment to describe the operating costs of denitrification and desulfurization. To simplify the model, the denitrification and desulfurization cost database is referenced, and M0 is selected as the desulfurization cost per unit of power generation, and M1 is the denitrification cost per unit of power generation. The calculation formula for the operating cost of environmental protection equipment is:

[0153] The total cost of thermal power generation is: C DG,i (P i )=C CS,DG,i (P i )+C EC,DG,i (P i ) (15)

[0154] The maintenance cost and depreciation cost function of the energy storage device is: C OM,k (P k )=K OM,k |P k |Δt (16) C DP,k (P k )=(E rated,k C E,k +P rated,k C P,k )L loss,k (17)

[0155] In formula (16) and formula (17), K OM,k E is the unit power operation and maintenance cost coefficient of the kth battery group; rated,k 、P rated,k are the rated capacity and rated power of the kth battery pack respectively; C E,k 、C P,k are the present values ​​of the unit capacity and unit power installation costs of the kth battery pack respectively; L loss,k is the life loss coefficient of the kth battery pack;

[0156] S1.4 Constraints, including:

[0157] The power balance constraint is:

[0158] In formula (18), P d P is the power demanded by the electric load; loss For network loss;

[0159] The operating constraints of wind power generation are: 0≤P Wt,l ≤P Wt,rate,l ,l=1,2,…,N l (19)

[0160] The maximum and minimum output constraints of thermal power plants are: P T,i,min ≤P T,i (t)≤P T,i,max (20)

[0161] Where, P T,i,max and P T,i,min are the upper and lower limits of the output of thermal power plant i, respectively;

[0162] The ramp constraint of the thermal power plant is: |P T,i (t)-P T,i (t-1)|≤P T,i speed (twenty one)

[0163] Where, P T,i speed is the upward or downward ramp rate limit of thermal power plant i;

[0164] The battery's cyclic charge and discharge and discharge depth will affect its life, so its operating state must be constrained, including SOC, charging current I c and discharge current I d , the battery operation constraint is: SOC min ≤SOC≤SOC max (22) I c ≤I c,max (23) I d ≤I d,max (twenty four)

[0165] In formula (22)-(24), SOC min SOC is the lower limit of the battery state of charge; max is the upper limit of the battery state of charge; I c,max The upper limit of the battery charging current; I d,max It is the upper limit of battery discharge current.

[0166] S2. Consider the battery life factor in the economic dispatch of the microgrid and apply the weighted throughput method to predict the battery life;

[0167] The weighted throughput method is a summary of battery experiments. It assumes that the total charge and discharge capacity of a battery over its lifetime does not change much at different depths of discharge. The total lifetime throughput is considered a constant value, namely, the average of the throughputs at different depths of discharge. The throughput method is simple and easy to implement, meets many computational requirements, and has experimental support, making it an excellent method for models that only require rough calculations. However, in actual operation, many factors affect battery life. By weighting these factors, an improved weighted throughput method is developed. This method maintains the simplicity of the throughput method while improving model accuracy.

[0168] In this step, the specific method for predicting the battery life includes the following steps:

[0169] S2.1. Battery life prediction model based on improved throughput method:

[0170] The life of a battery is generally expressed by the number of cycles at different depths of discharge given by the manufacturer. The function of the total number of cycles of the battery is as follows:

[0171] In formula (25), h DOD is the depth of discharge of the battery; a1-a5 are parameters provided by the manufacturer;

[0172] However, batteries are usually charged and discharged frequently, and the depth of discharge cannot be guaranteed to be the same each time, which makes it difficult to calculate the number of cycles. The throughput method is a simple and easy life prediction method proposed to address the difficulty in counting the number of cycles in battery operation. The life calculation formula is as follows:

[0173] In formula (26), E throughput is the total energy throughput during the battery life; E rated is the rated capacity of the battery; n test is the number of tests on different depths of discharge of the battery; h DODi is the discharge depth of the battery in the i-th test; N i is the total number of cycles of the battery in the i-th test; the number of cycles at different depths of discharge and the corresponding total energy throughput are provided by the manufacturer;

[0174] Assuming that the total energy throughput of the battery is a certain value, the life loss coefficient formula is as follows:

[0175] In formula (27), E loss To calculate the energy throughput of the battery during the cycle, the following formula is shown:

[0176] In formula (28), T is the calculation time period; P(t) is the charge and discharge power of the battery at time t;

[0177] The simple throughput method can roughly calculate the battery life, but the actual battery life is affected by several factors. The most important factor in economic scheduling is the battery's SOC value. When a battery charges or discharges the same amount of energy, the loss of battery life is greater when the state of charge is lower, and less when the state of charge is higher. The battery equivalent energy throughput after considering the battery life loss weight is:

[0178] In formula (29), f(SOC(t)) is the battery life loss weight related to the state of charge;

[0179] The improved throughput method, while maintaining its original simplicity, takes into account the state of charge factor that affects battery life, making the model more accurate.

[0180] S2.2. Battery life loss treatment:

[0181] In order to calculate the battery life more accurately, the battery state of charge factor is taken into account. However, this introduces a high-order variable, see formula (29), which makes formula (17) a nonlinear constraint. In order to simplify the model, formula (29) is reduced in order and piecewise linearized. This embodiment introduces the cumulative function f of the battery life loss weight a (SOC), the formula is as follows:

[0182] The linear approximation function of formula (30) can be written as:

[0183] In formula (31) and formula (32), z k A variable introduced during the process of reducing the order and piecewise linearization, z k ≥0;

[0184] Formula (29) can be rewritten as a linear constraint form:

[0185] In formula (33), E rated is the rated capacity of the battery; SOC(t) and SOC(t-1) represent the battery charge capacity at time t and time t-1 respectively; SOC(0) is the initial remaining capacity of the battery before economic dispatch.

[0186] S3. Establish an electricity price response model based on consumer psychology principles and integrate price-based demand response into economic scheduling;

[0187] Among them, price-based demand response includes peak-valley time-of-use electricity prices, peak electricity prices and real-time electricity prices. Its essence is to change the user's electricity consumption structure and electricity consumption methods through the guiding role of electricity prices, so that users reduce electricity consumption when the system load is peak and increase electricity consumption when the system load is at a low point. This can effectively reduce the peak-valley difference of the system load.

[0188] In this step, the specific method includes the following steps:

[0189] S3.1. Electricity Price Response Model Based on Consumer Psychology Principles:

[0190] Electricity is a special commodity, and electricity users are consumers. The psychological changes of users when using electricity conform to the principles of consumer psychology. According to the principles of consumer psychology, users' electricity usage behavior responds to fluctuations in electricity prices in three areas, namely the response saturation area, the response linear area, and the response dead area.

[0191] The response saturation zone means that when the electricity price increases to a certain threshold, users will not respond to the change in electricity price, that is, they will no longer reduce their electricity consumption due to the increase in electricity price;

[0192] The linear response region means that users are relatively sensitive to electricity price fluctuations. As electricity prices increase, electricity consumption will decrease. Conversely, as electricity prices decrease, electricity consumption will increase. The relationship between electricity consumption and electricity prices is linear.

[0193] The response dead zone means that when the electricity price drops to a certain threshold, users will no longer respond to changes in electricity prices on the basis of meeting their daily needs, that is, they will no longer increase their electricity consumption due to the reduction in electricity prices.

[0194] Furthermore, modeling methods based on the demand elasticity matrix require analyzing extensive market operating data to determine the self-elasticity coefficient and the mutual elasticity coefficient. These two elasticity coefficients are significantly correlated with load fluctuations, while demand load is significantly affected by seasons, weather, and holidays. Therefore, the model stability is poor, and regular revisions of the demand response parameters are required. The key to demand response modeling methods based on statistical principles lies in determining the appropriate curve model. The rationality of the curve model directly affects the accuracy of the user response calculation results. The selection of the curve model requires continuous trial and error, which is computationally intensive and may not always be the case. Compared to the previous two demand response modeling methods, the demand response modeling method based on consumer psychology principles takes into account user consumption psychology and is more consistent with consumer behavior. Furthermore, this modeling method is less affected by uncertainties such as season and weather. Therefore, in this proposal, a demand response model based on consumer psychology principles is adopted for economic dispatch.

[0195] A piecewise linear function can be used to describe how users' electricity consumption responds to changes in electricity prices. This function contains three important parameters: the dead zone threshold, the saturation zone threshold, and the linear zone slope. For example, consider peak-valley time-of-use electricity pricing. Assuming a higher price during peak load periods, users will reduce their electricity consumption to reduce their bills. This means they shift their electricity consumption from peak periods to off-peak and valley periods. The amount of this shift is described by the load shift rate:

[0196] In formula (34), a pvj is the dead zone threshold; b pvj is the saturation threshold; k pvj is the slope of the linear region; ΔP pvj represents the difference between peak and valley electricity prices; j represents the jth type of user (enterprise, school, resident, etc.); Indicates the maximum value of load transfer rate;

[0197] After obtaining the load transfer rate, the system power consumption in each period after implementing demand response can be calculated as follows:

[0198] In formula (35), the subscript 0 indicates that the demand response project is not implemented; the superscript t indicates the tth period; represents the electricity consumption after the implementation of the demand response project; Indicates the electricity consumption when the demand response project is not implemented; L pav Indicates the average electricity consumption during peak hours when the demand response project is not implemented; L fav represents the average electricity consumption during normal hours when the demand response project is not implemented; T P 、T f 、T v Represent peak, flat and valley time sets respectively; λ pvj is the load transfer rate from peak period to valley period; fvj is the load transfer rate from normal period to off-peak period; pfj is the load transfer rate from peak period to normal period; T is the scheduling period, which is generally 24 hours;

[0199] S3.2. Use peak and valley time-of-use electricity prices (TOU) to guide and change users’ electricity consumption structure and usage patterns:

[0200] In order to avoid users' resistance to the implementation of demand response projects due to unreasonable peak and valley time-of-use prices (TOU), which hinders the implementation of demand response projects, the concept of user electricity satisfaction is introduced into power generation planning. User electricity satisfaction describes the user's electricity consumption experience during the implementation of demand response, including electricity consumption mode satisfaction ω and electricity bill expenditure satisfaction μ, as shown in the following formula: TOU =v1ω+v2μ (36)

[0201] In formula (36), v1 is the weight coefficient of satisfaction with electricity consumption mode; v2 is the weight coefficient of satisfaction with electricity expenditure;

[0202] The implementation of TOU (Time-of-Use) electricity pricing will change users' electricity usage habits. Users will adjust their electricity usage structure appropriately according to the difference between peak and off-peak electricity prices to reduce electricity bills. Before TOU was implemented, users would reasonably arrange their electricity usage based on their own schedules and electricity needs. At this time, their electricity usage satisfaction was the highest, with a maximum value of 1. However, after TOU was implemented, users were forced to transfer electricity from high-price periods to normal and off-peak periods, or from normal periods to off-peak periods. The expression for electricity usage satisfaction is as follows:

[0203] Through the leverage of TOU, users will change their most comfortable electricity consumption structure, which will correspondingly reduce their electricity bills. The electricity bill satisfaction μ describes the changes in users' electricity bills before and after TOU implementation, as shown below:

[0204] In formula (38), p0, p t Respectively represent the electricity prices before and after TOU implementation; μ = 1 means that the electricity expenditure before and after TOU implementation remains unchanged, μ> 1 means that the electricity expenditure before and after TOU implementation has decreased, and μ< 1 means that the electricity expenditure before and after TOU implementation has increased; thus: p t,min ≤p t ≤p t,max (39)

[0205] In formula (39), p t,min and p t,max Respectively represent the maximum and minimum values ​​after TOU implementation.

[0206] S4. Extend the Giant Trevally Optimizer (GTO) to the field of multi-objective optimization, and use GTO to solve the wind-thermal-storage optimal economic dispatch problem taking into account electricity price demand response, obtaining the optimal solution and thus achieving optimal economic dispatch;

[0207] Among them, GTO is a new metaheuristic algorithm, mainly inspired by the process of giant trevally preying on seabirds. Giant trevally is a large marine fish belonging to the family of trevallies, also known as the giant kingfish. Its daily diet includes fish, cephalopods, crustaceans, and seabirds. As the search space expands, giant trevally move within its ecosystem and between habitats. Giant trevally hunt alone or in groups. Group hunting is more efficient, and the most effective member in capturing prey is the leader or first predator. After defining the hunting area, giant trevally begins to stalk its prey and then jumps out of the water to attack the prey. These novel hunting strategies - mobile foraging patterns, selecting suitable areas based on the amount of food, and jumping out of the water to attack and capture prey - are the main inspiration for the design of GTO.

[0208] In this step, as shown in FIG2 , the specific method includes the following steps:

[0209] S4.1. Population initialization:

[0210] In GTO, each giant trevally is a feasible solution or candidate solution to the optimization problem; from a mathematical point of view, each member of the population is a vector, which constitutes the algorithm's population matrix:

[0211] In formula (40), X is the candidate solution of GTO; X i is the i-th candidate solution of GTO; Dim is the number of decision variables of the given problem; N is the number of GTO population; x i,j is the jth variable value of the i-th candidate solution;

[0212] In the solution space of the problem, positions must be randomly assigned to each giant trevally before they can begin to play a role; this random assignment must cover all feasible areas in the N×Dim dimensional search space, as shown in the following formula: i,j =Minimum j +(Maximum j -Minimum j )×R (41)

[0213] Where, i=1,2,...,N;j=1,2,...,Dim;R is a random number in [0,1],Maximum j and Minimum j is the upper and lower bounds of the population of the j-th dimension problem;

[0214] The objective function can be evaluated using each candidate solution:

[0215] In formula (42), F i represents the function value of the i-th member, and F represents the fitness function;

[0216] S4.2. Mathematical model of GTO:

[0217] GTO mimics the behavior of giant trevally hunting seabirds; therefore, GTO can be roughly divided into three steps: extensive search using Levy flight, defining the hunting area, and pursuing and attacking prey by jumping out of the water. The first step of GTO is the exploration phase, while the second and third steps are the development phase.

[0218] Specifically, the mathematical model of GTO is operated in three steps, including the following steps:

[0219] Step 1: Search widely; use the following formula to imitate the feeding movement of giant trevally: X(t+1)=Best p ×R+((Maximum-Minimum)×R+Minimum)×Levy(Dim) (43)

[0220] In formula (43), X(t+1) is the position of the giant trevally in the next iteration, Best p is the best position from the previous search, and Levy(Dim) is a Levy flight, a special type of non-Gaussian random process whose step size is determined by the Levy distribution. The algorithm's ability to perform global searches is due to its large random step size. Furthermore, the main advantage of using Levy flights is avoiding local optima and accelerating convergence.

[0221] Levy flight is calculated using the following formula:

[0222] In formula (44), step is the step size, which is fixed at 0.01; β is the Levy flight distribution function exponent, which ranges from 0 to 2 and is preferably set to 1.5 in this embodiment; u and v are random numbers normally distributed in the range (0, 1); σ is calculated by the following formula:

[0223] In formula (45), Γ() is the gamma function;

[0224] Step 2: Select hunting area: Giant trevally determine and select the best area in the selected search space based on the amount of food (seabirds) where they can hunt prey; the following formula simulates this behavior: X(t+1)=Best p ×A×R+Mean_Info-X i (t)×R (46)

[0225] In formula (46), Best pis the current optimal solution, A is the position change control parameter, ranging from 0.3 to 0.4; Mean_Info is the mean information, indicating that these giant trevally have used up all the previously available information, calculated by the following formula:

[0226] Among them, X i (t) is the current position of the i-th giant trevally;

[0227] Step 3: Chasing and attacking prey: This step represents the development phase of the GTO, as shown in Figures 3 and 4. To simulate the behavior of the giant trevally when chasing and attacking prey, the GTO assumes that the giant trevally will be affected by light refraction and experience visual distortion. Snell's law is used for calculations; Snell's law is: η1sinθ1=η2sinθ2 (48)

[0228] Where η1 = 1.00029, η2 = 1.33; θ1 and θ2 represent the angle of incidence and the angle of refraction, respectively;

[0229] Simulate the behavior of giant trevally when chasing and jumping out of the water: X(t+1)=ψ+Γ+Η (49)

[0230] In formula (49), X(t+1) is the acceleration solution generated by the attack; ψ is the launch velocity; Γ represents the visual distortion; and H is the jump slope function, which enables the algorithm to adaptively perform the appropriate transition from the exploration phase to the exploitation phase. It can be calculated using the following formula:

[0231] In formula (50), t and T are the current number of iterations and the maximum number of iterations, respectively;

[0232] The optimal solution obtained after the algorithm iterative convergence is what we are looking for.

[0233] As shown in FIG5 , this embodiment further provides an optimal economic dispatch platform device, which includes a processor, a memory, and a computer program stored in the memory and running on the processor.

[0234] The processor includes one or more processing cores, and the processor is connected to the memory through a bus. The memory is used to store program instructions. When the processor executes the program instructions in the memory, the steps of the above-mentioned wind-thermal-storage optimal economic dispatch method based on the giant scad optimizer and taking into account the electricity price demand response are implemented.

[0235] Alternatively, the memory may be implemented by any type of volatile or non-volatile memory device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0236] In addition, the present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the above-mentioned wind-fire-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account the electricity price demand response.

[0237] Optionally, the present invention also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the steps of the above-mentioned wind-thermal-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account electricity price demand response.

Claims

1. A wind-fire-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account the electricity price demand response, characterized in that: The steps include: S1. Starting from the goal of minimizing the total power generation cost, a wind-fire-storage joint economic dispatch model is established, and various constraints in actual operation are considered; S2. Consider the battery life factor in the economic dispatch of the microgrid and use the weighted throughput method to predict the battery life; S3. Establish an electricity price response model based on the principles of consumer psychology and integrate price-based demand response into economic dispatch; S4. Extend the Giant Trevally Optimizer (GTO) to the field of multi-objective optimization, and use GTO to solve the wind-fire-storage optimal economic dispatch problem taking into account electricity price demand response, obtain the optimal solution, and thus achieve optimal economic dispatch.

2. The wind-fire-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account the electricity price demand response according to claim 1 is characterized in that: In step S1, the specific method for establishing the wind-fire-storage joint economic dispatch model includes the following steps: S1.

1. Wind power modeling based on Weibull distribution; the probability density function that obeys the Weibull distribution is defined as follows: In formula (1), the subscript t represents the time interval per hour, v is the wind speed, and v t is the wind speed at time t, k t is the shape parameter of the wind speed probability distribution function at time t, which determines the overall shape of the wind speed probability distribution density curve, c t is the scale parameter at time t, reflecting the average wind speed of the wind field, which plays the role of enlarging or shrinking the curve, but does not affect the shape of the distribution density curve; the corresponding cumulative distribution function is: Assuming that the output characteristic of the wind turbine is a simplified piecewise linear function, its output power for a given wind speed v is defined as follows: In formula (3), P Wt is the output power of a single wind turbine; is the rated power of the fan; v in is the cut-in wind speed of the wind turbine; v r is the rated wind speed of the wind turbine; v out is the cut-out wind speed of the wind turbine; When the output power of the fan is within a continuously changing range, its Weibull probability density function is as follows: In formula (4), S1.2, battery model; the formula for calculating SOC using the ampere-hour method is: In formula (5), SOC0 is the initial state of charge; C A is the rated capacity of the battery; η is the charge and discharge coefficient of the battery; I is the charge and discharge current of the battery; t0 is the initial time of calculation; t is the end time of the calculation; In economic dispatch, the change of SOC is determined by the charge and discharge current; assuming the system rated voltage is U n , the load power required is P LD , the remaining power of the system is P e , then: Discharge current I d : Charging current I c : Among them, U n is the system rated voltage, P LD is the load demand power, Pe is the system surplus power; S1.

3. Wind-heat-storage joint economic dispatch model: This model aims to minimize the economic cost of the system, and its objective function is as follows: Min C total =C Wt,l (P l )+C DG,i (P i )+C OM,k (P k )+C DP,k (P k ) (8) In formula (8), C Wt,l (P l ) is the cost of wind power generation, l=1,2,..,N l ; C DG,i (P i ) is the cost of thermal power generation, i=1,2,...,N i ; C OM,k (P k ) and C DP,k (P k ) are the maintenance cost and depreciation cost of the kth battery pack, k=1,2,...,N k ; Typically, wind farms are privately owned, and grid operators need to purchase electricity from wind power producers under power purchase agreements and pay a certain fee; the fee consists of three parts. The first part is the direct cost of purchasing wind power: C d,Wt (P j )=d Wt,l P Wt,l, l=1,2,…,N l (9) In formula (9), d Wt,l is the direct cost coefficient of wind power; The second part is the penalty cost caused by underestimating wind power output. Due to the large volatility of wind power generation, wind power forecasts usually have some errors; when the planned wind power generation is lower than the available wind power generation, the excess wind energy will be wasted: In formula (10), K ue,Wt,l is the underestimated coefficient of wind power; P Wt,rate,l is the rated power of the first wind turbine; P Wt,l is the planned output power of the first wind turbine; The third component is the cost of ancillary services due to overestimation of wind power output; if the planned wind power generation is higher than the actual available wind power generation, then the load demand should be met by obtaining backup power from elsewhere: In formula (11), K oe,Wt,l is the overestimation factor of wind power; The total cost of wind power generation is then: C Wt,l (P l )=C d,Wt (P j )+C ue,Wt,l (P l )+C oe,Wt,l (P l ),l=1,2,…,N l (12) The total cost of a thermal power plant is mainly the economic cost of coal consumption and the operating cost of environmental protection equipment. The economic cost of coal consumption of a single thermal power plant depends on its power generation capacity, which is generally approximated by quadratic function fitting. The calculation formula for the economic cost of coal consumption of all thermal power plants is: In formula (13), P T,i (t) is the power generation of thermal power plant i in period t; T is the number of dispatching periods; N T is the number of thermal power plants; a i 、b i 、c i is the operating cost coefficient of thermal power plant i; The main pollutants emitted by thermal power plants are NO x and SO2; to reduce pollutant emissions, coal-fired boilers are transformed into ultra-low emissions and denitrification and desulfurization equipment is installed; the environmental cost per unit of power generation of the equipment is used to describe the operating costs of denitrification and desulfurization. To simplify the model, the denitrification and desulfurization cost database is referenced, M0 is selected as the desulfurization cost per unit of power generation, and M1 is the denitrification cost per unit of power generation. The calculation formula for the operating cost of environmental protection equipment is: The total cost of thermal power generation is: C DG,i (P i )=C CS,DG,i (P i )+C EC,DG,i (P i ) (15) The maintenance cost and depreciation cost function of the energy storage device is: C OM,k (P k )=K OM,k |P k |Δt (16) C DP,k (P k )=(E rated,k C E,k +P rated,k C P,k )L loss,k (17) In formula (16) and formula (17), K OM,k E is the unit power operation and maintenance cost coefficient of the kth battery group; rated,k , P rated,k are the rated capacity and rated power of the kth battery group respectively; C E,k , C P,k are the unit capacity and unit power of the kth battery group respectively. The present value of the installation cost; L loss,k is the life loss coefficient of the kth battery pack; S1.4, Constraints, including: The power balance constraint is: In formula (18), P d P is the power demanded by the electrical load; loss For network loss; The operating constraints of wind power generation are: 0≤P Wt,l ≤P Wt,rate,l ,l=1,2,…,N l (19) The maximum and minimum output constraints of thermal power plants are: P T,i,min ≤P T,i (t)≤P T,i,max (20) Where P T,i,max and P T,i,min are the upper and lower limits of the output of thermal power plant i respectively; The ramp constraints of thermal power plants are: |P T,i (t)-P T,i (t-1)|≤P T,i speed (21) Where P T,i speed is the upward or downward ramp rate limit of thermal power plant i; The battery's cyclic charge and discharge and discharge depth will affect its life, so its operating state must be constrained, including SOC, charging current I c and the discharge current I d , the operating constraints of the battery are: SOC min ≤SOC≤SOC max (22) I c ≤I c,max (23) I d ≤I d,max (24) In formula (22)-(24), SOC min SOC is the lower limit of the battery state of charge; max I is the upper limit of the battery charge state; c,max The upper limit of the battery charging current; I d,max It is the upper limit of battery discharge current.

3. The wind-fire-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account the electricity price demand response according to claim 2 is characterized in that: In step S2, the specific method for predicting the life of the battery includes the following steps: S2.

1. Battery life prediction model based on improved throughput method: The life of a battery is generally expressed by the number of cycles at different depths of discharge given by the manufacturer. The total number of cycles of the battery is expressed as follows: Down: In formula (25), h DOD is the discharge depth of the battery; a1-a5 are parameters provided by the manufacturer; The throughput method is a simple and easy life prediction method proposed to address the difficulty in counting the number of cycles in battery operation. The calculation formula for its life is as follows: In formula (26), E throughput is the total energy throughput during the battery life; E rated is the rated capacity of the battery; n test is the number of tests on different depths of discharge of the battery; h DODi is the discharge depth of the battery in the i-th test; N i is the total number of cycles of the battery in the i-th test; Assuming that the total energy throughput of the battery is a certain value, the life loss coefficient formula is as follows: In formula (27), E loss To calculate the energy throughput of the battery during the cycle, it is shown in the following formula: E loss =∫0 T |P(t)|dt (28) In formula (28), T is the calculation time period; P(t) is the charge and discharge power of the battery at time t; The simple throughput method can roughly calculate the service life of the battery, but the actual service life of the battery will be affected by some factors. The most important influencing factor in economic scheduling is the SOC value of the battery. The battery equivalent energy throughput after considering the battery life loss weight is: E loss =∫0 T |P(t)|f(SOC(t))dt (29) In formula (29), f(SOC(t)) is the battery life loss weight related to the state of charge; S2.

2. Battery life loss processing: In order to simplify the model, equation (29) is downgraded and piecewise linearized, and the cumulative function f of the battery life loss weight is introduced a (SOC), the formula is as follows: f a (SOC)=∫0 SOC f(x)dx (30) The linear approximation function of formula (30) can be written as: In formula (31) and formula (32), z k A variable introduced during the process of downscaling and piecewise linearization, z k ≥0; Formula (29) can be rewritten as a linear constraint form: In formula (33), E rated is the rated capacity of the battery; SOC(t) and SOC(t-1) represent the battery charge capacity at time t and time t-1 respectively; SOC(0) is the initial remaining capacity of the battery before economic dispatch.

4. The wind-fire-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account the electricity price demand response according to claim 3 is characterized in that: In step S3, a specific method for establishing an electricity price response model based on the principle of consumer psychology and integrating price-based demand response into economic dispatch includes the following steps: S3.

1. Electricity price response model based on consumer psychology principles: According to the principles of consumer psychology, the user's electricity consumption behavior has three response areas to the fluctuation of electricity prices, namely the response saturation area, the response linear area and the response dead area. The demand response model based on the principles of consumer psychology is used to participate in economic dispatch; A piecewise linear function is used to describe the response behavior of user electricity consumption to changes in electricity prices. The piecewise linear function contains three important parameters: dead zone threshold, saturation zone threshold and linear zone slope. A higher electricity price is set during peak load. At this time, users will reduce electricity consumption to reduce electricity bills, that is, transfer electricity consumption during peak hours to normal and valley hours, and the amount of transfer is described by the load transfer rate: In formula (34), a pvj is the dead zone threshold; b pvj is the saturation zone threshold; k pvj is the slope of the linear region; ΔP pvj represents the peak-valley electricity price difference; j represents the jth type of user; Indicates the maximum value of load transfer rate; After obtaining the load transfer rate, the power consumption of the system in each period after implementing demand response can be calculated as follows: In formula (35), the subscript 0 indicates that no demand response project has been implemented; the superscript t indicates the tth period; represents the electricity consumption after the implementation of the demand response project; Represents the electricity consumption when the demand response project is not implemented; L pav Indicates the average electricity consumption during peak hours when the demand response project is not implemented; L fav represents the average electricity consumption during normal hours when no demand response project is implemented; T P , T f , T v Respectively represent the peak, flat and valley time sets; λ pvj is the load transfer rate from peak period to valley period; fvj is the load transfer rate from normal period to valley period; λ pfj is the load transfer rate from peak period to normal period; T represents the scheduling period; S3.

2. Change the electricity consumption structure and mode of users through peak and valley time-of-use electricity price TOU guidance: User electricity satisfaction describes the user's electricity experience during the implementation of demand response, including electricity consumption mode satisfaction ω and electricity expenditure satisfaction μ, as shown in the following formula: F TOU =v1ω+v2μ (36) In formula (36), v1 is the weight coefficient of satisfaction with electricity consumption mode; v2 is the weight coefficient of satisfaction with electricity expenditure; Before TOU is implemented, users will reasonably arrange their electricity usage according to their own work and rest time and electricity demand. At this time, the satisfaction with electricity usage is the highest, with a maximum value of 1. After TOU is implemented, users will be forced to transfer electricity from high-price periods to normal and valley periods or from normal periods to valley periods. The expression of electricity usage satisfaction is as follows: Through the leverage of TOU, users will change the most comfortable electricity consumption structure, and accordingly reduce electricity expenditure. The electricity expenditure satisfaction μ describes the changes in users' electricity expenditure before and after the implementation of TOU, as shown below: In formula (38), p0, p t They represent the electricity prices before and after the implementation of TOU; μ = 1 means that the electricity expenditure before and after the implementation of TOU remains unchanged, μ> 1 means that the electricity expenditure before and after the implementation of TOU has decreased, and μ< 1 means that the electricity expenditure before and after the implementation of TOU has increased; thus: p t,min ≤p t ≤p t,max (39) In formula (39), p t,min and p t,max Respectively represent the maximum and minimum values ​​after TOU implementation.

5. The wind-fire-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account the electricity price demand response according to claim 4 is characterized in that: In step S3.1, in the electricity price response model based on the principle of consumer psychology, the three response areas of the user's electricity consumption behavior to the fluctuation of electricity prices, namely, the response saturation area, the response linear area and the response dead area, specifically include: The response saturation zone means that when the electricity price increases to a certain threshold, users will not respond to the change in electricity price, that is, they will no longer reduce electricity consumption due to the increase in electricity price; The response linear region means that users are relatively sensitive to electricity price fluctuations at this time. The electricity consumption will decrease as the electricity price increases, and vice versa, the electricity consumption will increase as the electricity price decreases. The electricity consumption and electricity price are linearly related. The response dead zone means that when the electricity price drops to a certain threshold, users will no longer respond to changes in electricity prices on the basis of meeting their daily needs, that is, they will no longer increase their electricity consumption due to the reduction in electricity prices.

6. The wind-fire-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account the electricity price demand response according to claim 4 is characterized in that: In step S4, the Giant Trevally Optimizer (GTO) is used to solve the wind-fire-storage optimal economic dispatch problem taking into account the electricity price demand response. The specific method for obtaining the optimal solution includes the following steps: S4.

1. Population initialization: In GTO, each giant trevally is a feasible solution or candidate solution to the optimization problem; from a mathematical point of view, each member of the population is a vector, and these vectors constitute the algorithm's population matrix: In formula (40), X is the candidate solution of GTO; X i is the i-th candidate solution of GTO; Dim is the number of decision variables of the given problem; N is the number of GTO population; x i,j is the jth variable value of the i-th candidate solution; In the solution space of the problem, positions are randomly assigned to each giant trevally before they can begin to play a role; this random assignment must cover all feasible areas in the N×Dim dimensional search space, as shown in the following formula: X i,j =Minimum j +(Maximum j -Minimum j )×R (41) Where i = 1, 2, ..., N; j = 1, 2, ..., Dim; R is a random number in [0, 1], Maximum j and Minimum j is the upper and lower bounds of the population of the j-th dimension problem; The objective function can be evaluated using each candidate solution: In formula (42), F i represents the function value of the i-th member, and F represents the fitness function; S4.

2. Mathematical model of GTO: GTO mimics the behavior of giant trevally when hunting seabirds; GTO is roughly divided into the following three steps: using Levy flight to conduct a wide search, determine the hunting area, and chase and attack the prey by jumping out of the water.

7. The wind-fire-storage optimal economic dispatch method based on the giant trevally optimizer and taking into account the electricity price demand response according to claim 6 is characterized in that: In step S4.2, the specific method of calculating the mathematical model of GTO in three steps includes the following steps: Step 1: Search widely; use the following method to imitate the feeding movement of giant trevally: X(t+1)=Best p ×R+((Maximum-Minimum)×R+Minimum)×Levy(Dim) (43) In formula (43), X(t+1) is the position of the giant trevally in the next iteration, Best p is the best position of the last search, Levy(Dim) is the Levy flight, which is a special type of non-Gaussian random process, and its step size is determined by the Levy distribution; The Levy flight is calculated using the following formula: In formula (44), step is the step size; β is the distribution function index of Levy flight, which ranges from 0 to 2; u and v are random numbers with normal distribution in the range of (0, 1); σ is calculated by the following formula: In formula (45), Γ() is the gamma function; Step 2: Select hunting area: Giant trevally determine and select the best area in the selected search space based on the amount of food, where they can hunt prey; the following equation simulates this behavior: X(t+1)=Best p ×A×R+Mean_Info-X i (t)×R (46) In formula (46), Best p is the current optimal solution, A is the position change control parameter; Mean_Info is the mean information, indicating that these giant trevally have used up all the previously available information, calculated by the following formula: Among them, X i (t) is the current position of the ith giant trevally; Step 3: Chasing and attacking prey: This step represents the development stage of GTO. In order to simulate the behavior of giant trevally when chasing and attacking prey, GTO assumes that the giant trevally will be affected by light refraction and visual distortion occurs. Snell's law is used for calculation; Snell's law is: η1sinθ1=η2sinθ2 (48) Where, η1 = 1.00029, η2 = 1.33; θ1 and θ2 represent the angle of incidence and the angle of refraction respectively; Simulate the behavior of giant trevally when chasing and jumping out of the water: X(t+1)=ψ+Γ+H (49) In formula (49), X(t+1) is the acceleration solution generated by the attack; ψ is the launch speed; Γ represents the visual distortion; H is the jump slope function, which enables the algorithm to adaptively perform the appropriate transition from the exploration stage to the exploitation stage, which can be calculated using the following formula: In formula (50), t and T are the current iteration number and the maximum iteration number, respectively; When the algorithm converges after iteration, the optimal solution is obtained.

Images