Electrolytic aluminum load adjusting method based on master-slave game price feedback

Abstract

The invention discloses an electrolytic aluminum load adjusting method based on master-slave game price feedback, relates to a new energy power grid adjusting method, and aims to solve the problem that the safe operation of a power grid is affected due to insufficient new energy consumption caused by the fact that the existing electrolytic aluminum load power adopts fixed instruction and one-way control and cannot be dynamically adjusted along with wind power photovoltaic real-time fluctuation. The method comprises the following steps: acquiring regional power grid side data, electrolytic aluminum user side data and algorithm parameter data; constructing a master-slave game optimization model, and based on algorithm parameter data, carrying out iterative optimization on the double-layer optimization model by adopting a hybrid solving strategy combining a parallel particle swarm algorithm and a solver to obtain an optimal compensation price and an electrolytic aluminum optimal power regulation quantity under a Nash equilibrium condition; and the regional power grid implements excitation according to the optimal compensation price, and the electrolytic aluminum user executes real-time power adjustment according to the optimal power adjustment quantity to complete electrolytic aluminum load power adjustment. The method has the beneficial effects of improving new energy consumption efficiency and guaranteeing power grid safety.

Description

Technical Field

[0001] This invention relates to a method for regulating new energy power grids. Background Technology

[0002] With a high proportion of renewable energy integrated into the power system, wind and solar power generation exhibit significant intermittency, volatility, and randomness. Large-scale grid connection can easily lead to power imbalances, power flow exceeding limits, and wind and solar curtailment, severely restricting renewable energy consumption and the safe and economical operation of the power grid. Electrolytic aluminum loads are high-power, adjustable, and energy-intensive industrial loads. Their electrolytic cells possess short-term power adjustment and temperature buffering characteristics, making them a high-quality, adjustable resource for participating in grid dispatch, smoothing renewable energy fluctuations, and improving consumption levels.

[0003] Currently, the control method for electrolytic aluminum loads to participate in grid regulation mostly adopts a combination of fixed power commands and fixed compensation, which has obvious technical defects:

[0004] 1. Fixed power commands are issued directly by the grid side without taking into account the real-time power of new energy sources, system reserves, line power flow and other operating conditions. This makes it impossible to achieve dynamic matching between load regulation and new energy consumption, resulting in high wind and solar curtailment rates and poor overall system economy.

[0005] 2. The lack of a mechanism for coordinating interests between the power grid and users means that power regulation does not take into account the constraints of electrolytic aluminum production processes, output benefits, electricity costs, and regulation costs. Users are prone to production cuts, efficiency declines, or economic losses, resulting in low willingness to participate in regulation and difficulty in releasing the adjustable potential.

[0006] 3. Existing regulation methods are mostly one-way command control, lacking a decision-making framework of two-way interaction, interest game, and iterative optimization, and cannot form a closed-loop regulation mechanism of "grid pricing guidance - user response power - grid optimization adjustment";

[0007] 4. Existing game-theoretic scheduling methods do not fully incorporate the process characteristics of electrolytic aluminum loads, such as power constraints, temperature constraints, adjustable rate constraints, and maximum number of adjustments, nor do they fully consider system safety constraints such as thermal power unit ramping, power balance, and branch power flow, resulting in insufficient model practicality.

[0008] 5. Existing solution methods mostly employ a single optimization algorithm, which converges slowly, is prone to getting trapped in local optima, and cannot quickly obtain the optimal power and compensation strategy that satisfies Nash equilibrium.

[0009] In summary, existing methods for regulating the load power of electrolytic aluminum cannot simultaneously achieve the four objectives of efficient consumption of new energy, lowest grid operating costs, guaranteed user revenue, and satisfaction of system security constraints. There is an urgent need for a method for regulating the load power of electrolytic aluminum that is based on master-slave game theory, takes into account both grid dispatching leadership and user autonomous response, and implements two-level collaborative optimization, so as to achieve a win-win situation for the grid and electrolytic aluminum users and support the safe, stable and economical operation of a high-proportion new energy power system. Summary of the Invention

[0010] The purpose of this invention is to solve the problem that the existing electrolytic aluminum load power adopts fixed commands and unidirectional control, which cannot dynamically adjust with the real-time fluctuations of wind and solar power, resulting in insufficient absorption of new energy and thus affecting the safe operation of the power grid. This invention proposes an electrolytic aluminum load regulation method based on master-slave game price feedback.

[0011] The electrolytic aluminum load regulation method based on master-slave game price feedback described in this invention includes the following steps:

[0012] Acquire regional power grid data, electrolytic aluminum user-side data, and algorithm parameter data;

[0013] A master-slave game optimization model is constructed, which includes an upper-level optimization model and a lower-level optimization model. The upper-level optimization model takes regional power grid side data as input and outputs compensation price, while the lower-level optimization model takes electrolytic aluminum user side data as input and outputs electrolytic aluminum power adjustment amount.

[0014] Based on the algorithm parameter data, a hybrid solution strategy combining parallel particle swarm optimization and solver is adopted to iteratively optimize the master-slave game optimization model and obtain the optimal compensation price and the optimal power adjustment amount of electrolytic aluminum under the Nash equilibrium condition.

[0015] The regional power grid implements incentives based on the optimal compensation price, and electrolytic aluminum users perform real-time power adjustments according to the optimal power regulation amount to complete the load power regulation of electrolytic aluminum.

[0016] Furthermore, the regional power grid-side data includes new energy power generation forecast data, thermal power unit operating parameter data, power grid operation constraint data, and time-of-use electricity price data; the electrolytic aluminum user-side data includes electrolytic aluminum load base power data, electrolytic cell process data, electrolytic cell temperature data, production and economic data, and compensation price boundary; the algorithm parameter data includes parallel particle swarm inertia weight, first learning factor, second learning factor, third learning factor, local and global balance coefficients, and convergence threshold.

[0017] Furthermore, the objective function expression of the upper-level optimization model is:

[0018]

[0019] in, This indicates the total operating cost of the regional power grid; This indicates the cost of thermal power units in the regional power grid; This indicates the penalty cost for wind curtailment in the regional power grid; This indicates the cost of curtailment penalties for solar power curtailment in the region. This indicates the cost of load adjustment for electrolytic aluminum production. For the running cycle; This is the first power generation cost coefficient for thermal power units; This is the second power generation cost coefficient for thermal power units; This is the third power generation cost coefficient for thermal power units; For thermal power units in Power generation at any given moment; For distributed wind turbine units in Power generation at any given moment; for Predicted wind power generation at any given time; The cost coefficient for wind curtailment penalties for wind turbine units in the regional power grid; For distributed photovoltaic units in Photovoltaic power at all times; for Predicted photovoltaic power generation at any given time; The cost coefficient for curtailment penalties of photovoltaic units in the regional power grid; For electrolytic aluminum load at Response and power adjustment in real time; For the power grid The price for load adjustment compensation of electrolytic aluminum is set at all times.

[0020] Furthermore, the constraints of the upper-level optimization model include regional power grid power balance constraints, thermal power unit power constraints, wind and solar power consumption constraints, branch power flow constraints, and electrolytic aluminum load regulation constraints.

[0021] Furthermore, the electrolytic aluminum load adjustment constraint satisfies:

[0022]

[0023] in, The coefficient before temperature change; To ensure that the electrolytic aluminum load is The rate of temperature change over time; The specific heat capacity of the electrolytic material in the electrolytic cell; The mass of the electrolytic material in the electrolytic cell; This represents the minimum load power for electrolytic aluminum production. This represents the maximum load power of the electrolytic aluminum plant. For electrolytic aluminum in Real-time load power; For electrolytic aluminum load at Adjust power constantly; This represents the minimum load temperature for electrolytic aluminum. This represents the maximum load temperature for electrolytic aluminum. For electrolytic aluminum load at Temperature at any given time; This represents the minimum price for load adjustment compensation in electrolytic aluminum production. This represents the maximum compensation price for load regulation in electrolytic aluminum production. For electrolytic aluminum load at Adjustment and compensation prices at any time.

[0024] Furthermore, the objective function expression of the lower-level optimization model is:

[0025]

[0026] in, Total revenue for electrolytic aluminum users; This represents the maximum total revenue for electrolytic aluminum users. For the production benefits of electrolytic aluminum users; This represents the maximum production revenue for electrolytic aluminum users. Electricity costs for electrolytic aluminum loads; The cost of raw materials in the electrolytic aluminum production process; Profit per unit output of electrolytic aluminum; For electrolytic aluminum load DC side current at any given time; The rated current efficiency for electrolytic aluminum load; The rated temperature for electrolytic aluminum production; For the time-of-use electricity pricing system Electricity price at any given time; The price of consumables used in the electrolytic aluminum production process; In order to be in The quality of consumables consumed at all times.

[0027] Furthermore, the constraints of the lower-level optimization model include the power of the electrolytic aluminum load, the temperature of the electrolytic aluminum load, and the maximum number of adjustments to the electrolytic aluminum load within the operating cycle; wherein, the constraint formula for the maximum number of adjustments to the electrolytic aluminum load within the operating cycle is:

[0028]

[0029] in, For electrolytic aluminum load at The state is constantly adjusted, with 1 indicating adjustment and 0 indicating no adjustment. This represents the maximum number of adjustments to the electrolytic aluminum load within one operating cycle.

[0030] Furthermore, the specific steps of the hybrid solution strategy include:

[0031] Parameter initialization: Based on the decision variable dimension of the upper-level optimization model, initialize the particle swarm size, parallel particle swarm inertia weight, first learning factor, second learning factor and third learning factor, and generate a uniformly distributed initial particle swarm based on feasible region constraints.

[0032] Upper-level optimization model iteration: The upper-level optimization model uses parallel computing to synchronously update the particle velocity and position vector of each particle in the uniformly distributed initial particle swarm, and outputs the initial feasible solution set of electrolytic aluminum load adjustment compensation price;

[0033] Information transmission: The initial feasible solution set of the electrolytic aluminum load adjustment compensation price output by the upper-level optimization model is embedded as the boundary condition into the lower-level optimization model, and the lower-level optimization model calls the solver to solve for the electrolytic aluminum power adjustment amount;

[0034] Adaptive solution evaluation: The objective function of the upper-level optimization model is calculated based on the particle fitness value, and a local optimum detection mechanism is introduced; when it is determined that a particle is trapped in a local extremum, the solution is recalculated, otherwise the current particle state is maintained and iteration continues.

[0035] Convergence determination: The calculation is terminated when the rate of change of the global optimal solution for three consecutive generations is less than the relative error threshold or the maximum number of iterations is reached. The output is the optimal compensation price and the optimal power adjustment amount of electrolytic aluminum that meet the Nash equilibrium condition.

[0036] Furthermore, the specific formula for synchronously updating the particle velocity and position vectors of a uniformly distributed initial particle swarm in parallel computing is as follows:

[0037]

[0038] in, For particle swarm Particles in exist The speed of time; Inertial weights; For particle swarm Particles in exist The speed of time; For particle swarm Particles in Historically best position; For particle swarm Particles in exist The iteration position at any given moment; This is a balance coefficient between local and global factors; This is the historical best for particle swarm optimization. This is the globally optimal historical position. For particle swarm Particles in exist The iteration position at any given moment; As the first learning factor; As the second learning factor; As the third learning factor; The first random number between 0 and 1; The second random number between 0 and 1; It is a third random number between 0 and 1.

[0039] Compared with the prior art, the present invention has the following advantages:

[0040] To improve the efficiency of renewable energy consumption and ensure grid security; to dynamically adjust electrolytic aluminum power in accordance with real-time fluctuations in wind and solar power, accurately match power balance, and significantly reduce wind and solar curtailment rates. To strictly meet system safety constraints such as thermal power unit ramping and branch power flow, to avoid power imbalance and power flow exceeding limits, and to support the stable operation of a high-proportion renewable energy grid.

[0041] Establish a two-way benefit coordination mechanism to enhance user participation; achieve closed-loop regulation of "grid pricing guidance - user response power - grid optimization adjustment" through master-slave game theory, balancing the lowest grid cost with the maximum user benefit. Fully consider the electrolytic aluminum production process and economic constraints to ensure user output, efficiency and profit, and solve the problem of user losses due to regulation and low participation enthusiasm.

[0042] The model is more practical as it closely aligns with the characteristics of the electrolytic aluminum process; it fully incorporates process constraints such as electrolytic aluminum power, temperature, adjustable rate, and maximum number of adjustments, making the adjustment strategy more closely aligned with actual production. It also simultaneously considers various safety constraints on the power grid side, overcoming the limitations of traditional game theory models that are detached from real-world operating conditions.

[0043] The algorithm optimizes solution efficiency and effectiveness, quickly obtaining the optimal strategy. It employs a hybrid strategy combining parallel particle swarm optimization and a solver, introducing local optimum detection and dual convergence criteria, resulting in faster convergence and less susceptibility to local optima. It can rapidly solve for the optimal compensation price and power regulation under Nash equilibrium, meeting the real-time dispatch requirements of the power grid. Attached Figure Description

[0044] Figure 1 This is a flowchart of a method for adjusting the load of electrolytic aluminum based on master-slave game price feedback, as described in Specific Implementation Method 1.

[0045] Figure 2 This is a master-slave game framework diagram in Implementation Method 1;

[0046] Figure 3 This is a flowchart of the hybrid solution strategy in Implementation Method 8;

[0047] Figure 4 This is a schematic diagram showing the power output of each unit in Scheme 1 of Implementation Method 9;

[0048] Figure 5 This is a schematic diagram showing the power output of each unit in Scheme 2 of Implementation Method 9;

[0049] Figure 6 This is a schematic diagram showing the power output of each unit in Scheme 3 of Implementation Method 9;

[0050] Figure 7 This is a schematic diagram showing the power output of each unit in Scheme 4 of Implementation Method 9;

[0051] Figure 8 This is a comparative diagram of the new energy consumption situation under different schemes in Implementation Method Nine;

[0052] Figure 9 This is a schematic diagram illustrating the adjustment of compensation prices under different schemes in Implementation Method Nine. Detailed Implementation

[0053] Specific Implementation Method 1: Combination Figures 1 to 2 This embodiment describes a method for regulating the load of electrolytic aluminum based on master-slave game price feedback, which includes the following steps:

[0054] Acquire regional power grid data, electrolytic aluminum user-side data, and algorithm parameter data;

[0055] A master-slave game optimization model is constructed, which includes an upper-level optimization model and a lower-level optimization model. The upper-level optimization model takes regional power grid side data as input and outputs compensation price, while the lower-level optimization model takes electrolytic aluminum user side data as input and outputs electrolytic aluminum power adjustment amount.

[0056] Based on the algorithm parameter data, a hybrid solution strategy combining parallel particle swarm optimization and solver is adopted to iteratively optimize the master-slave game optimization model and obtain the optimal compensation price and the optimal power adjustment amount of electrolytic aluminum under the Nash equilibrium condition.

[0057] The regional power grid implements incentives based on the optimal compensation price, and electrolytic aluminum users perform real-time power adjustments according to the optimal power regulation amount to complete the load power regulation of electrolytic aluminum.

[0058] In this implementation, a master-slave game optimization model is established to achieve two-way interactive regulation between grid-led pricing and user-initiated responses. This forms a price feedback closed loop, allowing electrolytic aluminum load regulation to dynamically match wind and solar power fluctuations. Ultimately, this achieves four unified goals: renewable energy consumption, grid security, user benefits, and economic optimization.

[0059] Specific Implementation Method Two: This implementation method further defines the electrolytic aluminum load regulation method based on master-slave game price feedback described in Specific Implementation Method Two. In this implementation method, the regional power grid side data includes new energy power generation forecast data, thermal power unit operating parameter data, power grid operation constraint data, and time-of-use electricity price data; the electrolytic aluminum user side data includes electrolytic aluminum load base power data, electrolytic cell process data, electrolytic cell temperature data, production and economic data, and compensation price boundary; the algorithm parameter data includes parallel particle swarm inertia weight, first learning factor, second learning factor, third learning factor, local and global balance coefficients, and convergence threshold.

[0060] In this embodiment, the new energy power generation forecast data includes Forecast value of wind power generation at any time Real-time photovoltaic power generation forecast and operating cycle; thermal power unit operating parameter data, including minimum power value, maximum power value, and operating parameters of the thermal power unit. Rotating reserve capacity at any given time, thermal power units in Downward spinning reserve capacity, upward ramp rate of thermal power units, downward ramp rate of thermal power units, dispatch time interval, and system variables of thermal power units within a given time period; grid operation constraint data includes system... Total load power at any given time, upper limit of power flow for each branch, sensitivity matrix, curtailment penalty cost coefficient for photovoltaic units in the regional power grid, and curtailment penalty cost coefficient for wind turbine units in the regional power grid; time-of-use pricing data includes the time-of-use pricing system in... Electricity price at any given time; basic power data for electrolytic aluminum load includes the electricity price at the time of electrolytic aluminum loading. At any given time, the minimum and maximum load power of electrolytic aluminum, and the load power of electrolytic aluminum at various times. Power is adjusted in real time; electrolytic cell process data includes the specific heat capacity and mass of the electrolytic materials in the cell; electrolytic cell temperature data includes the minimum and maximum electrolytic aluminum load temperatures, and the electrolytic aluminum load at various temperatures. The rated temperature and electrolytic aluminum load at various times during production. Rate of temperature change over time; production and economic data including electrolytic aluminum load. DC side current at any given time, rated current efficiency of electrolytic aluminum load, production profit per unit output of electrolytic aluminum, price of consumables in the electrolytic aluminum production process, and so on. The parameters include the mass of consumables consumed at any given time and the maximum number of adjustments to the electrolytic aluminum load within a scheduling cycle; the compensation price boundary includes the minimum and maximum compensation prices for electrolytic aluminum load adjustments; the convergence threshold includes the relative error threshold and the maximum number of iterations; where the relative error threshold is... .

[0061] In this implementation, the data boundaries for the grid side, user side, and algorithm side are clearly defined, ensuring complete model input. It comprehensively covers key information such as wind and solar power forecasts, thermal power parameters, electrolyzer processes, temperature, and economic costs. This provides standardized and structured data support for subsequent modeling and solving.

[0062] Specific Implementation Method Three: This implementation method further defines the electrolytic aluminum load regulation method based on master-slave game price feedback described in Specific Implementation Method One. In this implementation method, the objective function expression of the upper-level optimization model is:

[0063]

[0064] in, This indicates the total operating cost of the regional power grid; This indicates the cost of thermal power units in the regional power grid; This indicates the penalty cost for wind curtailment in the regional power grid; This indicates the cost of curtailment penalties for solar power curtailment in the region. This indicates the cost of load adjustment for electrolytic aluminum production. For the running cycle; This is the first power generation cost coefficient for thermal power units; This is the second power generation cost coefficient for thermal power units; This is the third power generation cost coefficient for thermal power units; For thermal power units in Power generation at any given moment; For distributed wind turbine units in Power generation at any given moment; for Predicted wind power generation at any given time; The cost coefficient for wind curtailment penalties for wind turbine units in the regional power grid; For distributed photovoltaic units in Photovoltaic power at all times; for Predicted photovoltaic power generation at any given time; The cost coefficient for curtailment penalties of photovoltaic units in the regional power grid; For electrolytic aluminum load at Response and power adjustment in real time; For the power grid The price for load adjustment compensation of electrolytic aluminum is set at all times.

[0065] In this implementation, the upper-level optimization model aims to minimize the total grid cost, while also considering the costs of thermal power, wind and solar curtailment penalties, and load regulation costs. The objective function is clearly quantified and can be directly used for dispatch decisions, improving the overall economic efficiency of the system. The model transforms renewable energy consumption into a penalty cost, forcibly guiding a reduction in wind and solar curtailment.

[0066] Specific Implementation Method Four: This implementation method further defines the electrolytic aluminum load regulation method based on master-slave game price feedback described in Specific Implementation Method Three. In this implementation method, the constraints of the upper-level optimization model include regional power grid power balance constraints, thermal power unit power constraints, wind and solar power consumption constraints, branch power flow constraints, and electrolytic aluminum load regulation constraints.

[0067] In this embodiment, the power balance constraint of the regional power grid is that the regional power grid must maintain real-time power balance at any given time. This balance is achieved through dynamic matching between the power of wind power, photovoltaic and thermal power units on the supply side and the regulating load and conventional load of electrolytic aluminum on the demand side.

[0068]

[0069] in, For the system in Total load power at any given time.

[0070] The power constraints of thermal power units are as follows: The power constraints of thermal power units in the regional power grid include upper and lower limits of unit power, unit ramp rate constraints, and unit spinning reserve capacity constraints.

[0071]

[0072] in, This represents the minimum power output of a thermal power unit. This represents the maximum power output of the thermal power unit. For thermal power units in Upward rotation reserve capacity within a given timeframe; For thermal power units in Downward rotation reserve capacity within a given timeframe; The upward gradient rate of the thermal power unit; The downward ramp rate of the thermal power unit; The scheduling time interval; For thermal power unit system variables, The value range is 0-1;

[0073] The power constraint of wind and solar power integration is that the power of wind and solar generator units is limited by wind and solar resources, and there is a power range constraint on the units.

[0074]

[0075] Branch power flow constraints are defined as follows: Branch power flow constraints reflect the physical transmission bottleneck of the regional power grid, requiring that the transmission power of each line should be strictly limited within its thermal stability limit or the maximum transmission capacity allowed by static safety under any operating condition.

[0076]

[0077] in, This is the sensitivity matrix; This is the power matrix for thermal power units; This is the power matrix of the wind turbine generator; For photovoltaic power units; For the regional power grid load power matrix; This is the matrix representing the maximum power flow value of the line.

[0078] In this implementation, safety constraints such as power balance, thermal power ramping, wind and solar power, and branch power flow are incorporated to ensure reliable grid operation. The constraint system is complete, conforms to actual grid operation rules, and avoids exceeding limits and overload.

[0079] Specific Implementation Method Five: This implementation method further defines the electrolytic aluminum load adjustment method based on master-slave game price feedback described in Specific Implementation Method Four. In this implementation method, the electrolytic aluminum load adjustment constraint satisfies:

[0080]

[0081] in, The coefficient before temperature change; To ensure that the electrolytic aluminum load is The rate of temperature change over time; The specific heat capacity of the electrolytic material in the electrolytic cell; The mass of the electrolytic material in the electrolytic cell; This represents the minimum load power for electrolytic aluminum production. This represents the maximum load power of the electrolytic aluminum plant. For electrolytic aluminum in Real-time load power; For electrolytic aluminum load at Adjust power constantly; This represents the minimum load temperature for electrolytic aluminum. This represents the maximum load temperature for electrolytic aluminum. For electrolytic aluminum load at Temperature at any given time; This represents the minimum price for load adjustment compensation in electrolytic aluminum production. This represents the maximum compensation price for load regulation in electrolytic aluminum production. For electrolytic aluminum load at Adjustment and compensation prices at any time.

[0082] In this embodiment, a power-temperature coupling constraint for electrolytic aluminum is established, which aligns with the thermal buffering characteristics of the electrolytic cell. Upper and lower limits and an average constraint for compensation prices are set to prevent prices from becoming too high or too low, protecting the interests of both parties. The adjustment amount is strongly correlated with temperature changes to avoid damage to the production process due to excessive adjustment.

[0083] Specific Implementation Method Six: This implementation method further defines the electrolytic aluminum load regulation method based on master-slave game price feedback described in Specific Implementation Method Five. In this implementation method, the objective function expression of the lower-level optimization model is:

[0084]

[0085] in, Total revenue for electrolytic aluminum users; This represents the maximum total revenue for electrolytic aluminum users. For the production benefits of electrolytic aluminum users; This represents the maximum production revenue for electrolytic aluminum users. Electricity costs for electrolytic aluminum loads; The cost of raw materials in the electrolytic aluminum production process; Profit per unit output of electrolytic aluminum; For electrolytic aluminum load DC side current at any given time; The rated current efficiency for electrolytic aluminum load; The rated temperature for electrolytic aluminum production; For the time-of-use electricity pricing system Electricity price at any given time; The price of consumables used in the electrolytic aluminum production process; In order to be in The quality of consumables consumed at all times.

[0086] In this implementation, the lower-level optimization model aims to maximize total user revenue while considering production revenue, electricity costs, and raw material costs. It incorporates the electrolytic aluminum current efficiency and temperature effects into the revenue calculation, aligning with actual production efficiency. This truly optimizes from the user's perspective, increasing their willingness to participate in adjustments.

[0087] Specific Implementation Method Seven: This implementation method further defines the electrolytic aluminum load regulation method based on master-slave game price feedback described in Specific Implementation Method Six. In this implementation method, the constraints of the lower-level optimization model include the power of the electrolytic aluminum load, the temperature of the electrolytic aluminum load, and the maximum number of electrolytic aluminum adjustments within the operating cycle; wherein, the constraint formula for the maximum number of electrolytic aluminum adjustments within the operating cycle is:

[0088]

[0089] in, For electrolytic aluminum load at The state is constantly adjusted, with 1 indicating adjustment and 0 indicating no adjustment. This represents the maximum number of adjustments to the electrolytic aluminum load within one operating cycle.

[0090] In this implementation, a maximum number of adjustment cycles is constrained to protect the electrolytic cell equipment and prevent frequent adjustments from damaging its lifespan. The triple constraints of power, temperature, and number of adjustments make the user-side model more closely resemble industrial settings, ensuring that adjustment behavior is sustainable and can be executed long-term.

[0091] Detailed Implementation Method 8: Combination Figure 3 This embodiment further defines the electrolytic aluminum load adjustment method based on master-slave game price feedback described in Specific Embodiment Seven. In this embodiment, the specific steps of the hybrid solution strategy include:

[0092] Parameter initialization: Based on the decision variable dimension of the upper-level optimization model, initialize the particle swarm size, parallel particle swarm inertia weight, first learning factor, second learning factor and third learning factor, and generate a uniformly distributed initial particle swarm based on feasible region constraints.

[0093] Upper-level optimization model iteration: The upper-level optimization model uses parallel computing to synchronously update the particle velocity and position vector of each particle in the uniformly distributed initial particle swarm, and outputs the initial feasible solution set of electrolytic aluminum load adjustment compensation price;

[0094] Information transmission: The initial feasible solution set of the electrolytic aluminum load adjustment compensation price output by the upper-level optimization model is embedded as the boundary condition into the lower-level optimization model, and the lower-level optimization model calls the solver to solve for the electrolytic aluminum power adjustment amount;

[0095] Adaptive solution evaluation: The objective function of the upper-level optimization model is calculated based on the particle fitness value, and a local optimum detection mechanism is introduced; when it is determined that a particle is trapped in a local extremum, the solution is recalculated, otherwise the current particle state is maintained and iteration continues.

[0096] Convergence determination: The calculation is terminated when the rate of change of the global optimal solution for three consecutive generations is less than the relative error threshold or the maximum number of iterations is reached. The output is the optimal compensation price and the optimal power adjustment amount of electrolytic aluminum that meet the Nash equilibrium condition.

[0097] In this implementation, a hybrid strategy of parallel particle swarm optimization and solver is adopted, resulting in faster solution speed and higher accuracy. A local optimum detection mechanism is added to avoid getting trapped in local optima. A dual convergence criterion is used to ensure stable and reliable results, achieving Nash equilibrium.

[0098] Detailed Implementation Method Nine: Combination Figures 4 to 9This embodiment further defines the electrolytic aluminum load adjustment method based on master-slave game price feedback described in Specific Embodiment Eight. In this embodiment, the specific formula for parallel calculation and synchronous updating of the particle velocity and position vectors of the uniformly distributed initial particle swarm is as follows:

[0099]

[0100] in, For particle swarm Particles in exist The speed of time; Inertial weights; For particle swarm Particles in exist The speed of time; For particle swarm Particles in Historically best position; For particle swarm Particles in exist The iteration position at any given moment; This is a balance coefficient between local and global factors; This is the historical best for particle swarm optimization. This is the globally optimal historical position. For particle swarm Particles in exist The iteration position at any given moment; As the first learning factor; As the second learning factor; As the third learning factor; The first random number between 0 and 1; The second random number between 0 and 1; It is a third random number between 0 and 1.

[0101] In this embodiment, the calculation example is as follows:

[0102] Scheme 1 considers the scenario where electrolytic aluminum load does not participate in the regional power grid's renewable energy consumption; Scheme 2 considers the scenario where, under a fixed compensation price, the regional power grid defaults to allowing electrolytic aluminum load to participate in regulation entirely according to the grid's wishes, with electrolytic aluminum load participating in regional power grid regulation and using the load elasticity model proposed in this paper; Scheme 3 considers the economic benefits of electrolytic aluminum users under a fixed compensation price; and Scheme 4 considers the scenario where, under a dynamic compensation price mechanism, the electrolytic aluminum load elasticity model proposed in this paper is used to participate in the regional power grid's renewable energy consumption. The regional power grid unit power and demand response compensation price under these four different scenarios are as follows: Figure 4 , Figure 5 , Figure 6 and Figure 7 As shown;

[0103] Combination Figure 4 , Figure 5 , Figure 6 and Figure 7 It can be seen that the power generation side of the regional power grid includes thermal power units, wind power units, and photovoltaic units, while the load side consists of electrolytic aluminum loads and other basic loads. During operation, the power generation and load power are balanced in real time. When the electrolytic aluminum load does not participate in the regional power grid regulation, the system operating status is as follows: Figure 4 As shown in the diagram. In this situation, the regional power grid relies solely on thermal power units for grid regulation and renewable energy consumption. However, due to the limited regulation capacity of thermal power units, severe wind and solar power curtailment occurs. When the electrolytic aluminum load participates in regulation entirely according to grid dispatch instructions under a fixed compensation price, the system operating status is as follows: Figure 5 As shown in the diagram. Compared to the other three scenarios, this mode achieves the optimal level of renewable energy absorption in the regional power grid, fully releasing the regulation potential of the electrolytic aluminum load. However, this scheme only considers the operating cost-effectiveness of the power grid side and does not consider the economic benefits of electrolytic aluminum industrial users, resulting in the inability to effectively guarantee user-side benefits. Under the fixed compensation price mechanism, when considering the benefits of electrolytic aluminum industrial users, the system operating state is as follows: Figure 6 As shown in the figure. Analysis results indicate that, due to the need to consider user economic benefits, the regulation capacity of the electrolytic aluminum load is somewhat limited, resulting in both its actual regulation volume and renewable energy consumption falling between Scheme 1 (load does not participate in grid regulation) and Scheme 2 (load is entirely regulated according to grid demand). Under the dynamic compensation price mechanism, when the electrolytic aluminum load participates in regional grid regulation while also considering its own economic benefits, the system operating state is as follows: Figure 7 As shown in the figure. Research shows that, considering both user economic constraints, Scheme 4 and Scheme 3 have similar regulation effects, with their actual load regulation and renewable energy consumption falling between Scheme 1 (load does not participate in regulation) and Scheme 2 (load is entirely regulated according to grid demand). However, due to the time-of-use pricing mechanism for electrolytic aluminum, the fixed compensation price used in Scheme 3 cannot cover users' electricity costs and production losses during the 12-15 and 19-22 periods, leading to reduced participation in regulation and consequently affecting renewable energy consumption efficiency. In contrast, the dynamic compensation mechanism implemented in Scheme 4 can effectively coordinate the interests of the grid and users through price elasticity, ensuring the economic benefits of electrolytic aluminum users while reducing system operating costs. The renewable energy consumption situation of the regional power grid and the established electrolytic aluminum demand response compensation prices under the four schemes are shown in the figure. Figure 8 and Figure 9 As shown.

[0104] This study yielded four schemes for electrolytic aluminum compensation pricing, corresponding power conditions of various generating units, and electrolytic aluminum load adjustment. After finally solving for the system-side cost and the comprehensive revenue of electrolytic aluminum, it was concluded that the pricing strategy proposed in this implementation method is superior, and the optimal power for electrolytic aluminum load is achieved.

[0105] In this implementation, a formula for updating the velocity and position of a parallel particle swarm is given, and the algorithm can be directly implemented through programming. Comparison of four schemes verifies that dynamic pricing and master-slave game theory yield the best results. This demonstrates that our method comprehensively outperforms the fixed-instruction and fixed-price schemes in terms of absorption capacity, system cost, and user benefits.

[0106] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for regulating the load of electrolytic aluminum based on price feedback in a master-slave game, characterized in that, Includes the following steps: Acquire regional power grid data, electrolytic aluminum user-side data, and algorithm parameter data; A master-slave game optimization model is constructed, which includes an upper-level optimization model and a lower-level optimization model. The upper-level optimization model takes regional power grid side data as input and outputs compensation price, while the lower-level optimization model takes electrolytic aluminum user side data as input and outputs electrolytic aluminum power adjustment amount. Based on the algorithm parameter data, a hybrid solution strategy combining parallel particle swarm optimization and solver is adopted to iteratively optimize the master-slave game optimization model and obtain the optimal compensation price and the optimal power adjustment amount of electrolytic aluminum under the Nash equilibrium condition. The regional power grid implements incentives based on the optimal compensation price, and electrolytic aluminum users perform real-time power adjustments according to the optimal power regulation amount to complete the load power regulation of electrolytic aluminum.

2. The electrolytic aluminum load adjustment method based on master-slave game price feedback according to claim 1, characterized in that, The regional power grid side data includes new energy power generation forecast data, thermal power unit operating parameter data, power grid operation constraint data, and time-of-use electricity price data; the electrolytic aluminum user side data includes electrolytic aluminum load base power data, electrolytic cell process data, electrolytic cell temperature data, production and economic data, and compensation price boundary; the algorithm parameter data includes parallel particle swarm inertia weight, first learning factor, second learning factor, third learning factor, local and global balance coefficients, and convergence threshold.

3. The electrolytic aluminum load adjustment method based on master-slave game price feedback according to claim 2, characterized in that, The objective function expression of the upper-level optimization model is: in, This indicates the total operating cost of the regional power grid; This indicates the cost of thermal power units in the regional power grid; This indicates the penalty cost for wind curtailment in the regional power grid; This indicates the cost of curtailment penalties for solar power curtailment in the region. This indicates the cost of load adjustment for electrolytic aluminum production. For the running cycle; This is the first power generation cost coefficient for thermal power units; This is the second power generation cost coefficient for thermal power units; This is the third power generation cost coefficient for thermal power units; For thermal power units in Power generation at any given moment; For distributed wind turbine units in Power generation at any given moment; for Predicted wind power generation at any given time; The cost coefficient for wind curtailment penalties for wind turbine units in the regional power grid; For distributed photovoltaic units in Photovoltaic power at all times; for Predicted photovoltaic power generation at any given time; The cost coefficient for curtailment penalties of photovoltaic units in the regional power grid; For electrolytic aluminum load at Response and power adjustment in real time; For the power grid The price for load adjustment compensation of electrolytic aluminum is set at all times.

4. The electrolytic aluminum load adjustment method based on master-slave game price feedback according to claim 3, characterized in that, The constraints of the upper-level optimization model include regional power grid power balance constraints, thermal power unit power constraints, wind and solar power consumption constraints, branch power flow constraints, and electrolytic aluminum load regulation constraints.

5. The electrolytic aluminum load adjustment method based on master-slave game price feedback according to claim 4, characterized in that, The electrolytic aluminum load regulation constraint satisfies: in, The coefficient before temperature change; To ensure that the electrolytic aluminum load is The rate of temperature change over time; The specific heat capacity of the electrolytic material in the electrolytic cell; The mass of the electrolytic material in the electrolytic cell; This represents the minimum load power for electrolytic aluminum production. This represents the maximum load power of the electrolytic aluminum plant. For electrolytic aluminum in Real-time load power; For electrolytic aluminum load at Adjust power constantly; This represents the minimum load temperature for electrolytic aluminum. This represents the maximum load temperature for electrolytic aluminum. For electrolytic aluminum load at Temperature at any given time; This represents the minimum price for load adjustment compensation in electrolytic aluminum production. This represents the maximum compensation price for load regulation in electrolytic aluminum production. For electrolytic aluminum load at Adjustment and compensation prices at any time.

6. The electrolytic aluminum load adjustment method based on master-slave game price feedback according to claim 5, characterized in that, The objective function expression of the lower-level optimization model is: in, Total revenue for electrolytic aluminum users; This represents the maximum total revenue for electrolytic aluminum users. For the production benefits of electrolytic aluminum users; This represents the maximum production revenue for electrolytic aluminum users. Electricity costs for electrolytic aluminum loads; The cost of raw materials in the electrolytic aluminum production process; Profit per unit output of electrolytic aluminum; For electrolytic aluminum load DC side current at any given time; The rated current efficiency for electrolytic aluminum load; The rated temperature for electrolytic aluminum production; For the time-of-use electricity pricing system Electricity price at any given time; The price of consumables used in the electrolytic aluminum production process; In order to be in The quality of consumables consumed at all times.

7. The electrolytic aluminum load regulation method based on master-slave game price feedback according to claim 6, characterized in that, The constraints of the lower-level optimization model include the power of the electrolytic aluminum load, the temperature of the electrolytic aluminum load, and the maximum number of adjustments to the electrolytic aluminum load within the operating cycle; wherein, the constraint formula for the maximum number of adjustments to the electrolytic aluminum load within the operating cycle is: in, For electrolytic aluminum load at The state is constantly adjusted, with 1 indicating adjustment and 0 indicating no adjustment. This represents the maximum number of adjustments to the electrolytic aluminum load within one operating cycle.

8. The electrolytic aluminum load adjustment method based on master-slave game price feedback according to claim 7, characterized in that, The specific steps of the hybrid solution strategy include: Parameter initialization: Based on the decision variable dimension of the upper-level optimization model, initialize the particle swarm size, parallel particle swarm inertia weight, first learning factor, second learning factor and third learning factor, and generate a uniformly distributed initial particle swarm based on feasible region constraints. Upper-level optimization model iteration: The upper-level optimization model uses parallel computing to synchronously update the particle velocity and position vector of each particle in the uniformly distributed initial particle swarm, and outputs the initial feasible solution set of electrolytic aluminum load adjustment compensation price; Information transmission: The initial feasible solution set of the electrolytic aluminum load adjustment compensation price output by the upper-level optimization model is embedded as the boundary condition into the lower-level optimization model, and the lower-level optimization model calls the solver to solve for the electrolytic aluminum power adjustment amount; Adaptive solution evaluation: The objective function of the upper-level optimization model is calculated based on the particle fitness value, and a local optimum detection mechanism is introduced; when it is determined that a particle is trapped in a local extremum, the solution is recalculated, otherwise the current particle state is maintained and iteration continues. Convergence determination: The calculation is terminated when the rate of change of the global optimal solution for three consecutive generations is less than the relative error threshold or the maximum number of iterations is reached. The output is the optimal compensation price and the optimal power adjustment amount of electrolytic aluminum that meet the Nash equilibrium condition.

9. The electrolytic aluminum load regulation method based on master-slave game price feedback according to claim 8, characterized in that, The specific formula for synchronously updating the particle velocity and position vectors of a uniformly distributed initial particle swarm in parallel computing is as follows: in, For particle swarm Particles in exist The speed of time; Inertial weights; For particle swarm Particles in exist The speed of time; For particle swarm Particles in Historically best position; For particle swarm Particles in exist The iteration position at any given moment; This is a balance coefficient between local and global factors; This is the historical best for particle swarm optimization. This is the globally optimal historical position. For particle swarm Particles in exist The iteration position at any given moment; As the first learning factor; As the second learning factor; As the third learning factor; The first random number between 0 and 1; The second random number between 0 and 1; It is a third random number between 0 and 1.

Images