A Method for Air Conditioning Parameter Identification and Modeling Based on a Multi-Mechanism Fusion Gray Wolf Optimization Algorithm

Abstract

This invention relates to the field of air conditioning load modeling and parameter identification technology, specifically a method for air conditioning parameter identification and modeling based on a multi-mechanism fusion gray wolf optimization algorithm. The method first collects air conditioning operation data and performs preprocessing and normalization. Then, based on the traditional gray wolf algorithm, it integrates a composite chaotic initialization strategy, a cross-optimization strategy, and a Cauchy inverse cumulative distribution variation strategy to construct a multi-mechanism fusion optimization algorithm. Next, a first-order equivalent thermal parameter model is established, and the algorithm is used to identify the equivalent thermal resistance R and equivalent heat capacity C parameters, with the goal of minimizing the mean square error between the predicted and measured temperatures. Finally, based on the identified parameters, air conditioning load aggregation modeling is performed, and the impact of the diversity of set temperature and initial indoor temperature on aggregated power fluctuations is analyzed. Through algorithm improvement and comprehensive aggregation characteristic analysis, theoretical and technical support is provided for the flexible scheduling of air conditioning loads in modern power systems.

Description

Technical Field

[0001] This invention relates to the field of air conditioning load modeling and parameter identification technology, specifically to an air conditioning parameter identification and modeling method based on a multi-mechanism fusion gray wolf optimization algorithm. Background Technology

[0002] With the acceleration of global urbanization and the improvement of living standards, air conditioning load accounts for more than 50% of building energy consumption, making it a key demand response resource for maintaining power grid supply and demand balance and frequency regulation. Accurately identifying the thermodynamic parameters of air conditioning systems and establishing precise aggregated load models are of great significance for power grid dispatching, demand response, and virtual power plant operation.

[0003] Existing methods for identifying air conditioning load parameters face several technical shortcomings: First, traditional optimization algorithms such as particle swarm optimization and genetic algorithms are prone to getting trapped in local optima when dealing with the nonlinear characteristics of air conditioning loads, resulting in low convergence accuracy and poor initial population quality. Second, existing equivalent thermal parameter models mostly use typical parameter values, which differ significantly from actual conditions. Furthermore, while higher-order models offer higher theoretical accuracy, they suffer from high computational complexity and poor real-time performance. Third, existing aggregated modeling research neglects the combined impact of user-set temperatures and the diversity of initial indoor temperatures on aggregated power fluctuations, lacking a quantitative evaluation mechanism. These limitations lead to insufficient accuracy in air conditioning load modeling, affecting the accuracy of power grid dispatching. Summary of the Invention

[0004] This invention provides a method for air conditioning parameter identification and modeling based on a multi-mechanism fusion gray wolf optimization algorithm. Through algorithm improvement and comprehensive aggregation characteristic analysis, it provides theoretical and technical support for the flexible scheduling of air conditioning loads in modern power systems.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for air conditioning parameter identification and modeling based on a multi-mechanism fusion gray wolf optimization algorithm includes:

[0007] S100: Collects indoor temperature, outdoor temperature and cooling capacity data during the operation of the air conditioning system, performs data preprocessing, and obtains a normalized dataset.

[0008] S200: Based on the traditional gray wolf optimization algorithm, a multi-mechanism fusion gray wolf optimization algorithm is constructed by integrating a composite chaotic initialization strategy, a cross-optimization strategy, and a Cauchy inverse cumulative distribution mutation strategy.

[0009] S300: Based on the dataset, establish a first-order equivalent thermal parameter model describing the thermodynamic characteristics of the air conditioning system, set the equivalent thermal resistance R and equivalent heat capacity C as parameters to be identified, and use the multi-mechanism fusion gray wolf optimization algorithm to identify the RC parameters. With the goal of minimizing the mean square error between the predicted indoor temperature and the measured indoor temperature, the optimal parameter values ​​are obtained through iterative search.

[0010] S400: Based on the identified optimal parameter values, perform air conditioning load aggregation modeling, analyze the impact of the diversity of set temperature and initial indoor temperature on the aggregation power fluctuation, and output the air conditioning load aggregation characteristic analysis results.

[0011] As a preferred embodiment of the present invention, the step of obtaining the normalized dataset includes:

[0012] The system collects indoor and outdoor temperature data during air conditioner operation using a temperature sensor, and collects air conditioner operating current data using a current sensor and converts it into cooling capacity data.

[0013] The sliding window mean method was used to average the collected data within a time window.

[0014] Box plots were used to identify and remove outliers from the dataset.

[0015] The processed data were normalized using the min-max normalization method to obtain normalized datasets of indoor temperature, outdoor temperature, and cooling capacity.

[0016] As a preferred embodiment of the present invention, the composite chaos initialization strategy includes:

[0017] A composite chaotic system is constructed using Tent mapping, Logistic mapping, and Cosine mapping, and the appropriate chaotic mapping is selected for iteration based on the individual position value.

[0018] The ergodicity and randomness of chaotic mapping are used to generate an initial population that is uniformly distributed and highly diverse.

[0019] The fitness of the generated chaotic initial population is pre-evaluated, and individuals with low fitness are removed to reduce redundant computation in subsequent iterations.

[0020] As a preferred embodiment of the present invention, the cross-sectional optimization strategy includes:

[0021] The horizontal crossover operation randomly pairs individuals within the population and performs arithmetic crossover operations sequentially on all dimensions of the paired individuals, generating new offspring individuals through information exchange between individuals.

[0022] The longitudinal crossover operation performs an arithmetic crossover operation between two different dimensions for each individual in the population, promoting the stagnant dimension to escape premature convergence.

[0023] After the crossover operation is completed, a competitive selection mechanism is adopted. The fitness values ​​are compared to determine whether to adopt a new solution, ensuring that a high-quality solution replaces the parent individual.

[0024] As a preferred embodiment of the present invention, the Cauchy inverse cumulative distribution variation strategy includes:

[0025] Establish the cumulative distribution function of the standard Cauchy distribution, and use the heavy-tailed property of the Cauchy distribution to determine the probability distribution characteristics of the mutation operation;

[0026] The inverse transformation method is used to convert uniformly distributed random numbers into variable-length random numbers that follow a Cauchy distribution;

[0027] In the later stages of algorithm iteration, Cauchy distribution variation perturbation is introduced into the population individuals to enhance the algorithm's ability to escape local optima.

[0028] As a preferred embodiment of the present invention, the step of establishing a first-order equivalent thermal parameter model describing the thermodynamic characteristics of the air conditioning system, and setting the equivalent thermal resistance R and equivalent heat capacity C as parameters to be identified, includes:

[0029] The air conditioning system is simplified into a first-order equivalent thermal parameter model, ignoring the temperature difference between the inner and outer walls and the temperature difference between the air and the wall surface.

[0030] Based on the principle of energy balance, a differential equation relationship is established between indoor temperature change, indoor-outdoor temperature difference, and air conditioning cooling capacity;

[0031] The differential equation is discretized using the forward Euler method to obtain a discretized indoor temperature prediction equation suitable for numerical solution.

[0032] Determine the physical meaning of equivalent thermal resistance R, which characterizes the resistance to heat transfer between indoor and outdoor environments, and equivalent heat capacity C, which characterizes the building structure's ability to store heat. Set R and C as model parameters to be identified.

[0033] Establish a logical relationship between the air conditioner's on / off status and the indoor temperature control range. When the indoor temperature reaches the set upper limit, the air conditioner starts cooling; when it reaches the set lower limit, the air conditioner stops operating.

[0034] As a preferred embodiment of the present invention, the step of obtaining the optimal parameter value through iterative search includes:

[0035] A composite chaotic initialization strategy that integrates multiple mechanisms with the gray wolf optimization algorithm is used to generate candidate solutions for initial RC parameters;

[0036] In each iteration, the cross-sectional optimization and Cauchy inverse cumulative distribution mutation operations are performed sequentially to update the position of the individual gray wolves;

[0037] The optimal, second-best, and third-best individuals are updated based on the objective function value, guiding the entire population to search for the optimal solution.

[0038] Determine if the iteration termination condition is met. If it is met, output the current optimal RC parameter value. If not, continue iterative optimization.

[0039] As a preferred embodiment of the present invention, the step of air conditioning load aggregation modeling includes:

[0040] Using the optimal equivalent thermal resistance R and equivalent heat capacity C obtained from the identification, a power aggregation calculation model for multiple air conditioning loads is established.

[0041] Set up multiple aggregation scenarios, including scenarios where both the set temperature and the initial indoor temperature are varied, scenarios where only the initial indoor temperature is varied, scenarios where only the set temperature is varied, and comparison scenarios where all temperature parameters are fixed.

[0042] Different set temperature values ​​and initial indoor temperature values ​​are assigned to each air conditioning load using a random distribution method;

[0043] Calculate the duty cycle and average power of the air conditioning load under each scenario, and summarize the data to obtain the aggregated total power change over time.

[0044] As a preferred embodiment of the present invention, the output air conditioning load aggregation characteristic analysis results include:

[0045] The peak-to-average power ratio, standard deviation, and coefficient of variation of the polymerization power under various polymerization scenarios are calculated as volatility evaluation indicators.

[0046] By comparing the volatility indicators of different scenarios, the suppression effect of temperature diversity and initial indoor temperature diversity on polymerization power fluctuations was quantified.

[0047] The dominant and secondary roles of initial indoor temperature diversity and set temperature diversity in the influence of volatility were analyzed.

[0048] Generate an aggregated characteristic analysis report that includes volatility index values, percentage reduction in volatility, and a ranking of the importance of influencing factors.

[0049] As a preferred embodiment of the present invention, the computer-readable storage medium stores a computer program, which, when executed by a processor, implements a method for identifying and modeling air conditioning parameters based on a multi-mechanism fusion gray wolf optimization algorithm.

[0050] The beneficial effects of this invention are:

[0051] 1. This invention overcomes the shortcomings of existing GWO algorithms, such as being prone to getting trapped in local optima, slow convergence speed, and insufficient population diversity, by organically integrating composite chaotic initialization, cross-cutting, and Cauchy inverse CDF mutation mechanisms. This achieves high accuracy and high stability in parameter identification. Experimental results show that this invention outperforms existing methods in terms of convergence speed and mean square error, demonstrating an overall optimization effect of "stability, accuracy, and speed."

[0052] 2. This invention establishes an air conditioning load aggregation model considering the diversity of setpoint temperature and initial indoor temperature based on the identified optimal RC parameters. Through multi-scenario comparative analysis, it innovatively reveals the suppression mechanism of user diversity on aggregated power fluctuations. The study found that the diversity of initial indoor temperature is the dominant influencing factor, while the diversity of setpoint temperature is a secondary influencing factor. The fluctuation suppression effect is optimal when both types of diversity are considered simultaneously. This finding establishes a quantitative evaluation index system including peak-to-average power ratio, standard deviation, and coefficient of variation, providing direct theoretical basis and technical support for the formulation of power grid demand response strategies and the operation and management of virtual power plants. Attached Figure Description

[0053] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0054] Figure 1 This is a schematic diagram of the structure of the air conditioning parameter identification and modeling method based on the multi-mechanism fusion gray wolf optimization algorithm of the present invention;

[0055] Figure 2 This is a diagram of the first-order equivalent thermal parameter model of the present invention;

[0056] Figure 3 This is a graph showing the indoor temperature changes during the air conditioning control period of this invention;

[0057] Figure 4 This is a graph showing the indoor and outdoor temperature and power curves of the present invention. Detailed Implementation

[0058] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0059] Example 1: As Figure 1 As shown, the present invention provides a method for air conditioning parameter identification and modeling based on a multi-mechanism fusion gray wolf optimization algorithm, comprising:

[0060] S100: Collects indoor temperature, outdoor temperature and cooling capacity data during the operation of the air conditioning system, performs data preprocessing, and obtains a normalized dataset.

[0061] Furthermore, the step of deriving the normalized dataset includes:

[0062] The system collects indoor and outdoor temperature data during air conditioner operation using a temperature sensor, and collects air conditioner operating current data using a current sensor and converts it into cooling capacity data.

[0063] The sliding window mean method was used to average the collected data within a time window.

[0064] Box plots were used to identify and remove outliers from the dataset.

[0065] The processed data were normalized using the min-max normalization method to obtain normalized datasets of indoor temperature, outdoor temperature, and cooling capacity.

[0066] Specifically, a data acquisition system was built, including a DC switching power supply, a three-phase power acquisition module, a switching current transformer, and a data logger. The indoor temperature sensor was installed at an appropriate distance from the air conditioner vent, while the outdoor temperature sensor was installed on the well-ventilated exterior of the building, avoiding direct sunlight which could affect measurement accuracy. The current sensor collected the air conditioner's operating current through the current transformer, and converted the current data into corresponding cooling capacity data according to the power calculation formula. An appropriate data sampling frequency was set; a lower sampling frequency was used for temperature data to ensure data stability, while a higher sampling frequency was used for current data to capture details of power changes. The air conditioner was kept running normally during the acquisition process to ensure the authenticity and validity of the data.

[0067] The original data is processed using a sliding window mean method. A fixed time window length is set, and the average value of all data points within the window is calculated as the representative value for that time period, effectively reducing data noise and data volume. Outlier detection employs a box plot statistical method, calculating the quartiles Q1, Q3, and interquartile range (IQR) of the dataset. Data points deviating from the normal distribution range are identified as outliers and removed from the dataset, ensuring data quality. Normalization is performed using a min-max standardization method, which uses linear transformation to uniformly scale data of different dimensions to the same numerical range. For each data category, its minimum and maximum values ​​are identified, and the data are transformed according to the normalization formula to eliminate dimensional differences between data points. The final result is a standardized dataset of indoor temperature, outdoor temperature, and cooling capacity. This dataset serves as input data for a first-order equivalent thermal parameter model, supporting subsequent RC parameter identification and aggregation modeling.

[0068] S200: Based on the traditional gray wolf optimization algorithm, a multi-mechanism fusion gray wolf optimization algorithm is constructed by integrating a composite chaotic initialization strategy, a cross-optimization strategy, and a Cauchy inverse cumulative distribution mutation strategy.

[0069] Based on the traditional gray wolf optimization algorithm, a multi-mechanism fusion gray wolf optimization algorithm is constructed through the following three steps:

[0070] Furthermore, the composite chaotic initialization strategy includes:

[0071] A composite chaotic system is constructed using Tent mapping, Logistic mapping, and Cosine mapping, and the appropriate chaotic mapping is selected for iteration based on the individual position value.

[0072] The ergodicity and randomness of chaotic mapping are used to generate an initial population that is uniformly distributed and highly diverse.

[0073] The fitness of the generated chaotic initial population is pre-evaluated, and individuals with low fitness are removed to reduce redundant computation in subsequent iterations.

[0074] First, a Tent-Logistic-Cosine complex chaotic system is established. The initial population size is set to 2 (corresponding to the two parameters to be identified: equivalent thermal resistance R and equivalent heat capacity C).

[0075] For each individual and each dimension, an initial position value is generated using a composite chaotic system. During the chaotic iteration process, the appropriate chaotic mapping is determined based on the current position value.

[0076] when At that time, a Tent-Logistic hybrid mapping is used:

[0077] ;

[0078] when At this time, Cosine mapping is introduced to smooth the transition:

[0079] ;

[0080] in, For the position of the kth generation individual, The position of the individual in the (k+1)th generation is given, and c∈[0,1] is the control parameter (tests show that the population distribution is most uniform and the MSE error is smallest when c=0.5).

[0081] After generating the initial chaotic population, an initial population screening operation is performed to improve population quality. The fitness value of all individuals is calculated, which is the mean squared error (MSE) of the corresponding RC parameters for each individual. The average fitness value of the population is calculated, and a screening threshold of 0.8 times the average fitness value is set. Individuals with fitness values ​​below this threshold are identified as low-fitness individuals, and 30% of these individuals are removed, retaining the individuals with higher fitness as the final initial population. This screening process reduces redundant computation in subsequent iterations and improves the overall efficiency of the algorithm.

[0082] By applying the aforementioned composite chaotic initialization strategy, the generated initial population exhibits rich diversity and uniform distribution. In the gray wolf optimization algorithm, this results in a more dispersed distribution of the alpha wolf pack (α wolf, β wolf, δ wolf), increasing the number of hunting paths and expanding the hunting range. When the algorithm searches for the optimal RC parameters for the air conditioning load, it can more accurately pinpoint the location of the global optimum, ultimately improving the algorithm's global search capability and convergence speed, laying a solid foundation for accurately identifying the equivalent thermal resistance and equivalent heat capacity parameters of the air conditioning load.

[0083] Furthermore, the cross-sectional optimization strategy includes:

[0084] The horizontal crossover operation randomly pairs individuals within the population and performs arithmetic crossover operations sequentially on all dimensions of the paired individuals, generating new offspring individuals through information exchange between individuals.

[0085] The longitudinal crossover operation performs an arithmetic crossover operation between two different dimensions for each individual in the population, promoting the stagnant dimension to escape premature convergence.

[0086] After the crossover operation is completed, a competitive selection mechanism is adopted. The fitness values ​​are compared to determine whether to adopt a new solution, ensuring that a high-quality solution replaces the parent individual.

[0087] Specifically, the cross-sectional strategy includes two operations: horizontal cross-sectional and vertical cross-sectional, which achieve efficient fusion of population information through information interaction between dimensions.

[0088] The process of lateral crossover involves randomly pairing individuals within the population, with the parent individuals in the random pairing being... and For the D-th dimension parameter, the parent individual , The parameter in this dimension is denoted as , Perform arithmetic crossover operation on the offspring individuals. , New parameters in the D-th dimension , The calculation formula is:

[0089] ;

[0090] ;

[0091] in These are random weights, used to dynamically allocate the proportion of parent individual parameters in the offspring. This is a perturbation factor, introducing random fluctuations into parameter interactions. Through this operation, different individuals generate parameter perturbations in the same dimension, promoting information fusion and diffusion within the population search space. It should be noted that when… and When this happens, the above formula degenerates into a simple average crossover; introducing... The differences between parents can be amplified or reversed to generate more exploratory solutions.

[0092] The vertical crossover process involves performing arithmetic crossover operations between two different dimensions for each individual in the population. Let the individual... The peacekeeping The dimension parameters are respectively and When the generated random number is less than the preset probability threshold r, vertical crossover is performed. The crossover formula is:

[0093] ;

[0094] in These are random weights used for linear mixing between two dimensions within the body. For the cross-th A new interpretation of wei.

[0095] For the core parameters of air conditioning load, since the equivalent thermal resistance R has a higher weight in influencing the air conditioning temperature response, it needs to be treated differently during the vertical cross-parameter process: when the cross-parameter involves the equivalent thermal resistance R parameter, it is assigned a weight coefficient of 1.2; when the cross-parameter involves the equivalent heat capacity C parameter, the weight coefficient is 1.0.

[0096] After updating the vertical crossover parameters, a competitive selection process is performed to determine whether to adopt the new solution. The fitness values ​​of the new individual and the original individual are calculated; if the new solution has better fitness, the individual's position is updated. ;in The fitness function is used to quantitatively evaluate the quality of solutions. This selection process ensures that high-quality new solutions replace their parent solutions, ensuring that updates consistently improve the algorithm's solution quality.

[0097] During the iteration process, the crossover probability is dynamically adjusted with the number of iterations. Crossover probability in the early stages of iteration (first 40% of iterations). Enhance global search capabilities to expand the solution space; crossover probability in the later stages of iteration (last 60% of iterations) Improve the accuracy of local development to optimize the best solution.

[0098] Furthermore, the Cauchy inverse cumulative distribution variation strategy includes:

[0099] Establish the cumulative distribution function of the standard Cauchy distribution, and use the heavy-tailed property of the Cauchy distribution to determine the probability distribution characteristics of the mutation operation;

[0100] The inverse transformation method is used to convert uniformly distributed random numbers into variable-length random numbers that follow a Cauchy distribution;

[0101] In the later stages of algorithm iteration, Cauchy distribution variation perturbation is introduced into the population individuals to enhance the algorithm's ability to escape local optima.

[0102] Specifically, we establish the probability density function and cumulative distribution function of the standard Cauchy distribution. The probability density function of the Cauchy distribution is:

[0103] ;

[0104] in, This is a location parameter used to determine the location of the distribution peak; This is a scale parameter used to define the half-width at half the maximum value. When When the distribution is a standard Cauchy distribution, its probability density function is:

[0105] ;

[0106] The formula for calculating its cumulative distribution function is as follows:

[0107] ;

[0108] When the cumulative distribution function of the Cauchy distribution has an inverse function, the inverse transform method can be used to generate random numbers that follow a uniform distribution. The inverse function of the cumulative distribution function of the Cauchy distribution is defined as:

[0109] ;

[0110] ;

[0111] in, It is a uniformly distributed random number whose value ranges from [0,1]. Let be the dimension of the function. The location parameter 'a' takes a value of 0, consistent with the standard Cauchy distribution; the scale parameter... A value of 0.01 is used to prevent the algorithm from being too aggressive during the search process, thus ensuring that it operates effectively within the design scope.

[0112] By organically integrating the above three strategies, the multi-mechanism fusion gray wolf optimization algorithm effectively solves the shortcomings of the traditional gray wolf algorithm in air conditioning RC parameter identification. The traditional gray wolf algorithm, in the later stages of iteration, suffers from excessive shrinkage of the search region due to the linear decrease in the convergence factor, easily getting trapped in local optima and struggling to effectively handle complex nonlinear optimization problems such as air conditioning RC parameter identification. The composite chaotic initialization strategy provides a good search starting point for the algorithm by generating a uniformly distributed, high-quality initial population; the cross-optimization strategy establishes a balance between global exploration and local development through information exchange between individuals and dimensions; and the Cauchy inverse cumulative distribution mutation strategy injects exploration momentum into the later stages of the algorithm using the heavy-tailed distribution characteristics, enhancing the ability to escape local optima. The three mechanisms work synergistically, significantly improving the algorithm's search performance and convergence accuracy, providing strong technical support for the accurate identification of air conditioning load RC parameters.

[0113] S300: Based on the dataset, establish a first-order equivalent thermal parameter model describing the thermodynamic characteristics of the air conditioning system, set the equivalent thermal resistance R and equivalent heat capacity C as parameters to be identified, and use the multi-mechanism fusion gray wolf optimization algorithm to identify the RC parameters. With the goal of minimizing the mean square error between the predicted indoor temperature and the measured indoor temperature, the optimal parameter values ​​are obtained through iterative search.

[0114] Based on the preprocessed dataset, a first-order equivalent thermal parameter model is established and RC parameters are identified through the following steps:

[0115] Furthermore, the step of establishing a first-order equivalent thermal parameter model describing the thermodynamic characteristics of the air conditioning system, and setting the equivalent thermal resistance R and equivalent heat capacity C as parameters to be identified, includes:

[0116] The air conditioning system is simplified into a first-order equivalent thermal parameter model, ignoring the temperature difference between the inner and outer walls and the temperature difference between the air and the wall surface.

[0117] Based on the principle of energy balance, a differential equation relationship is established between indoor temperature change, indoor-outdoor temperature difference, and air conditioning cooling capacity;

[0118] The differential equation is discretized using the forward Euler method to obtain a discretized indoor temperature prediction equation suitable for numerical solution.

[0119] Determine the physical meaning of equivalent thermal resistance R, which characterizes the resistance to heat transfer between indoor and outdoor environments, and equivalent heat capacity C, which characterizes the building structure's ability to store heat. Set R and C as model parameters to be identified.

[0120] Establish a logical relationship between the air conditioner's on / off status and the indoor temperature control range. When the indoor temperature reaches the set upper limit, the air conditioner starts cooling; when it reaches the set lower limit, the air conditioner stops operating.

[0121] Specifically, the air conditioning system is simplified into a first-order equivalent thermal parameter model, such as... Figure 2 As shown, the temperature difference between the inner and outer walls, as well as the temperature difference between the air and the wall surface, are ignored. This simplification is based on the engineering application requirements of actual air conditioning loads. Although the theoretical accuracy is reduced compared to the third-order and second-order equivalent thermal parameter models, it significantly improves computational efficiency and meets the real-time requirements of large-scale air conditioning load aggregation analysis.

[0122] Based on the principle of energy balance, a differential equation is established to relate indoor temperature changes to the indoor-outdoor temperature difference and the air conditioning cooling capacity. The heat balance equation for the air conditioning system is:

[0123] ;

[0124] ;

[0125] Among them, the equivalent thermal resistance R represents the resistance to heat transfer between the indoor and outdoor environments. Its value reflects the thermal insulation performance of the building structure; a larger R value indicates a better thermal insulation effect. The equivalent heat capacity C represents the building structure's ability to store heat. Its value is related to the mass and specific heat capacity of the building materials; a larger C value indicates a stronger thermal inertia of the building. R and C are set as the model parameters to be identified. It refers to the cooling capacity of the air conditioner; It refers to the indoor air temperature; It is the outdoor temperature; It refers to the air conditioner's energy efficiency ratio; This refers to the air conditioning power. The above basic model is applicable to scenarios where indoor and outdoor air are fully mixed and there is no forced ventilation (doors and windows are open for ≤5 minutes / hour).

[0126] In special scenarios with forced ventilation (doors and windows open for more than 5 minutes per hour), a ventilation heat loss correction factor needs to be introduced to correct the model's heat balance relationship. The formula for calculating the ventilation heat loss correction factor is as follows:

[0127] ;

[0128] in This is the correction factor for ventilation heat loss. For ventilation volume, air density, The specific heat capacity of air at constant pressure. This represents the indoor and outdoor temperature difference. It should be noted that in actual calculations, the molecular part... This is expressed as the heat transfer power generated by ventilation per second, and its unit can be unified to kW through conversion, along with the denominator. The units are consistent. The corrected heat balance equation is:

[0129] ;

[0130] After introducing this correction factor, the model can more accurately describe the indoor temperature change pattern under ventilation conditions, and the temperature prediction error of the corrected model can be reduced by 15%.

[0131] The differential equations are discretized using the forward Euler method. The simulation step size is then set. This transforms continuous-time differential equations into discretized forms suitable for numerical solutions:

[0132] ;

[0133] Where t is the time step index;

[0134] The air conditioning state model is set as follows: This indicates the air conditioner's operating status: '1' indicates it's active, and '0' indicates it's in standby mode. It establishes a logical relationship between the air conditioner's on / off status and the indoor temperature control range: when the indoor temperature reaches the set upper limit... When the air conditioner starts cooling ( When the set lower limit is reached... When the air conditioner stops running ( 0).

[0135] Lianli Air Conditioner Switch State Model: When the air conditioner is off :

[0136] ;

[0137] When the air conditioner is turned on :

[0138] ;

[0139] Furthermore, the objective function is to minimize the mean square error between the predicted and measured indoor temperatures. The mean square error is defined as the average of the squares of the differences between the predicted and actual indoor temperatures.

[0140] ;

[0141] in The indoor temperature is predicted based on a first-order equivalent thermal parameter model. This refers to the actual measured indoor temperature. This represents the total number of data points. This objective function quantifies the model's prediction accuracy. The smaller the value, the higher the degree of fit between the model and the actual system.

[0142] Furthermore, the step of obtaining the optimal parameter value through iterative search includes:

[0143] A composite chaotic initialization strategy that integrates multiple mechanisms with the gray wolf optimization algorithm is used to generate candidate solutions for initial RC parameters;

[0144] In each iteration, the cross-sectional optimization and Cauchy inverse cumulative distribution mutation operations are performed sequentially to update the position of the individual gray wolves;

[0145] The optimal, second-best, and third-best individuals are updated based on the objective function value, guiding the entire population to search for the optimal solution.

[0146] Determine if the iteration termination condition is met. If it is met, output the current optimal RC parameter value. If not, continue iterative optimization.

[0147] Specifically, a composite chaotic initialization strategy integrating multiple mechanisms of the gray wolf optimization algorithm is used to generate candidate solutions for the initial RC parameters. A search boundary is set, and the search range for the equivalent thermal resistance R is defined as follows: The search range for the equivalent heat capacity C is This range is determined based on engineering experience values ​​for typical building air conditioning systems.

[0148] In each iteration, the core operations of the algorithm are executed sequentially: First, according to the position update mechanism of the traditional gray wolf algorithm, the distance between the gray wolf individual and the α wolf, β wolf, and δ wolf is calculated, and the individual position is updated; then, the horizontal and vertical crossover optimization operation is performed, which realizes information interaction between individuals through horizontal crossover and breaks the correlation between dimensions through vertical crossover; finally, in the later stage of the iteration, the Cauchy inverse cumulative distribution mutation operation is introduced to inject random perturbation into the population.

[0149] Update the best, second-best, and third-best individuals based on the objective function value. Calculate the value corresponding to each candidate solution. Value, will Let the smallest individual be α wolf (the optimal solution). The second smallest individual is designated as β wolf (suboptimal solution). The third smallest individual is designated as the δ wolf (the third optimal solution). These three alpha wolves guide the entire population in the search towards the optimal solution.

[0150] Set the iteration termination condition: terminate the iteration when the maximum number of iterations (1000) is reached (based on the central limit theorem in statistics, sample size > 30 is sufficient); or terminate the iteration when the improvement of the optimal solution in 50 consecutive iterations is less than 1000. Early termination of the iteration indicates that the algorithm has converged. Check if the iteration termination condition is met; if so, output the current optimal RC parameter values ​​(the solution corresponding to α wolf); otherwise, continue iterative optimization.

[0151] Through the iterative search process described above, the multi-mechanism fusion gray wolf optimization algorithm can efficiently find the optimal solution within the search space of RC parameters. The minimum parameter combination ultimately yields equivalent thermal resistance R and equivalent heat capacity C values ​​that accurately describe the thermodynamic characteristics of the air conditioning system, providing precise basic parameters for subsequent air conditioning load aggregation modeling. It should be noted that the algorithm searches for the optimal solution on normalized data, but the final output optimal parameter values ​​(R and C) are values ​​with actual physical dimensions after inverse normalization processing.

[0152] S400: Based on the identified optimal parameter values, perform air conditioning load aggregation modeling, analyze the impact of the diversity of set temperature and initial indoor temperature on the aggregation power fluctuation, and output the air conditioning load aggregation characteristic analysis results.

[0153] Based on the aforementioned optimal R and C parameters that represent the average thermodynamic characteristics of the target group of air conditioners, an aggregation model containing N air conditioners is established. In the aggregation, the set temperature and initial indoor temperature of each air conditioner can be set to different values ​​(to simulate diversity), but their thermal resistance R and heat capacity C all adopt the identified unified optimal values.

[0154] Furthermore, the steps of the air conditioning load aggregation modeling include:

[0155] Using the optimal equivalent thermal resistance R and equivalent heat capacity C obtained from the identification, a power aggregation calculation model for multiple air conditioning loads is established.

[0156] Set up multiple aggregation scenarios, including scenarios where both the set temperature and the initial indoor temperature are varied, scenarios where only the initial indoor temperature is varied, scenarios where only the set temperature is varied, and comparison scenarios where all temperature parameters are fixed.

[0157] Different set temperature values ​​and initial indoor temperature values ​​are assigned to each air conditioning load using a random distribution method;

[0158] Calculate the duty cycle and average power of the air conditioning load under each scenario, and summarize the data to obtain the aggregated total power change over time.

[0159] Specifically, based on the aforementioned parameter identification results, a duty cycle calculation model for a single air conditioner load is established. Assuming the outdoor temperature remains constant within a given time period, when t=0, the indoor temperature drops to the set lower limit, at which point the air conditioner turns off (s(t)=0), and then the indoor temperature begins to rise when t>0. Similarly, when t=0, the indoor temperature rises to the set upper limit, at which point the air conditioner turns on (s(t)=1), and then the indoor temperature begins to decrease when t>0. Figure 3 As shown;

[0160] Under constant outdoor temperature, the control cycle of air conditioning load Closing time and opening time The calculation formula is:

[0161] ;

[0162] ;

[0163] ;

[0164] in and These are the user-defined upper and lower limits of indoor temperature. Outdoor temperature This refers to the cooling capacity of the air conditioner.

[0165] The formula for calculating the duty cycle D of the air conditioning load is:

[0166] ;

[0167] in Let i be the rated power of the i-th air conditioner. This represents the duty cycle of the air conditioning load. Average power:

[0168] ;

[0169] In this embodiment, the aggregation scenario parameters are set as follows: number of aggregated air conditioners N = 1000, outdoor temperature... The allowable deviation of the air conditioner set temperature is 0.5℃, the temperature lag interval is 1℃, the time interval is 1 minute, and the simulation duration is 2 hours (covering 3-4 complete air conditioner switching cycles).

[0170] Four different aggregation scenarios were set up for comparative analysis:

[0171] Scenario 1: The user sets a temperature that is a random integer that is uniformly distributed within the range of 22 to 26°C, and the initial indoor temperature is a random integer that is uniformly distributed within the range of 22 to 28°C.

[0172] Scenario 2: The user sets the temperature to a fixed 23℃, and the initial indoor temperature is a random integer that is evenly distributed within the range of 22 to 28℃.

[0173] Scenario 3: The user sets a temperature that is a random integer that is evenly distributed within the range of 22 to 26℃, and the initial indoor temperature is fixed at 26℃.

[0174] Scenario 4: The user sets the temperature to a constant 23℃, and the initial indoor temperature is fixed at 26℃.

[0175] Different setpoint temperature values ​​and initial indoor temperature values ​​are assigned to each air conditioning load using a random distribution method. For parameters requiring diversity, a uniformly distributed random number generator is used to assign parameter values ​​to each air conditioner within a specified range, ensuring the randomness and uniformity of parameter distribution.

[0176] Calculate the duty cycle and average power of the air conditioning load under each scenario. Based on the set temperature range and initial indoor temperature of each air conditioner, and considering the outdoor temperature conditions, calculate the on-time, off-time, and duty cycle of each air conditioner during the simulation period. The average power of a single air conditioner is the product of its rated power and duty cycle. Summarize the power contributions of all air conditioners to obtain a data sequence of the aggregated total power changing over time.

[0177] Furthermore, the analysis results of the output air conditioning load aggregation characteristics include:

[0178] The peak-to-average power ratio, standard deviation, and coefficient of variation of the polymerization power under various polymerization scenarios are calculated as volatility evaluation indicators.

[0179] By comparing the volatility indicators of different scenarios, the suppression effect of temperature diversity and initial indoor temperature diversity on polymerization power fluctuations was quantified.

[0180] The dominant and secondary roles of initial indoor temperature diversity and set temperature diversity in the influence of volatility were analyzed.

[0181] Generate an aggregated characteristic analysis report that includes volatility index values, percentage reduction in volatility, and a ranking of the importance of influencing factors.

[0182] Specifically, the volatility evaluation indicators of aggregation power under various aggregation scenarios are calculated. Peak-to-average power ratio (PAPR) C, standard deviation (σ), and coefficient of variation (CV) are selected as the core indicators for quantifying power volatility characteristics.

[0183] The formula for calculating the peak-to-average power ratio (C) is:

[0184] ;

[0185] in This represents the peak value of the polymerization power. This represents the average polymerization power. The peak-to-average power ratio (PAPR) reflects extreme power fluctuations; a higher value indicates more significant power fluctuations.

[0186] The formula for calculating the standard deviation σ is:

[0187] ;

[0188] in, Polymerization power at the 1st The value at each sampling time. : Number of sample points (total number of time points);

[0189] The coefficient of variation is the ratio of standard deviation to average power. It eliminates the influence of magnitude and reflects the relative fluctuation of power. The higher the value, the greater the power fluctuation and the worse the system stability.

[0190] By comparing volatility indicators across different scenarios, the effects of temperature diversity and initial indoor temperature diversity on suppressing polymerization power fluctuations were quantified. Using scenario 4 as the baseline scenario, the percentage reduction in volatility of the other three scenarios relative to the baseline scenario in terms of peak-to-average power ratio (C), standard deviation (σ), and coefficient of variation (CV) was calculated.

[0191] ;

[0192] This study analyzes the dominant and secondary roles of initial indoor temperature diversity and setpoint temperature diversity in the fluctuation of air conditioning load. By comparing the fluctuation suppression effects of Scenario 2 (initial indoor temperature diversity only) and Scenario 3 (setpoint temperature diversity only), the relative importance of the two factors is determined. The results show that initial indoor temperature diversity has a more significant impact on the fluctuation of aggregated air conditioning load; considering only this factor reduces the standard deviation by 66.78% and the coefficient of variation by 65.05%. The impact of setpoint temperature diversity is relatively weak; considering only this factor reduces the standard deviation and coefficient of variation by 49.45% and 40.77%, respectively.

[0193] Generate an aggregated characteristic analysis report containing volatility index values, volatility reduction percentages, and a ranking of the importance of influencing factors. The report includes: specific values ​​for peak-to-average ratio (PMR), standard deviation (σ), and coefficient of variation (CV) under four scenarios; the degree of volatility improvement in each scenario relative to the baseline scenario; the conclusion that initial indoor temperature diversity is the dominant influencing factor, while temperature diversity is considered a secondary influencing factor; and the optimal volatility suppression effect achieved when both diversity factors are considered simultaneously, providing a theoretical basis for power grid dispatching and demand response strategy development.

[0194] Example 2:

[0195] The data collection for this field study took place in a university laboratory in eastern China. The collected data included outdoor temperature, indoor temperature, and indoor current. The experimental subject was a Gree KFR-72LW / (72569S)Bb-2 split-type floor-standing heat pump air conditioner. This model is a mainstream household split-type heat pump, covering over 80% of household user scenarios. During peak summer electricity consumption, the operation of air conditioner clusters causes significant fluctuations in the power grid. To assess the dispatchability potential of the air conditioner cluster and verify the effectiveness of the method of this invention, indoor and outdoor temperature and current data were collected from the air conditioners in the building (temperature sampling interval 5s, current sampling interval 1s, approximately 690 valid samples obtained after preprocessing). The indoor and outdoor temperature and power curves are shown below. Figure 4 As shown.

[0196] First, parameters are identified using a first-order RC model. Throughout the optimization process, the objective fitness function is the mean squared error (MSE) metric. Algorithms used for comparison include GWO, PSO, GA, and WOA. It is worth noting that because the initial values ​​are randomized during the iterative optimization process, the initial positions of the algorithm iterations are different for each algorithm. Table 1 shows a comparison of the optimization algorithm results.

[0197] surface Optimization Algorithm Results Table

[0198]

[0199] The results above show that during the identification of R and C parameters using different algorithms, the mean squared error (MSE) gradually converged to 0.005855, and both R and C converged to fixed values. The MF-GWO algorithm showed an advantage over other algorithms, with a lower MSE value, indicating that the improved algorithm effectively increased the accuracy of model identification. Furthermore, almost all algorithms converged to the same value, suggesting they are closer to the true values ​​and the identification results are more accurate. Therefore, the parameter R in the subsequent air conditioning load aggregation process was set to 0.6157℃・kW−1, and C was set to 94.9725((kW·h)·C−1).

[0200] Subsequently, the parameters set in the aggregation considering real-time indoor temperature changes were: N=1000, C=94.9725, R=0.6157, energy efficiency ratio (COP)=3.28, power (P)=35 kW, allowable deviation of air conditioner set temperature of 0.5℃, set temperature (To)=36℃, temperature lag interval of 1, and time interval of 1 minute. The results showed:

[0201] surface Peak-to-peak ratio, standard deviation (o), and coefficient of variation (CV) under different scenarios

[0202]

[0203] surface Peak-to-peak ratio, standard deviation (o), and coefficient of variation (CV) under different scenarios

[0204]

[0205] This practical example demonstrates that the method of this invention can not only obtain more accurate thermal parameters in complex environments, but also quantitatively reveal the impact of different user diversity on load aggregation fluctuations. In particular, when considering the difference between the setpoint temperature and the initial room temperature, the power curve is smoother and the overall adjustable capacity is more reliable, providing a direct reference for grid demand response and virtual power plant dispatch.

[0206] It should be noted that for common air conditioning load scenarios such as office buildings and small commercial outlets, this method can be reused without changing the algorithm and constraint settings of Example 1, only the sampling data and aggregation scale of the corresponding scenario need to be replaced:

[0207] When there is prolonged ventilation or frequent door opening (strong ventilation) on site, a ventilation term can be introduced according to the heat loss correction method given in Example 1 to further reduce temperature prediction deviation and enhance the credibility of polymerization assessment.

[0208] When the aggregation scale, setting strategy, or comfort constraints are adjusted, simply replace the distribution and interval based on the four scenarios in this embodiment to output a wave suppression index (C / σ / CV) that can be directly used for scheduling evaluation.

[0209] This embodiment does not repeat the same or similar content as Embodiment 1 in terms of model establishment, objective function, parameter boundaries, and iteration process; anything not mentioned herein shall be based on Embodiment 1.

[0210] Example 3:

[0211] In a third embodiment of the present invention, based on the same inventive concept, the present invention proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the identification and modeling method described in the above embodiments.

[0212] It should be understood that the air conditioning parameter identification and modeling method based on the multi-mechanism fusion gray wolf optimization algorithm of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system.

[0213] Specifically, the present invention can be implemented by a computer-readable storage medium storing a computer program, which, when executed by a processor, implements all the steps of the air conditioning parameter identification and modeling method based on the multi-mechanism fusion gray wolf optimization algorithm.

[0214] The computer-readable storage medium includes, but is not limited to, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code. The computer program includes program code for implementing functional modules such as data acquisition and preprocessing, construction of a multi-mechanism fusion gray wolf optimization algorithm, establishment of a first-order equivalent thermal parameter model, RC parameter identification and optimization, air conditioning load aggregation modeling, and aggregation characteristic analysis.

[0215] If a hardware implementation is adopted, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits with logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) with suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), etc. Hardware implementations are particularly suitable for applications requiring high-speed processing of large-scale air conditioning load data.

[0216] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An air conditioner parameter identification and modeling method based on a multi-mechanism fusion gray wolf optimization algorithm, characterized in that, The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm.

2. The multi-mechanism fusion grey wolf optimization algorithm-based air conditioner parameter identification and modeling method according to claim 1, characterized in that, The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization algorithm. The application relates to an air conditioner system load aggregation modeling method based on a multi-mechanism fusion grey wolf optimization The indoor temperature and outdoor temperature data during the operation of the air conditioner are collected by a temperature sensor, and the current data of the air conditioner is collected by a current sensor and converted into refrigeration capacity data; The collected data is processed by the sliding window mean method to obtain the average value within the time window; The box plot method is used to identify and eliminate abnormal values in the data set; The processed data is normalized by the minimum-maximum normalization method to obtain the normalized indoor temperature, outdoor temperature and refrigeration capacity data set.

3. The multi-mechanism fusion grey wolf optimization algorithm-based air conditioner parameter identification and modeling method according to claim 1, characterized in that, The step of establishing a first-order equivalent thermal parameter model describing the thermodynamic characteristics of the air conditioning system, and setting the equivalent thermal resistance R and the equivalent thermal capacity C as the parameters to be identified, comprises: The air conditioning system is simplified as a first-order equivalent thermal parameter model, and the temperature difference between the inner wall and the outer wall and the temperature difference between the air and the wall are ignored; According to the energy balance principle, a differential equation relationship between the indoor temperature change and the indoor and outdoor temperature difference and the air conditioner refrigeration capacity is established; The differential equation is discretized by the forward Euler method to obtain a discretized indoor temperature prediction equation suitable for numerical solution; The physical meanings of the equivalent thermal resistance R representing the heat transfer resistance between the indoor and outdoor environments and the equivalent thermal capacity C representing the heat storage capacity of the building structure are determined, and R and C are set as the model parameters to be identified; The logical relationship between the air conditioner on-off state and the indoor temperature control interval is established, and the air conditioner starts refrigeration when the indoor temperature reaches the set upper limit, and stops running when the set lower limit is reached.

4. The multi-mechanism fusion grey wolf optimization algorithm-based air conditioner parameter identification and modeling method according to claim 1, characterized in that, The step of air conditioning load aggregation modeling comprises: Using the identified optimal equivalent thermal resistance R and equivalent thermal capacity C parameter values, a multi-air-conditioner load aggregation power calculation model is established; A plurality of aggregation scenarios are set, including setting temperature and initial indoor temperature diversification scenarios, only initial indoor temperature diversification scenarios, only setting temperature diversification scenarios, and comparison scenarios with fixed temperature parameters; Different setting temperature values and initial indoor temperature values are assigned to each air conditioning load in a random distribution manner; The duty cycle and average power of the air conditioning load in each scenario are calculated, and the data of the total aggregated power varying with time is obtained.

5. The multi-mechanism fusion grey wolf optimization algorithm-based air conditioner parameter identification and modeling method according to claim 1, characterized in that, The output air conditioning load aggregation characteristic analysis results comprise: The peak-to-average ratio, standard deviation and coefficient of variation of the aggregated power in each aggregation scenario are calculated as the volatility evaluation indexes; By comparing the volatility evaluation indexes of different scenarios, the suppression effect of setting temperature diversity and initial indoor temperature diversity on the volatility of the aggregated power is quantified; The dominant role and auxiliary role of initial indoor temperature diversity and setting temperature diversity on volatility are analyzed; An aggregation characteristic analysis report containing the values of the volatility evaluation indexes, the percentage of volatility reduction and the importance ranking of the influencing factors is generated.

6. A computer readable storage medium characterized by, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the air conditioning parameter identification and modeling method based on the multi-mechanism fusion grey wolf optimization algorithm according to any one of claims 1 to 5.

Images