Improved particle swarm optimization algorithm-based integrated energy system optimal scheduling method and system

Abstract

The application discloses an improved particle swarm algorithm-based comprehensive energy system optimal scheduling method and system, and belongs to the technical field of comprehensive energy system control. The method comprises the following steps: constructing a comprehensive energy system multi-objective optimization model with the output of a generator unit and the number of various energy supply devices of the comprehensive energy system as decision variables, and minimizing resource consumption while taking into account sociality and reliability; the model can reduce resource consumption while taking into account sociality and reliability, and can better improve the overall benefit of the comprehensive energy system; and the improved particle swarm algorithm is used to solve the comprehensive energy system multi-objective optimization model, and the optimal particle swarm is obtained, that is, the optimal operation scheduling plan of the comprehensive energy system; the improved particle swarm algorithm adopts an elite particle set updating mode, so that the exploration and exploitation capabilities of the particle swarm algorithm are balanced, the problems of traditional particle swarm algorithms, such as easy acquisition of bad solutions, low search efficiency and premature convergence, are effectively solved, and the convergence performance of the algorithm is greatly improved.

Description

Technical Field

[0001] This invention belongs to the field of integrated energy system control technology, and more specifically, relates to an optimized scheduling method and system for integrated energy systems based on an improved particle swarm optimization algorithm. Background Technology

[0002] Rational energy management of integrated energy systems is an effective way to reduce the impact of distributed energy fluctuations on the power grid, promote the development and application of renewable energy, alleviate fossil fuel shortages, and reduce carbon emissions. Therefore, designing reasonable and effective optimization and scheduling methods for integrated energy systems is of great significance for accelerating the construction of low-carbon integrated energy systems.

[0003] Currently, there is extensive research on optimal scheduling methods for integrated energy systems. Mainstream methods include mathematical optimization methods such as nonlinear programming, second-order cone programming, and mixed-integer programming, as well as heuristic algorithms such as genetic algorithms and particle swarm optimization. Patent CN115619015A provides an optimal operation method for integrated energy systems that considers user demand response. It uses piecewise linearization to linearize the nonlinear constraints in the objective function and optimal scheduling conditions, and then uses a mixed-integer linear programming algorithm to solve the optimal scheduling model of the integrated energy system. Patent CN112241508B provides an optimal scheduling method for integrated energy systems that considers economic benefits and carbon emissions, and uses a particle swarm optimization algorithm based on a comprehensive crossover operator to solve the optimal scheduling problem of integrated energy systems.

[0004] However, the above methods have the following drawbacks and shortcomings: they neglect considerations of reliability and social impact; and the traditional particle swarm optimization algorithm currently used for solving integrated energy system operation scheduling models is prone to obtaining poor solutions and has low exploration efficiency when dealing with complex and variable integrated energy optimization scheduling problems. In addition, existing improved particle swarm optimization algorithms rely too much on random exploration and cannot effectively balance the maintenance of local optima with the exploration of global optima, which can easily lead to premature convergence and getting trapped in local optima. Further improvements are urgently needed. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to provide a comprehensive energy system optimization scheduling method and system based on an improved particle swarm optimization algorithm, thereby solving the problems of existing comprehensive energy optimization scheduling methods being prone to obtaining bad solutions and having low exploration efficiency.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a comprehensive energy system optimization scheduling method based on an improved particle swarm optimization algorithm, the method comprising:

[0007] S1. Construct a multi-objective optimization model for the integrated energy system, taking the output of generator sets and the number of energy supply devices in the integrated energy system as decision variables, with the goal of minimizing resource consumption while also considering social and reliability considerations;

[0008] S2. An improved particle swarm optimization algorithm is used to solve the multi-objective optimization model of the integrated energy system to obtain the optimal particle swarm, which is the optimal operation scheduling plan of the integrated energy system. The improved particle swarm optimization algorithm adopts an elite particle set update method.

[0009] Preferably, step S2 specifically includes:

[0010] S21. Randomly generate an initial particle swarm according to the aforementioned decision variables, calculate the comprehensive resource consumption, sociality, and reliability evaluation index cs of each particle, and simultaneously establish an elite particle set E. k :

[0011]

[0012] Among them, E k Let c be the set of elite particles in the k-th iteration. e For the threshold coefficient of the elite set, For the historical best information of particle i in the first k rounds, gbest k .s represents the global optimal information for the entire particle swarm in the first k rounds, and .s represents the comprehensive evaluation index of the corresponding particle.

[0013] S22. Calculate the self-awareness coefficient of the particle swarm:

[0014]

[0015] in, Let be the self-awareness coefficient vector at round k, and ne be the total number of particles in the particle swarm. R is the self-awareness coefficient of the j-th particle in round k. k Let R be a nonnegative random matrix of size ne × ne. k Each element in the matrix follows a Gaussian distribution on [0,1]. Let be the fitness vector at round k. Let be the fitness of the j-th particle in round k, where j = 1, 2, ..., ne;

[0016] S23. Calculate the population cognition coefficient of the particle swarm:

[0017]

[0018] in, The population cognition coefficient of the particle swarm. Self-awareness coefficient vector The maximum value in;

[0019] S24. Update the velocity and position information of the particle swarm according to the following formula:

[0020]

[0021]

[0022] in, Let i be the velocity of particle i in round k+1. Let i be the velocity of particle i in wheel k. Let i be the position of particle i in round k+1. Let EM be the position of particle i in wheel k. k For the elite particle set E k The transformed vector has and Same dimension, Inertial weight;

[0023] S25. Update inertia weights

[0024]

[0025] Among them, w min For the minimum inertia weight, w max For the maximum inertia weight, k max Let k be the maximum number of training rounds, and k be the current training round number.

[0026] S26. Compare the fitness of the local optimal particle swarm and the global optimal particle swarm in the current loop. If the fitness of the local optimal particle swarm is better than that of the global optimal particle swarm, then update the current local optimal particle swarm to the global optimal particle swarm; otherwise, do not update.

[0027] S27. Repeat S22 to S26 until the maximum number of training iterations is reached to obtain the optimal particle swarm.

[0028] It should be noted that this invention improves the original particle swarm optimization algorithm by using an elite particle update method. The improved particle swarm optimization algorithm can adaptively update its own self-cognition coefficient and the population cognition coefficient, resulting in better algorithm convergence. At the same time, during the training process, the elite particle set filters out most of the inferior particles, which greatly avoids the situation where the algorithm converges to an inferior solution due to premature convergence.

[0029] Preferably, The calculation formula is as follows:

[0030]

[0031] Preferably, the multi-objective optimization model of the integrated energy system includes: an objective function and constraints;

[0032] The objective function is as follows:

[0033] Min F = F nor -α|P SP (t)+P wT (t)+P DG (t)-P Load (t)-P losses (t)+P BAT (t)-P dump (t)|

[0034] The constraints are as follows:

[0035]

[0036] Where α>0 is the penalty factor, used to control the penalty term for not satisfying power balance, P SP (t) represents the electrical power output by the distributed photovoltaic unit at time t, P wT (t) represents the electrical power output of the wind turbine at time t, P DG (t) represents the electrical power output by the generator at time t, P Load (t) represents the load power at time t, P losses (t) represents the network loss power at time t, P BAT P(t) represents the electrical power output by the distributed energy storage at time t. dump (t) represents the excess electrical energy at time t; N m ∈{N s N w N D N B N l}, This represents the minimum number of corresponding devices. This represents the maximum number of corresponding devices. This is the minimum output of the generator set. This represents the maximum output of the generator set.

[0037] It should be noted that the present invention preferably constructs the objective function in the manner described above, while considering a comprehensive evaluation index that minimizes resource consumption while taking into account social and reliability factors and power balance. Power balance constraints are incorporated into the objective function as a penalty term. When power balance is satisfied, the penalty term is 0; otherwise, the penalty value is adjusted according to the α coefficient. The objective function constructed in the above manner can effectively avoid the impact on the power grid caused by power overruns after scheduling. At the same time, reducing equality constraints is also beneficial for improving the search and update of the particle swarm algorithm.

[0038] Preferably, the comprehensive evaluation index F nor The calculation formula is as follows:

[0039]

[0040] Among them, LCOE max To achieve energy equalization, the maximum resource consumption is required, LPSP max HDI represents the maximum probability of power loss. max λ1 is the maximum value of the Human Development Index, λ2 is the resource consumption weight of the energy equalization coefficient, λ3 is the power supply loss probability weight, and the sum of the three is 1.

[0041] It should be noted that the present invention preferably constructs the comprehensive evaluation index in the above manner, while ensuring that the energy balance resource consumption (LCOE) is minimized, the power supply loss probability (LPSP) index is minimized, and the human development index (HDI) index is maximized.

[0042] Preferably, the formula for calculating the resource cost (LCOE) of energy equalization is as follows:

[0043]

[0044] NPC = (N s + w + D + B + l )*NPC cap +PC rep +PC op )

[0045]

[0046] Among them, P Load () represents the load power at time t, T is the total duration to be optimized, NPC is the difference between the inflow and outflow of resource consumption in the integrated energy system, and C RF N is the capital recovery factor for the integrated energy system. s N represents the number of solar photovoltaic units. w N represents the number of wind turbine units. D N represents the number of distributed generator sets. B N represents the number of distributed energy storage devices. l For the number of inverters, NPC cap NPCs expend resources for capital rep To reset the resources consumed, NPC op The resources consumed for operation and maintenance are ξ, where ξ is the proportionality coefficient and ny is the number of years.

[0047] Preferably, the power supply loss probability LPSP calculation formula is as follows:

[0048]

[0049] Where T is the total duration that needs to be optimized.

[0050] Preferably, the Human Development Index (HDI) is calculated using the following formula:

[0051]

[0052]

[0053]

[0054] in, The maximum surplus coefficient, ε is the maximum load factor. Pop P represents the number of users in an integrated energy system. Load For the total load power, P Dump For excess electrical energy, k1 and k2 are constant coefficients less than 1, and T is the total duration to be optimized.

[0055] To achieve the above objectives, in a second aspect, the present invention provides a comprehensive energy system optimization scheduling system based on an improved particle swarm optimization algorithm, comprising: a processor and a memory; the memory for storing computer execution instructions; and the processor for executing the computer execution instructions such that the method described in the first aspect is executed.

[0056] In summary, the technical solutions conceived by this invention have the following beneficial effects compared with the prior art:

[0057] This invention proposes an optimized scheduling method and system for integrated energy systems based on an improved particle swarm optimization (PSO) algorithm. It constructs a multi-objective optimization model for the integrated energy system, using generator output and the number of energy-supplying devices as decision variables. This model aims to minimize resource consumption while considering social and reliability considerations, thus improving the overall efficiency of the integrated energy system. The improved PSO algorithm is used to solve the multi-objective optimization model, yielding the optimal particle swarm, which represents the optimal operation and scheduling plan for the integrated energy system. The improved PSO algorithm employs an elite particle set update method, balancing the exploration and development capabilities of the algorithm. This effectively solves the problems of traditional PSO algorithms, such as obtaining poor solutions, low search efficiency, and premature convergence, and significantly improves the algorithm's convergence performance. Attached Figure Description

[0058] Figure 1 This is a flowchart of an integrated energy system optimization scheduling method based on an improved particle swarm optimization algorithm provided by the present invention. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0060] like Figure 1 As shown, this invention proposes a comprehensive energy system optimization scheduling method based on an improved particle swarm optimization algorithm, which mainly includes the following steps:

[0061] Step (1): Taking into account resource consumption, technical and social factors, construct a resource consumption, stability and social welfare index function to evaluate the scheduling effect of the integrated energy system;

[0062] Step (2): Based on the index function constructed in step (1), and considering the operational constraints of the integrated energy system, construct a multi-objective optimization model for the operation and scheduling of the integrated energy system;

[0063] Step (3): Use the improved particle swarm optimization algorithm to solve the multi-objective optimization model of the integrated energy system operation scheduling and obtain the optimal operation scheduling plan of the integrated energy system.

[0064] Furthermore, in step (1), considering resource consumption, technological and social factors, a resource consumption, stability and social welfare index function for evaluating the scheduling effect of the integrated energy system is constructed, as follows:

[0065] Step (1-1): Construct a resource consumption index function to evaluate the scheduling effect of the integrated energy system.

[0066] The resource consumption level (LCOE) for energy equalization is used as a resource consumption indicator in the comprehensive energy system evaluation model.

[0067]

[0068] Among them, P Load (t) represents the load power at time t, T represents the total duration to be optimized, NPC represents the difference between the inflow and outflow of resource consumption in the integrated energy system, and C represents the total load power at time t. RF The capital recovery factor for an integrated energy system is defined as follows:

[0069] NPC = (N s +N w +N D +N B +N l )*(NPC cap +NPC rep +NPC op (2)

[0070]

[0071] Where, N s N represents the number of solar photovoltaic units. w N represents the number of wind turbine units. D N represents the number of distributed generator sets. B N represents the number of distributed energy storage devices. l For the number of inverters, NPC cap NPCs expend resources for capital rep To reset the resources consumed, NPC op The resources consumed for operation and maintenance are ξ, where ξ is the proportionality coefficient and ny is the number of years.

[0072] Step (1-2): Construct a reliability index function to evaluate the scheduling effect of the integrated energy system.

[0073] The probability of power loss (LPSP) is used as a reliability indicator in the comprehensive evaluation model of the integrated energy system.

[0074]

[0075] Among them, P SP (t) represents the electrical power output by the distributed photovoltaic unit at time t, P wT (t) represents the electrical power output of the wind turbine at time t, P BAT (t) represents the electrical power output by the distributed energy storage at time t.

[0076] Steps (1-3): Construct a social index function to evaluate the scheduling effect of the integrated energy system.

[0077] Using the Human Development Index (HDI) as a social indicator in the integrated energy system assessment model:

[0078]

[0079] Where k1 and k2 are constant coefficients less than 1, here taken as 0.0978 and 0.0319 respectively. The maximum surplus coefficient, ε is the maximum load factor. Pop P represents the number of users in an integrated energy system. Load For the total load power, P Dump Excess electrical energy is defined as:

[0080]

[0081]

[0082] Among them, P DG(t) represents the electrical power output by the generator at time t.

[0083] Furthermore, in step (2), based on the index function constructed in step (1), and considering the operational constraints of the integrated energy system, a multi-objective optimization model for the operation and scheduling of the integrated energy system is constructed, as follows:

[0084] Step (2-1): Normalize the resource consumption, stability and social welfare index functions for evaluating the scheduling effect of the integrated energy system.

[0085] The energy equilibrium resource consumption (LCOE), power supply loss probability (LPSP), and human development index (HDI) in the integrated energy system assessment model are normalized as follows:

[0086]

[0087] Among them, F nor LCOE is a comprehensive evaluation index with unified dimensions. max To maximize energy balance, LPSP consumes resources. max For the maximum power supply loss probability, HDI max λ represents the maximum value of the Human Development Index. LCOE The resource consumption weight for the energy equalization coefficient, λ LPSP λ represents the probability weight of power supply loss. HDI Weights for the Human Development Index, and satisfying the following:

[0088] λ LCOE +λ LPSP +λ HDI =1 (9)

[0089] Step (2-2): Construct the objective function of the multi-objective operation optimization model of the integrated energy system.

[0090] The objective function of the multi-objective optimization model for the integrated energy system is established as follows:

[0091] Min F = F nor -α|P SP (t)+P wT (t)+P DG (t)-P Load (t)-P losses (t)+P BAT (t)-P dump (t)| (10)

[0092] Where α>0 is the penalty factor, mainly used to control the penalty term for not satisfying power balance, P losses (t) represents the network loss power at time t, and the power balance condition is:

[0093] P SP (t)+P wT (t)+P DG (t)=P Load (t)+P losses (t)-P BAT (t)+P dump (t) (11)

[0094] Step (2-3): Construct the constraints of the multi-objective operation optimization model of the integrated energy system.

[0095] Establish the constraint functions for the multi-objective optimization model of the integrated energy system as follows:

[0096]

[0097] Where, N m ∈{N s N w N D N B N l}, This represents the minimum number of corresponding devices. This represents the maximum number of corresponding devices. This is the minimum output of the generator set. The decision variables for the multi-objective operation optimization model of the integrated energy system are: (The maximum output of the generator unit is given.)

[0098] X = {P} DG (t),N s N w N D N B N l} (13)

[0099] Furthermore, in step (3), the steps of solving the multi-objective optimization model of the integrated energy system using the improved particle swarm optimization algorithm are as follows:

[0100] Step (3-1): Initialize the particle swarm and construct an elite particle set.

[0101] Initialize the particle swarm. According to the decision variables shown in formula (13), randomly generate the initial particle swarm, and calculate the resource consumption of each particle according to formula (14). The calculation formula is as follows:

[0102] cs = F nor -α|P SP (t)+P wT (t)+P DG (t)-P Load (t)-P losses(t)+P BAT (t)-P dump (t)| (14)

[0103] Simultaneously, establish the following elite particle set:

[0104]

[0105] Among them, E k Let c be the set of elite particles in the k-th iteration. e For the threshold coefficient of the elite set, For the historical best information of particle i in the first k rounds, gbest k This represents the global optimal information for the entire particle swarm in the first k rounds, where .cs represents the resource consumption value of each particle.

[0106] Step (3-2): Calculate the self-awareness coefficient.

[0107] The self-awareness coefficient of the particle swarm is calculated using the following formula:

[0108]

[0109] in, Let be the self-awareness coefficient vector at round k, and ne represent the total number of particles in the particle swarm. R is the self-awareness coefficient of the j-th particle in round k. k Let R be a nonnegative random matrix of size ne × ne. k Each element in the matrix follows a Gaussian distribution on [0,1]. Let be the fitness vector at round t. Let be the fitness of the j-th particle in round k, and its calculation formula is:

[0110]

[0111] Step (3-3): Calculate the population cognition coefficient.

[0112] The population cognition coefficient of the particle swarm is calculated using the following formula:

[0113]

[0114] in, The population cognition coefficient of the particle swarm. Represents the vector of self-awareness coefficients The maximum value in.

[0115] Step (3-4): Update the velocity and position information of the particle swarm.

[0116] The velocity and position information of the particle swarm are calculated using the following formula:

[0117]

[0118]

[0119] in, Let i be the velocity of particle i in round k+1. Let i be the velocity of particle i in wheel k. Let i be the position of particle i in round k+1. Let EM be the position of particle i in wheel k. k For the elite particle set E k The transformed vector has and Same dimension, This is the inertial weight.

[0120] Step (3-5): Update the inertia weights.

[0121] Update the inertia weights according to the following formula:

[0122]

[0123] Among them, w min For the minimum inertia weight, w max For the maximum inertia weight, k max is the maximum number of training rounds, and k is the current training round.

[0124] Step (3-6): Update the global optimal particle.

[0125] Compare the fitness of the local optimal particle swarm and the global optimal particle swarm in the current loop. If the fitness of the local optimal particle swarm is better than that of the global optimal particle swarm, then update the current local optimal particle swarm to the global optimal particle swarm; otherwise, do not update it.

[0126] Step (3-7): Repeat steps (3-2) to (3-6) until the maximum number of training iterations is reached to obtain the optimal particle swarm, which is the optimal operation and scheduling plan of the integrated energy system.

[0127] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A comprehensive energy system optimization scheduling method based on an improved particle swarm optimization algorithm, characterized in that, The method includes: S1. Construct a multi-objective optimization model for the integrated energy system, taking the output of generator sets and the number of energy supply devices in the integrated energy system as decision variables, with the goal of minimizing resource consumption while also considering social and reliability considerations; S2. An improved particle swarm optimization algorithm is used to solve the multi-objective optimization model of the integrated energy system to obtain the optimal particle swarm, which is the optimal operation scheduling plan of the integrated energy system. The improved particle swarm optimization algorithm adopts an elite particle set update method. Step S2 specifically includes: S21. Randomly generate an initial particle swarm according to the aforementioned decision variables, and calculate the comprehensive resource consumption, sociality, and reliability evaluation index of each particle. At the same time, an elite particle collection was established. : in, For the first A collection of elite particles in a round of iteration. For the threshold coefficient of the elite set, For particles in front The best historical information in the cycle, For the entire particle swarm in front Global optimal information in the round, This is the comprehensive evaluation index for the corresponding particle; S22. Calculate the self-awareness coefficient of the particle swarm: in, For the first Self-awareness coefficient vector during the round. The total number of particles in the particle swarm. for The first round The self-awareness coefficient of an individual particle. For size A nonnegative random matrix, Each element in the matrix follows a Gaussian distribution on [0,1]. For the first The fitness vector at each round, for The first round The fitness of each particle ; S23. Calculate the population cognition coefficient of the particle swarm: in, The population cognition coefficient of the particle swarm. Self-awareness coefficient vector The maximum value in; S24. Update the velocity and position information of the particle swarm according to the following formula: in, For particles exist Wheel speed, For particles exist Wheel speed, For particles exist The position of the wheel, For particles exist The position of the wheel, For a collection of elite particles The transformed vector has and Same dimension, Inertial weights; S25. Update inertia weights in, For minimum inertia weight, For maximum inertia weight, The maximum number of training rounds, This represents the current training round number; S26. Compare the fitness of the local optimal particle swarm and the global optimal particle swarm in the current loop. If the fitness of the local optimal particle swarm is better than that of the global optimal particle swarm, then update the current local optimal particle swarm to the global optimal particle swarm; otherwise, do not update. S27. Repeat S22~S26 until the maximum number of training iterations is reached to obtain the optimal particle swarm; The calculation formula is as follows: 。 2. The method as described in claim 1, characterized in that, The multi-objective optimization model for the integrated energy system includes: an objective function and constraints; The objective function is as follows: The constraints are as follows: in, This is a comprehensive evaluation index that normalizes energy equalization resource consumption, power supply loss probability, and the Human Development Index. , is a penalty factor used to control the penalty term for not satisfying power balance. For distributed photovoltaic units The electrical power output at all times For wind turbines The electrical power output at all times For the generator in The electrical power output at all times for Load power at any given time for Network loss power at any given time For distributed energy storage The electrical power output at all times for Excess electrical energy at any given moment; , The number of solar photovoltaic units, The number of wind turbine units, The number of distributed generator sets. For the quantity of distributed energy storage, The number of inverters, This represents the minimum number of corresponding devices. This represents the maximum number of corresponding devices. This is the minimum output of the generator set. This represents the maximum output of the generator set.

3. The method as described in claim 2, characterized in that, Comprehensive evaluation index The calculation formula is as follows: in, To achieve energy balance, resources are consumed. To achieve energy balance, the maximum amount of resources consumed. For the probability of power supply loss, This represents the maximum probability of power supply loss. The Human Development Index This represents the maximum value of the Human Development Index. The resource consumption weight for the energy equalization coefficient For power supply loss probability weights, The weights are determined by the Human Development Index, and the sum of the three factors is 1.

4. The method as described in claim 3, characterized in that, Energy equalization consumes resources The calculation formula is as follows: in, for Load power at any given time For the total duration that needs optimization, This is the difference between the inflow and outflow of resource consumption in a comprehensive energy system. The capital recovery factor for a comprehensive energy system. To consume resources for capital To reset the resources consumed, To consume resources for operation and maintenance, This is the proportionality coefficient. The number of years has elapsed.

5. The method as described in claim 3, characterized in that, Power loss probability The calculation formula is as follows: in, This represents the total duration that needs optimization.

6. The method as described in claim 3, characterized in that, Human Development Index The calculation formula is as follows: in, The maximum surplus coefficient, The maximum load factor, The number of users in an integrated energy system, For the total load power, For excess electrical energy, For constant coefficients less than 1, This represents the total duration that needs optimization.

7. A comprehensive energy system optimization scheduling system based on an improved particle swarm optimization algorithm, characterized in that, include: Processor and memory; The memory is used to store computer-executed instructions; The processor is configured to execute the computer execution instructions, such that the method described in any one of claims 1 to 6 is executed.

Images