A power distribution network optimization scheduling method based on improved frost and ice optimization algorithm containing biomass energy

Abstract

The application discloses a kind of biomass energy-based distribution network optimization scheduling method based on improved frost ice optimization algorithm, comprising the following steps: first step: constructing biomass power generation-based distribution network system model: second step: the multi-objective optimization model of distribution network is constructed, and the minimum total system cost is established as objective function, the total system cost is composed of distribution network operation cost and environmental protection cost weighted or combined;And establish power balance constraint, each generator output and climbing constraint, tie line transmission power constraint and energy storage device constraint;Third step, model solution: frost ice optimization algorithm is used to solve the planning model.The present application breaks through the passive situation of traditional distribution network highly dependent on diesel and energy storage, and constructs a biomass energy-based distribution network model with biogas units as controllable power sources, which can effectively coordinate multiple distributed power sources and effectively suppress the random fluctuations of wind and light power output.

Description

Technical Field

[0001] This invention discloses an optimized scheduling method for distribution networks containing biomass energy based on an improved frost and ice optimization algorithm, belonging to the field of power system technology. Background Technology

[0002] With the increasing depletion of fossil fuels and the intensification of environmental pollution, building a clean, low-carbon, safe, and efficient modern energy system has become a global consensus. In particular, the goal of "carbon peaking and carbon neutrality" has accelerated the penetration of renewable energy sources, such as wind and solar power, into the power system. Simultaneously, in the context of county-level development, with the gradual advancement of new rural construction, livestock farming is showing a trend towards large-scale development, and biogas power generation technology is continuously developing and improving. As a key link connecting energy production and consumption, the distribution network is gradually transforming from a traditional passive receiving-end network to an active distribution network that can absorb a high proportion of distributed power sources. However, wind and solar power generation have significant intermittency and randomness, and large-scale direct grid connection poses a severe challenge to the stable operation and economic dispatch of the distribution network.

[0003] For the optimal scheduling problem of distribution networks, scholars both domestically and internationally have conducted research from various aspects, including model construction and solution algorithms. Some existing technologies are based on edge computing, constructing a multi-regional energy autonomy framework and employing a proximal policy optimization (PPO) algorithm based on deep reinforcement learning to achieve optimal scheduling of the distribution network through cloud-edge collaboration, with the goal of minimizing operational scheduling costs. Other existing technologies use the total expenditure cost of the Active Distribution Network (ADN) as the objective function, comprehensively considering various constraints, constructing an ADN optimal scheduling model that incorporates distributed energy storage, and using an improved sparrow search algorithm to solve the ADN optimal scheduling model, obtaining the lowest total expenditure cost for the ADN. Existing technologies have proposed a two-layer model for distribution network security and economic dispatch that considers wind, solar, diesel, and energy storage. The upper layer constructs a multi-objective dispatch model that minimizes total operating cost, network loss, and voltage stability index under constraints such as power flow and microgrid interaction, and solves the model using an improved third-generation non-dominated sorting genetic algorithm. The lower layer constructs a dispatch model with the goal of minimizing the microgrid's own operating cost under operational constraints such as energy storage and distributed generation, and solves the model using an improved particle swarm optimization algorithm. These studies focus on minimizing operating cost or overall cost, without considering environmental costs, thus enabling the distribution network to operate under highly environmentally friendly conditions, which is significant for achieving the "dual-carbon" goal.

[0004] Existing power distribution network optimization scheduling technologies and solution algorithms suffer from the following bottlenecks and shortcomings that urgently need to be addressed in practical applications: (1) Scheduling model level: The potential of physical scheduling of biomass energy has not been fully explored, and there is a lack of in-depth environmental economic considerations.

[0005] Most existing scheduling models are only oriented towards minimizing "operating costs" or "overall economic costs," failing to deeply couple the treatment costs of pollutants such as SO2, NOx, and CO2 into the objective function. This leads to the system sacrificing environmental benefits in pursuit of low costs. Furthermore, existing research rarely considers biomass energy (biogas) as a highly flexible and controllable distributed power source for real-time global scheduling of the distribution network. It also fails to fully consider the nonlinear ramping physical constraints of biogas generator sets due to fermentation and gas supply systems. Consequently, when wind and solar resources fluctuate, the system becomes highly dependent on purchasing electricity from the main grid or diesel generator sets, failing to truly achieve low-carbon autonomy for county-level distribution networks.

[0006] (2) Optimization algorithm level: Traditional algorithms have serious problems of output exceeding limits and local optima when dealing with high-dimensional multi-constraint power physical scheduling.

[0007] Distribution network time-series scheduling is essentially an NP-hard problem involving hundreds of decision variables (e.g., 144 dimensions in the example), and is subject to strict Kirchhoff's laws and physical limits of equipment. Existing standard RIME algorithms and conventional heuristic algorithms suffer from three major drawbacks when applied to such scenarios: ① Population initialization uses purely random generation. The blindly generated initial scheduling scheme often violates power flow constraints and power balance in the distribution network, producing a large number of physically infeasible solutions, resulting in extremely low optimization efficiency in the early stages of the algorithm; ② The soft frost search mechanism blindly relies on global factors and lacks guidance from the relative distance between the current particle and the optimal solution. During the approximation of the optimal solution, the search step size cannot adaptively adjust according to the physical ramping capability of the generating units, leading to frequent and drastic fluctuations in unit output. This easily triggers ramping limits for biogas generators and diesel generators, resulting in forced corrections and system oscillations; ③ The hard frost puncture mechanism uses a direct replacement strategy. While this accelerates initial convergence, it rapidly destroys population diversity. When facing complex physical scheduling nodes such as the switching of energy storage device charging and discharging states and the reversal of tie-line power, it is prone to getting trapped in local solutions and cannot obtain a truly globally optimal scheduling strategy. Summary of the Invention

[0008] The technical problem this invention aims to solve is: first, to address the high cost and high pollution associated with peak-shaving resources in high-proportion renewable energy distribution networks. By constructing a multi-objective distribution network scheduling model incorporating biomass energy, the dispatchability of biogas power generation can be utilized to mitigate wind and solar power fluctuations, achieving a win-win situation for both environmental and economic benefits at the lowest overall cost.

[0009] Second, this paper addresses the problems of low optimization accuracy and susceptibility to local deadlock in traditional heuristic algorithms when solving time-series scheduling problems in high-dimensional distribution networks. The RIME algorithm is improved by introducing a four-fold strategy: Tent chaotic mapping, relative distance guidance, random reset, and Levy flight. Furthermore, the algorithm's search step size and position update operations are coupled at a low level with the power flow constraints of the distribution network and the ramp-up constraints of biomass power units. This significantly enhances the algorithm's global exploration and local development capabilities, ensuring that the output scheduling scheme can be truly implemented, reducing costs and increasing efficiency.

[0010] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: A method for optimal scheduling of distribution networks containing biomass energy based on an improved frost and ice optimization algorithm includes the following steps: Step 1: Construct a power distribution network system model that includes biomass power generation: Step 2: Construct a multi-objective optimization model for the power distribution network, and establish the objective function as minimizing the total system cost. The total system cost is composed of the weighted average or combined operating cost and environmental protection cost of the power distribution network. Also, establish power balance constraints, output and ramping constraints of each generator set, power transmission constraints of tie lines, and constraints of energy storage devices. The third step is model solving: The frost-ice optimization algorithm is used to solve this planning model, specifically as follows: (1) Initialization of the Frost Ice Population: ; In the formula: x ij For Frost Ice Population R Frost particles in the ice; (2) Soft Cream Search Mechanism: ; in: ; ; ; In the formula: R new ij For the updated version i The first frost body j The new location of the frost particle; R best.j For Frost Ice Population R The best frost body in the middle j The position of each particle; r 1 θ is a random number in the range (−1,1), which, together with cosθ, controls the diffusion direction of the particles, and θ changes with the number of iterations; β is an environmental factor, and its mathematical model is a step function; hThe adhesion degree is a random number within (0,1); U bij , L bij These are the upper and lower bounds of the particle escape space, respectively; r 2 Let be a random number within the range (0,1); E The adhesion coefficient, r 2 and E This jointly controls whether the position of the frost particles is updated; t is the current iteration number; T is the maximum iteration number; This is the floor function; w This is used to control the number of segments in the step function; (3) Mechanism of hard frost puncture: ; In the formula: r 3 A random number within the range (−1, 1); F normr ( S i ) represents the normalized value of the current frost body fitness, indicating the th i The probability of a frost body being selected; (4) Improve the greedy selection mechanism: By incorporating an active greedy selection mechanism into population updates, the fitness value of the updated or previous values ​​is selected.

[0011] Step 2.1 Establish objective functions: This includes three objective functions: operating cost, environmental cost, and total cost. The relationship between the three objective functions is as follows: ; In the formula, Z The total cost of the distribution network; f 1 Operating costs; f 2 For environmental protection costs; Operating costs are: ; ; In the formula, C t Grid , C t DE , C t BGS , C t BES They are respectivelyt Total cost of interaction between the distribution network and the main grid during the time period, total operating cost of diesel generator sets, total operating cost of biogas generator sets, and total operating cost of energy storage devices; P buy (t), P sell (t) are respectively t The power purchased and sold by the distribution network from the main grid at all times; c buy (t), c sell (t) are respectively t The electricity purchase price and electricity sales price of the distribution network to the main grid at all times; Environmental protection costs of power distribution networks: ; ; In the formula: C t Grid.EN The cost of treating pollutants for the main power grid; c Grid,k The first generated by the main power grid k Emission coefficients of pollutants; C Grid,k It is the first step in handling the main power grid. k Cost coefficients for pollutants of this type.

[0012] Step 2.2 Constraints: (1) Power balance constraint: ; In the formula, P Grid (t) is t The active power exchanged with the main grid during a given time period; a positive value indicates power injected from the main grid. P L (t) is t Load values ​​for a given time period; (2) Output constraints of biomass power generator sets: ; (3) Output constraints of diesel generator sets: ; (4) Transmission power constraints of tie lines: ; (5) Constraints of energy storage devices: ; In the formula: , These are the upper and lower limits of the output of the biomass power generator set; P BGS (t-1) represents the output of the biomass power generator unit during time period t-1; , These are the upper and lower limits of the diesel generator set's output; P DE (t-1) represents the output of the diesel generator set during the time period t-1; r BGS , r DE These are the maximum ramping power limits for biomass power generator sets and diesel power generator sets, respectively. , These are the upper and lower limits of the transmission power of the tie line, respectively. , These are the upper and lower limits of the energy storage device; SOC min (t), SOC max (t) represents the upper and lower limits of the energy storage capacity during time period t.

[0013] The improvements to the RIME algorithm include the following: Step 3.2.1: Set parameters and use Tent chaotic mapping to generate the initial population location; Step 3.2.2: Calculate the fitness value of all frost particles and record the current global best position; Step 3.2.3: Traverse each particle in the population and update its position through the soft frost search phase or the hard frost piercing phase; Step 3.2.4: Before the end of each iteration, perform a Levy flight perturbation on the current population to help the algorithm escape local optima; Step 3.2.5: Check if the updated particles have gone out of bounds, recalculate the fitness, and update the global optimal solution; Step 3.2.6: If the maximum number of iterations is reached, output the optimal scheduling scheme and the corresponding cost value; otherwise, return to step 3.2.2 to continue iterating.

[0014] In step 3.2.1, the Tent chaotic map initializes the population, and its mathematical expression is: ; In the formula: z k+1 For the first k+1 Iterative chaotic variables; z k For the first k The chaotic variable in the iteration; α is the chaotic control parameter; The expression for initializing frost particles is: ; In the formula: LB j For the first j The lower bound of the search space for dimensional variables; UB j For the first j The upper bound of the search space for dimensional variables.

[0015] In step 3.2.3, relative distance information is used to enable the search adjustment step size of the unit output to be adaptively adjusted according to the distance between the particle and the optimal solution. The expression for the distance guiding factor is: ; In the formula: rand s This is a random number with a value in the range [-1, 1]; round(.) is the rounding function. When the current scheduling scheme is far from the optimal scheme, the search step size is increased to fully utilize the biogas generator's climbing ability to accelerate the approach; when the distance is close, the algorithm adaptively decreases the search step size for fine-tuning. The improved soft frost ice update formula is: ; ; In the formula: rand A random number in the range [0,1]. Meanwhile, the hard frost mechanism was refactored, introducing a global random reset operation, which gives the scheduling scheme the ability to randomly reset output under the premise of satisfying constraints. The improved update logic is as follows: ; In the formula: r m , r 4 These are random numbers that follow a uniform distribution in the range [0,1]. p m This is the threshold for the probability of mutation.

[0016] In step 3.2.4, at the end of each iteration, a Levy flight mechanism is introduced, with the Levy step size formula as follows: ; In the formula: is the step size; c Take 1.5; u , v Random numbers that satisfy a normal distribution u~N(0, σ u 2 ) , v~N(0,1) ; Standard deviation s u The calculation formula is: ; Based on the generated step size S, new candidate solutions R' are generated using the distribution information of the current population. best : ; In the formula: f This is the step size control factor; R mean This represents the average position of all particles in the current iteration.

[0017] The beneficial effects of this invention are as follows: This application establishes a multi-objective optimization model for a distribution network that incorporates biomass energy, wind, solar, and energy storage, as well as traditional generating units. It proposes an IRIME algorithm that integrates the Tent chaotic mapping, distance guiding factor, and Levy flight strategy for solving the model. Compared with existing technologies, this application offers the following advantages: (1) Deeply activate the potential of biomass energy to achieve a dual cost reduction in environmental protection and economy.

[0018] This invention breaks through the passive situation of traditional power distribution networks that are highly dependent on diesel and energy storage. By using biogas generators as a controllable power source, the constructed power distribution network model containing biomass energy can effectively coordinate multiple distributed power sources and effectively smooth out random fluctuations in wind and solar power output.

[0019] Furthermore, the dispatch results demonstrate that, with the combination of off-peak electricity price energy storage charging and peak-peak discharge, biogas power generation, with its low pollution control cost advantage, effectively replaces high-polluting diesel power generation and grid-purchased electricity. Compared to traditional dispatch schemes, this scheme significantly reduces system environmental costs by 48.87% while minimizing environmental costs.

[0020] (2) The IRIME algorithm breaks through the optimization bottleneck under high-dimensional physical constraints and avoids scheduling instructions exceeding the limit. The proposed IRIME algorithm is not simply a collection of mathematical operators, but rather achieves a deep coupling between algorithm improvement and power physics constraints. It utilizes the Tent chaotic mapping to ensure the physical feasibility of the initial solution; it creatively controls the search step size through a distance guiding factor, a mechanism specifically designed to address the nonlinear ramping constraints caused by biomass gas valve regulation. By coupling the algorithm step size with the underlying mathematical understanding of biomass' physical ramping capability, it completely solves the system power oscillation problem frequently triggered by conventional algorithms when scheduling high-proportion biomass power distribution networks, perfectly matching the physical ramping constraints of biomass power units and avoiding the frequent output fluctuation exceeding limits problem in traditional algorithms; it breaks the deadlock state of energy storage and tie-line scheduling through random reset and Levy flight. Experiments show that IRIME reduces the total cost of solving this system by 14.27% compared to the traditional RIME, and the convergence curve shows that it still has the ability to continuously explore better solutions in the later stages of iteration, ensuring the extremely high stability and executability of the scheduling scheme in the actual power distribution network physical environment. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the power distribution network framework. Figure 2 To improve the RIME flowchart; Figure 3 This is a comparison of the convergence curves of the two algorithms. Detailed Implementation

[0022] A method for optimal scheduling of distribution networks containing biomass energy based on an improved frost and ice optimization algorithm includes the following steps: Step 1: Construct a power distribution network system model that includes biomass power generation: (1) Constructing the power distribution network system structure: A typical active distribution network framework mainly includes a higher-level power grid (Grid), new energy generating units represented by wind turbines (WT) and photovoltaics (PV), traditional generating units mainly composed of diesel engines (DE) and battery energy storage (BES), electricity loads, and an active distribution network dispatch and control center.

[0023] With the integration of biogas power generation technology, the optimal scheduling of the power distribution network needs to be adjusted. This application establishes an optimal scheduling model for the power distribution network composed of wind power, photovoltaic, diesel generator sets, biomass power generator sets, and energy storage devices. The power distribution network framework is as follows: Figure 1 As shown.

[0024] This application establishes an optimized scheduling model for a distribution network containing biomass energy, solves the scheduling model using algorithms before and after the improvement, provides a reasonable unit output plan, and verifies the effect of the algorithm improvement.

[0025] (2) Power generation model: ① Wind power generation model: The expression for the wind speed-power characteristic curve of a wind turbine (WT) is as follows: ; In the formula: P WT (t) It is the output power of WT; P r This is the rated power of WT; V(t) for t Wind speed at that time of day; V ci , V r , V co These are the cut-in wind speed, rated wind speed, and cut-out wind speed, respectively.

[0026] ② Photovoltaic power generation model: The output power model for photovoltaic power generation is shown below: ; In the formula: P PT (t) for t The output power of photovoltaics during a given period; R The photovoltaic output power under standard test conditions; q This is the derating factor for photovoltaic power, typically 0.8; I T (t) for t Actual solar radiation intensity during the time period; I STC Solar radiation intensity under standard test conditions; α P The temperature coefficient of the PV solar panel; T c (t) for t PV cell temperature over a period of time; T stc The temperature of the PV cell under standard testing conditions.

[0027] ③ Diesel generator model Diesel engines are a common type of fuel generator, which incurs fuel costs, maintenance costs, and pollutant treatment costs during operation.

[0028] ; In the formula: C DE.COM (t), C DE.F (t), C t DE.EN They are respectively t The operating and maintenance costs, fuel costs, and pollutant treatment costs of diesel engines at all times; P DE (t) It is a diesel engine t Electricity generation at any given moment; K DE.OM It is the operating and maintenance cost coefficient of the diesel engine; c de,k It is the first generated by the operation of the diesel engine k Emission coefficients of pollutants; C de,k It is the first step in handling diesel generator sets. k Cost coefficients for pollutants of this type; α、β、h The coefficients for the diesel engine are 0.00011, 0.1801, and 6, respectively.

[0029] ④ Biomass power generation system: Biogas generator sets (BGS) generate electricity using biomass gas produced through anaerobic fermentation. They are characterized by high dispatchability and rapid start-up, making them an important controllable distributed power source in the system. Their power generation efficiency model is shown below: ; In the formula: or BGS (t) is t The power generation efficiency of the biogas generator unit at all times; P BGS (t) is t The output power of the biogas generator unit at any given time; P rate This is the rated power of the biogas generator unit; a bg , b bg , c bg , d bg The coefficients are the efficiency fitting coefficients.

[0030] The operating cost of a biogas generator set mainly consists of fuel costs, maintenance costs, and pollutant treatment costs, which can be expressed mathematically as follows: ; In the formula: C BGS.F (t), C BGS.OM (t), C t BGS.EN They are respectively t The fuel cost, operation and maintenance cost, and pollutant treatment cost of diesel engines at all times; C gas The unit cost of biogas; LHV bio The lower heating value of biogas is usually affected by the methane concentration; K BGS.OM This is the operation and maintenance cost coefficient; c bg,k The first generation of biogas power generation k Emission coefficients of pollutants; C bg,k It is the first biogas generator set to be processed k Cost coefficients for pollutants of this type.

[0031] ⑤ Energy storage system: This application adopts a battery model, and the mathematical model is expressed as follows: ; In the formula: SOC(t) for t The remaining capacity of the battery during the specified period; P bes (t) for t The charging and discharging power of the battery is displayed at any time; a positive value indicates charging, and a negative value indicates discharging. or + , or - These represent the charge and discharge efficiencies, respectively.

[0032] The second step is to construct a multi-objective optimization model for the distribution network, establishing the objective function as minimizing the total system cost, which is composed of a weighted average or combined distribution network operating cost and environmental protection cost; and to establish constraints on power balance, output and ramping of each generator set, power transmission through tie lines, and energy storage devices.

[0033] Step 2.1 Objective function: This application considers three objective functions: operating cost, environmental cost, and total cost. The relationship between the three objective functions is as follows: ; In the formula, Z The total cost of the distribution network; f 1 Operating costs; f 2 For environmental protection costs.

[0034] Operating costs are: ; ; In the formula, C t Grid , C t DE , C t BGS , C t BES They are respectively t Total cost of interaction between the distribution network and the main grid during the time period, total operating cost of diesel generator sets, total operating cost of biogas generator sets, and total operating cost of energy storage devices; P buy (t), P sell (t) are respectively t The power purchased and sold by the distribution network from the main grid at all times; c buy (t), c sell (t) are respectively t The price at which the distribution network purchases electricity from the main grid and the price at which it sells electricity.

[0035] Environmental protection costs of power distribution networks ; ; In the formula: C t Grid.EN The cost of treating pollutants for the main power grid; c Grid,k The first generated by the main power grid k Emission coefficients of pollutants; C Grid,k It is the first step in handling the main power grid. k Cost coefficients for pollutants of this type.

[0036] 2.2 Constraints: (1) Power balance constraint ; In the formula, P Grid (t) is t The active power exchanged with the main grid during a given time period; a positive value indicates power injected from the main grid. P L (t) is tLoad values ​​for a given time period.

[0037] (2) Output constraints of biomass power generator sets: ; (3) Output constraints of diesel generator sets: ; (4) Transmission power constraints of tie lines: ; (5) Constraints of energy storage devices: ; In the formula: , These are the upper and lower limits of the output of the biomass power generator set; P BGS (t-1) represents the output of the biomass power generator unit during time period t-1; , These are the upper and lower limits of the diesel generator set's output; P DE (t-1) represents the output of the diesel generator set during the time period t-1; r BGS , r DE These are the maximum ramping power limits for biomass power generator sets and diesel power generator sets, respectively. , These are the upper and lower limits of the transmission power of the tie line, respectively. , These are the upper and lower limits of the energy storage device; SOC min (t), SOC max (t) represents the upper and lower limits of the energy storage capacity during time period t.

[0038] The third step is model solving: Distribution network optimization scheduling is essentially a high-dimensional nonlinear multi-constraint programming problem. Therefore, this application introduces an improved strategy based on the RIME algorithm, proposing an IRIME algorithm with stronger optimization performance to solve this planning model.

[0039] 3.1 Frost and Ice Optimization Algorithm (1) Initialization of the Frost Ice Population ; In the formula: x ij For Frost Ice Population R Frost particles in the middle.

[0040] (2) Soft Cream Search Mechanism ; in: ; ; ; In the formula: R new ij For the updated version i The first frost body j The new location of the frost particle; R best.j For Frost Ice Population R The best frost body in the middle j The position of each particle; r 1 θ is a random number in the range (−1,1), which, together with cosθ, controls the diffusion direction of the particles, and θ changes with the number of iterations; β is an environmental factor, and its mathematical model is a step function; h The adhesion degree is a random number within (0,1); U bij , L bij These are the upper and lower bounds of the particle escape space, respectively; r 2 Let be a random number within the range (0,1); E The adhesion coefficient, r 2 and E This jointly controls whether the position of the frost particles is updated; t is the current iteration number; T is the maximum iteration number; This is the floor function; w This is used to control the number of segments in the step function.

[0041] (3) Mechanism of hard frost puncture: ; In the formula: r 3 A random number within the range (−1, 1); F normr ( S i ) represents the normalized value of the current frost body fitness, indicating the th i The probability of a frost body being selected.

[0042] (4) Improve the greedy selection mechanism An improved greedy selection mechanism is introduced into population updates to enhance exploration efficiency by selecting the best fitness value, either the updated value or the value before the update. This operation can thus ensure that the population evolves towards a better outcome in each iteration.

[0043] 3.2 Improved RIME Algorithm While the standard frost-ice optimization algorithm possesses strong search capabilities, it still suffers from problems such as uneven initial population distribution, susceptibility to local optima, and insufficient convergence accuracy in the later stages when dealing with high-dimensional multi-objective optimization problems. To address these limitations, this application proposes an improved frost-ice optimization algorithm (IRIME), which introduces a Tent chaotic mapping, a distance guiding factor, a random reset mechanism, and a Levy flight strategy. Furthermore, it deeply integrates the mathematical mechanisms with the physical characteristics of the power distribution network. The specific improvement strategies are as follows: (1) Initialization of the population based on the Tent chaotic mapping of the physical feasible region of the distribution network: In the standard RIME algorithm, population initialization is typically performed using random generation, which can easily lead to uneven distribution of initial solutions in the search space, affecting the algorithm's ergodicity. To improve the quality of the initial population, this application introduces the Tent chaotic map for population initialization. The Tent map has good uniformity and ergodicity, and its mathematical expression is: ; In the formula: z k+1 For the first k+1 Iterative chaotic variables; z k For the first k The chaotic variable in the iteration; α is the chaotic control parameter.

[0044] Using chaotic sequences for population initialization instead of a random number generator ensures that the generated initial frost particles (i.e., the power output schemes of each unit) uniformly and broadly cover the physically feasible region that satisfies the power balance of the distribution network and the upper and lower limits of unit output, thus avoiding invalid searches. The expression for initializing frost particles is: ; In the formula: LB j For the first j The lower bound of the search space for dimensional variables; UB j For the first j Upper bound of the search space for dimensional variables; (2) Relative distance guiding soft frost search mechanism considering the climbing constraint of biogas generator unit In the soft-frost phase of standard RIME, the particle's movement step size is primarily determined by environmental factors and the global search range. However, equipment such as biogas generator sets have strict physical ramp-up rate limitations, and blindly searching with large step sizes can lead to scheduling command failures. The improved strategy proposed in this paper introduces relative distance information, enabling the search adjustment step size for unit output to be adaptively adjusted based on the distance between the particle and the optimal solution. The expression for the distance guiding factor is: ; In the formula: rand s This is a random number with a value in the range [-1, 1]; round(.) is the rounding function. When the current scheduling scheme is far from the optimal scheme, the search step size is increased to fully utilize the biogas generator's climbing ability to accelerate its approach; when the distance is close, the algorithm adaptively decreases the search step size for fine-tuning, thereby effectively avoiding the physical problem of generator output exceeding limits due to sudden changes in scheduling instructions. The improved soft-frost-ice update formula is: ; ; In the formula: rand A random number in the range [0, 1].

[0045] (3) Global random reset hard frost puncture mechanism to prevent local deadlock of unit output In the standard RIME hard frost puncture mechanism, particles are directly replaced with the position of the current optimal solution. In power dispatching, this direct replacement easily leads to premature convergence of the charging and discharging strategies of energy storage systems or the start-up and shutdown strategies of biomass power units, resulting in local dispatching deadlock. This application reconstructs the hard frost ice mechanism by introducing a global random reset operation, giving the dispatching scheme the ability to randomly reset its output while satisfying constraints. The improved update logic is as follows: ; In the formula: r m , r 4 These are random numbers that follow a uniform distribution in the range [0,1]. p m The mutation probability threshold is introduced to allow the unit output scheme to be randomly reset within the physical feasible domain when facing complex physical scheduling nodes such as the switching of energy storage equipment charging and discharging states and the reversal of tie line power, thereby avoiding the homogenization of scheduling strategies.

[0046] (4) Large-step perturbation of Levy flight for communication line interaction and biomass energy output At the end of each iteration, a Levy flight mechanism is introduced. This mechanism specifically addresses the large-step positional perturbation of biomass generator output and interconnected power when trapped in local optima. Utilizing the long-tail characteristic of the Levy distribution, the algorithm is given the ability to make long-distance jumps during search stagnation periods, mapping the mathematical characteristic of alternating long and short step sizes to wide-range adjustment and fine-tuning of physical generator output. This activates the environmentally friendly substitution effect of biomass energy under the combined charging and discharging of low-peak periods, thereby increasing the probability of finding the global optimum. The Levy step size formula is: ; In the formula: is the step size; c Take 1.5; u , v Random numbers that satisfy a normal distribution u~N(0, σ u 2 ) , v~N(0,1) .

[0047] Standard deviation s u The calculation formula is:

[0048] Based on the generated step size S, new candidate solutions R' are generated using the distribution information of the current population. best : ; In the formula: f This is the step size control factor; R mean This represents the average position of all particles in the current iteration.

[0049] The flowchart of the improved frost ice optimization algorithm (IRIME) is as follows: Figure 2 As shown, the specific steps are as follows: Step 3.2.1: Set parameters and use Tent chaotic mapping to generate the initial population location; Step 3.2.2: Calculate the fitness value of all frost particles and record the current global best position; Step 3.2.3: Traverse each particle in the population and update its position through the soft frost search phase or the hard frost piercing phase; Step 3.2.4: Before the end of each iteration, perform a Levy flight perturbation on the current population to help the algorithm escape local optima; Step 3.2.5: Check if the updated particles have gone out of bounds, recalculate the fitness, and update the global optimal solution; Step 3.2.6: If the maximum number of iterations is reached, output the optimal scheduling scheme and the corresponding cost value; otherwise, return to step 3.2.2 to continue iterating.

[0050] In summary, this invention updates particle positions based on an improved soft frost search mechanism or a hard frost puncture mechanism. During the soft frost search phase, the relative distance between the current particle's represented scheduling scheme and the globally optimal scheduling scheme is calculated. This relative distance is then mathematically correlated with the physical ramping power limits of the biomass generator set and the diesel generator set. The search step size for generator set output is adaptively adjusted. When the distance is greater, the step size is increased to fully utilize the physical ramping potential of the biomass generator set and accelerate convergence. When the distance is less, the step size is decreased for local refinement. This ensures that the generator set output always meets the physical ramping constraints of the distribution network during the process of approaching the optimal solution, thus avoiding violations of physical ramping constraints due to sudden changes in scheduling commands. During the hard frost puncture phase, a global random reset operation constrained by the physical feasible domain is introduced, giving particles the possibility of randomly resetting the unit output state. When the probability condition is met, the particle is no longer directly replaced by the current optimal solution, but is instead given the possibility of randomly resetting the unit output state in the global domain that satisfies the power flow of the distribution network and the extreme value constraints of equipment output. This is to maintain the diversity of the population in the process of solving high-dimensional multi-constraint problems and prevent deadlock and premature convergence of scheduling strategies at complex physical nodes such as energy storage charging and discharging state switching or tie line power reversal.

[0051] A Levy flight strategy is introduced to escape local optima. At the end of each iteration, position perturbations are performed using alternating long and short step sizes generated by the Levy distribution. This involves resetting the output of biomass generators and the interactive power of tie lines trapped in local optima with large step sizes, generating new candidate scheduling schemes. Iterative optimization is then performed, and the results are output. The updated particles are checked to ensure they meet the distribution network constraints, and the global optimal solution is updated. When the maximum number of iterations is reached, the minimum total system cost and the corresponding optimal output scheduling scheme for each generator in the distribution network are output. This optimal output scheduling scheme is then converted into actual control commands and sent to the underlying controllers of the corresponding wind turbines, photovoltaic generators, diesel generators, biomass generators, and energy storage devices to execute real-time physical scheduling of the distribution network.

[0052] To address the issue of biomass generator output and tie-line interaction power easily getting trapped in local optima during distribution network dispatch, the long-tail characteristic of the Levy distribution is used to generate search step sizes that alternate between long and short. The long step size enables the biomass generator output to overcome local extreme value barriers, while the short step size fine-tunes the current output scheme. This activates the environmentally friendly substitution effect of biomass energy under the combination of charging and discharging during off-peak periods during the search stagnation period.

[0053] Example overview and parameter configuration: The specific parameters of the generating units in the power distribution network system of this application are shown in Table 1, and the pollutant emission coefficients, costs, and energy storage parameters are shown in Tables 2 and 3.

[0054] Table 1 Unit Parameters

[0055] Optimize algorithm parameter settings: population size 50, maximum number of iterations 500, variable dimension 144; Tent chaos coefficient. α =0.7, soft cream step factor w =5, Hard Frost Reset Probability p m =0.1, perturbation step size factor φ= 0.01.

[0056] Table 2 Pollutant Emission Coefficients and Costs

[0057] Table 3 Energy Storage Parameters

[0058] Table 4 Time-of-use electricity prices

[0059] Dispatch Result Analysis: Based on the constructed distribution network model containing biomass energy, this application simulated and solved the output allocation strategy of each generator unit and energy storage in the distribution network under three different optimization objectives: minimizing total system cost, minimizing environmental cost, and minimizing operating cost. Experimental results show that the dispatch method of this invention can effectively coordinate various energy sources, and its specific operating characteristics are as follows: (1) Precise peak shaving and valley filling of energy storage devices: Regardless of the optimization target selected, the energy storage devices in the system have played a stable auxiliary regulation role. The dispatch strategy accurately executes the operation of charging during the low electricity price period and discharging during the peak period. Through cross-time period energy transfer, the economic cost and environmental cost are reduced simultaneously.

[0060] (2) Multi-source coordination ensures power supply reliability: Given the intermittent nature of photovoltaic and wind power, their total power generation cannot meet the rigid load demand of the system during certain periods. The scheduling model of this application can make dynamic decisions in real time. By combining the purchase of electricity from the main grid with the control of the output of distributed power sources (biogas and diesel generators), multi-source complementarity is formed, which effectively ensures the power balance of the distribution network around the clock.

[0061] (3) The environmental substitution effect of biomass energy is significant: In-depth analysis of the scheduling results shows that when the system takes "minimizing environmental costs" as its driving objective, the scheduling strategy actively and significantly increases the output power of biogas generator sets (BGS), making its output share significantly higher than that of diesel generator sets (DE) and the main grid's purchased power. Combined with the emission characteristics data of each unit, the sulfur dioxide (SO2) and nitrogen oxides (NOx) produced by BGS...x The emissions from biomass energy are far lower than those from the power grid and the main grid, while the unit treatment costs for these two types of high-polluting sources are very high. This dispatch result proves that connecting biomass energy as a core controllable power source to the active distribution network has environmental cost advantages.

[0062] Analysis of the improved RIME algorithm: The RIME algorithm and the IRIME algorithm were used to solve the problem respectively. Table 5 shows the comparison of the results of the two algorithms.

[0063] Table 5 Comparison of RIME and IRIME

[0064] As shown in Table 5, the improved RIME algorithm outperforms the traditional RIME algorithm in all three objective functions, reducing costs by 14.27%, 48.87%, and 14.3%, respectively. Therefore, the improved RIME algorithm exhibits better performance than the traditional RIME algorithm. A comparison of the convergence curves of the two algorithms is provided below. Figure 3 As shown.

[0065] from Figure 3 As can be seen, the convergence curve of the IRIME algorithm is generally below that of RIME throughout the entire iteration process, and the final total cost is significantly lower. This indicates that although the RIME algorithm can also converge, it tends to plateau earlier at higher costs and is prone to getting trapped in local optima; while the improved IRIME algorithm can continue to explore and find better solutions in the later stages of iteration, demonstrating its better global exploration ability and higher optimization accuracy. Therefore, the IRIME algorithm proposed in this application is superior to the RIME algorithm in reducing system costs.

[0066] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

Claims

1. A method for optimized scheduling of a distribution network containing biomass energy based on an improved frost and ice optimization algorithm, characterized in that: Includes the following steps: Step 1: Construct a power distribution network system model that includes biomass power generation: Step 2: Construct a multi-objective optimization model for the power distribution network, and establish the objective function as minimizing the total system cost. The total system cost is composed of the weighted average or combined operating cost and environmental protection cost of the power distribution network. Also, establish power balance constraints, output and ramping constraints of each generator set, power transmission constraints of tie lines, and constraints of energy storage devices. The third step is model solving: The frost-ice optimization algorithm is used to solve this planning model, specifically as follows: (1) Initialization of the Frost Ice Population: ； In the formula: x ij For Frost Ice Population R Frost particles in the ice; (2) Soft Cream Search Mechanism: ； in: ； ； ； In the formula: R new ij For the updated version i The first frost body j The new location of the frost particle; R best.j For Frost Ice Population R The best frost body in the middle j The position of each particle; r 1 θ is a random number in the range (−1,1), which, together with cosθ, controls the diffusion direction of the particles, and θ changes with the number of iterations; β is an environmental factor, and its mathematical model is a step function; h The adhesion degree is a random number within (0,1); U bij , L bij These are the upper and lower bounds of the particle escape space, respectively; r 2 Let be a random number within the range (0,1); E The adhesion coefficient, r 2 and E This jointly controls whether the position of the frost particles is updated; t is the current iteration number; T is the maximum iteration number; This is the floor function; w This is used to control the number of segments in the step function; (3) Mechanism of hard frost puncture: ； In the formula: r 3 A random number within the range (−1, 1); F normr ( S i ) represents the normalized value of the current frost body fitness, indicating the th i The probability of a frost body being selected; (4) Improve the greedy selection mechanism: By incorporating an active greedy selection mechanism into population updates, the fitness value of the updated or previous values ​​is selected.

2. The method for optimized scheduling of a distribution network containing biomass energy based on an improved frost and ice optimization algorithm according to claim 1, characterized in that: Step 2.1 Establish objective functions: This includes three objective functions: operating cost, environmental cost, and total cost. The relationship between the three objective functions is as follows: ； In the formula, Z The total cost of the distribution network; f 1 Operating costs; f 2 For environmental protection costs; Operating costs are: ； ； In the formula, C t Grid , C t DE , C t BGS , C t BES They are respectively t Total cost of interaction between the distribution network and the main grid during the time period, total operating cost of diesel generator sets, total operating cost of biogas generator sets, and total operating cost of energy storage devices; P buy (t), P sell (t) are respectively t The power purchased and sold by the distribution network from the main grid at all times; c buy (t), c sell (t) are respectively t The electricity purchase price and electricity sales price of the distribution network to the main grid at all times; Environmental protection costs of power distribution networks: ； ； In the formula: C t Grid.EN The cost of treating pollutants for the main power grid; γ Grid,k The first generated by the main power grid k Emission coefficients of pollutants; C Grid,k It is the first step in handling the main power grid. k Cost coefficients for pollutants of this type; Step 2.2 Constraints: (1) Power balance constraint: ； In the formula, P Grid (t) is t The active power exchanged with the main grid during a given time period; a positive value indicates power injected from the main grid. P L (t) is t Load values ​​for a given time period; (2) Output constraints of biomass power generator sets: ； (3) Output constraints of diesel generator sets: ； (4) Transmission power constraints of tie lines: ； (5) Constraints of energy storage devices: ； In the formula: , These are the upper and lower limits of the output of the biomass power generator set; P BGS (t-1) represents the output of the biomass power generator unit during time period t-1; , These are the upper and lower limits of the diesel generator set's output; P DE (t-1) represents the output of the diesel generator set during the time period t-1; r BGS , r DE These are the maximum ramping power limits for biomass power generator sets and diesel power generator sets, respectively. , These are the upper and lower limits of the transmission power of the tie line, respectively. , These are the upper and lower limits of the energy storage device; SOC min (t), SOC max (t) represents the upper and lower limits of the energy storage capacity during time period t.

3. The method for optimized scheduling of a distribution network containing biomass energy based on an improved frost and ice optimization algorithm according to claim 1, characterized in that: The improvements to the RIME algorithm include the following: Step 3.2.1: Set parameters and use Tent chaotic mapping to generate the initial population location; Step 3.2.2: Calculate the fitness value of all frost particles and record the current global best position; Step 3.2.3: Traverse each particle in the population and update its position through the soft frost search phase or the hard frost piercing phase; Step 3.2.4: Before the end of each iteration, perform a Levy flight perturbation on the current population to help the algorithm escape local optima; Step 3.2.5: Check if the updated particles have gone out of bounds, recalculate the fitness, and update the global optimal solution; Step 3.2.6: If the maximum number of iterations is reached, output the optimal scheduling scheme and the corresponding cost value; otherwise, return to step 3.2.2 to continue iterating.

4. The method for optimized scheduling of a distribution network containing biomass energy based on an improved frost and ice optimization algorithm according to claim 3, characterized in that: In step 3.2.1, the Tent chaotic map initializes the population, and its mathematical expression is: ； In the formula: z k+1 For the first k+1 Iterative chaotic variables; z k For the first k The chaotic variable in the iteration; α is the chaotic control parameter; The expression for initializing frost particles is: ； In the formula: LB j For the first j The lower bound of the search space for a dimensional variable; UB j For the first j The upper bound of the search space for dimensional variables.

5. A method for optimizing the scheduling of a distribution network containing biomass energy based on an improved frost and ice optimization algorithm according to claim 3, characterized in that: In step 3.2.3, relative distance information is used to enable the search adjustment step size of the unit output to be adaptively adjusted according to the distance between the particle and the optimal solution. The expression for the distance guiding factor is: ； In the formula: rand s This is a random number with a value in the range [-1, 1]; round(.) is the rounding function. When the current scheduling scheme is far from the optimal scheme, the search step size is increased to fully utilize the biogas generator's climbing ability to accelerate the approach; when the distance is close, the algorithm adaptively decreases the search step size for fine-tuning. The improved soft frost ice update formula is: ； ； In the formula: rand A random number in the range [0,1]. Meanwhile, the hard frost mechanism was refactored, introducing a global random reset operation, which gives the scheduling scheme the ability to randomly reset output under the premise of satisfying constraints. The improved update logic is as follows: ； In the formula: r m , r 4 These are random numbers that follow a uniform distribution in the range [0,1]. p m This is the threshold for the probability of mutation.

6. The method for optimized scheduling of a distribution network containing biomass energy based on an improved frost and ice optimization algorithm according to claim 3, characterized in that: In step 3.2.4, at the end of each iteration, a Levy flight mechanism is introduced, with the Levy step size formula as follows: ； In the formula: is the step size; γ Take 1.5; u , v Random numbers that satisfy a normal distribution u~N(0, σ u 2 ) , v~N(0,1) ; Standard deviation s u The calculation formula is: ； Based on the generated step size S, new candidate solutions R' are generated using the distribution information of the current population. best : ； In the formula: f This is the step size control factor; R mean This represents the average position of all particles in the current iteration.

Images