A neural network-based ventilation system energy-saving optimization method and system

Abstract

The application provides a ventilation system energy-saving optimization method and system based on a neural network, and the method comprises the following steps: collecting historical operation data of a fan, and performing mean value interpolation and wavelet transform filtering pretreatment on the historical operation data; an improved human evolution optimization algorithm (IHEOA) is obtained by optimizing a human evolution optimization algorithm (HEOA) by using a Gaussian variation strategy; parameters of a SOFTS model are optimized by using the IHEOA algorithm; an IHEOA-SOFTS ventilation system model is constructed, and the relationship between air volume opening degree, indoor temperature and energy consumption is established; a comprehensive objective function considering indoor temperature control accuracy and total power consumption is established, and the opening degree of each fan in the IHEOA-SOFTS ventilation system model is continuously optimized and adjusted by using the IHEOA algorithm, so that the optimal relationship between the fan opening degree, indoor temperature and total power consumption is determined, thereby realizing accurate control of the fan parameters and energy-saving optimization of the system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of ventilation system energy saving optimization, and particularly relates to a ventilation system energy saving optimization method and system based on a neural network. BACKGROUND

[0002] With the increasing global energy demand and the growing environmental problems, how to effectively save energy and reduce energy consumption has become the focus of various industries. The ventilation system of a building, as an important part of building environmental control, accounts for a considerable proportion of the total building energy consumption.

[0003] Traditional ventilation systems mainly rely on preset operation strategies and simple feedback control, which are difficult to respond to environmental changes and dynamically adjust in real time, resulting in energy waste and low operation efficiency. SUMMARY

[0004] The present application aims to at least partially solve one of the technical problems in the related art.

[0005] To this end, a first object of the present application is to provide a ventilation system energy saving optimization method based on a neural network.

[0006] A second object of the present application is to provide a ventilation system energy saving optimization system based on a neural network.

[0007] A third object of the present application is to provide an electronic device.

[0008] A fourth object of the present application is to provide a computer-readable storage medium.

[0009] A fifth object of the present application is to provide a computer program product.

[0010] To achieve the above objects, a first aspect of the present application provides a ventilation system energy saving optimization method based on a neural network, comprising:

[0011] Collecting historical operation data of the fan, and performing mean interpolation and wavelet transform filtering preprocessing on the historical operation data; the historical operation data includes the opening of each fan, external temperature, humidity, temperature at the fan, and total energy consumption;

[0012] An improved human evolution optimization algorithm IHEOA is obtained by optimizing the human evolution optimization algorithm HEOA using a Gaussian variation strategy;

[0013] The parameters of the SOFTS model are optimized by the IHEOA algorithm; the parameters include embedding dimension, core representation dimension, number of encoder layers, prediction window length, and learning rate;

[0014] An IHEOA-SOFTS ventilation system model is constructed to establish the relationship between the air volume opening degree and indoor temperature and energy consumption; the IHEOA-SOFTS ventilation system model takes the opening degree of each fan, external temperature and humidity as input, and takes the temperature at each fan and total energy consumption as output;

[0015] A comprehensive objective function considering indoor temperature control accuracy and total power consumption is established, and the opening degree of each fan in the IHEOA-SOFTS ventilation system model is continuously optimized and adjusted by using the IHEOA algorithm to minimize the comprehensive objective function, so as to determine the optimal relationship between the fan opening degree and indoor temperature and total power consumption.

[0016] Optionally, the mean interpolation and wavelet transform filtering pretreatment of the historical operation data comprises:

[0017] The mean interpolation is performed on the historical operation data, missing values in the data set are identified, the mean of all non-missing values in the data set is calculated, and the calculated mean is filled into the missing value position, and the calculation formula is as follows:

[0018]

[0019] Wherein, x i is a non-missing value, n is the total number of non-missing values, and mean is the mean of the position of the filled missing value;

[0020] The mean interpolated data set is subjected to wavelet transform filtering, and Daubechies wavelet is selected as the wavelet function;

[0021] The filtered data is subjected to discrete wavelet transform to obtain wavelet coefficients, and the calculation formula is as follows:

[0022]

[0023] ψ j,k (m)=2 -j / 2 ψ(2 -j ·m-k)

[0024] Wherein, W j,k is a wavelet coefficient at scale j and position k, j is a scale, k is a position parameter, m is a position index of the original data, and ψ j,k (m) is a wavelet function after scaling and translation, scaling and translation are realized by scaling the original wavelet function ψ by 2 -j / 2 times and translating 2 -j ·m-k, and f(m) is the value of the original data at position m;

[0025] The wavelet coefficients are subjected to hard threshold processing to remove noise and unimportant information, and the formula is as follows:

[0026]

[0027] where λ is a set threshold, if the absolute value of the wavelet coefficient W j,k is greater than the threshold λ, the coefficient is retained, otherwise it is set to 0;

[0028] The data is reconstructed using the wavelet coefficients processed by the threshold, and the formula is as follows:

[0029]

[0030] where f'(m) is the value of the reconstructed data at position m, obtained by multiplying all the reconstructed wavelet coefficients W th,j,k with the corresponding wavelet function ψ j,k (m) and summing over all scales j and positions k.

[0031] The overall reconstructed data is represented as:

[0032] f' = {f'(m) | m = 1, 2,..., M}

[0033] where M is the length of the signal, and f' is the overall reconstructed data.

[0034] Optionally, the human evolution optimization algorithm HEOA is optimized using a Gaussian mutation strategy to obtain an improved human evolution optimization algorithm IHEOA, comprising:

[0035] To simulate the chaotic stage in the early stage of human evolution, the population is initialized using a logistic chaotic mapping method, and the calculation formula is as follows:

[0036] x i = α·x i-1 ·(1-x i-1 ), 0 ≤ x0≤ 1, i = 1, 2,..., N, α = 4

[0037] where x i represents the i-th iteration value, x i-1 represents the previous iteration value, N is the population size, and α is the control parameter of the chaotic mapping; the chaotic value x i is mapped to the search space to obtain the position of the initial population, and the calculation formula is as follows:

[0038]

[0039] where x represents the initial position of the i-th individual, lb and ub are the lower and upper limits of the search space, respectively;

[0040] In each iteration of the human evolution optimization algorithm, the individual fitness of each individual in the population is calculated, and a Gaussian mutation strategy is applied to generate a normally distributed random number perturbation to fine-tune the new solution near the original solution, and the update formula is as follows:

[0041] X new = X old + Gaussian(0, σ)

[0042] where X new is the new solution after mutation, X old is the original solution before mutation, Gaussian(0, σ) represents a Gaussian distributed random number with mean 0 and variance σ 2 In the human exploration phase, the individual searches by the Gaussian mutation strategy, and the formula for updating the position is as follows:

[0043]

[0044] where β is an adaptive function, dim is the dimension of the problem, and represent the current position and the updated next position of the i-th individual, X best is the best position explored so far, is the average position of the current population, Levy represents the levy distribution, f jump is the jump coefficient, and rand is a random number in the range [0, 1];

[0045] In the human development phase, the population is divided into leaders, explorers, followers and failures, and the Gaussian mutation strategy is applied to each role, and the position is updated using the corresponding strategy;

[0046] Check the termination condition, if the termination condition is met, stop iteration and output the optimal solution.

[0047] Optionally, in the human development phase, the population is divided into leaders, explorers, followers and failures, and the Gaussian mutation strategy is applied to each role, and the position is updated using the corresponding strategy, including:

[0048] For each leader i in the population, apply the Gaussian mutation strategy, and the formula is as follows:

[0049]

[0050] where X i+1 is the new solution after mutation, X i is the original solution before mutation, is a Gaussian random variable with mean 0 and variance ;

[0051] The top 40% of pre-adapted individuals in the population are designated as leaders, and the position is updated using the leader strategy with a Gaussian mutation strategy, as follows:

[0052]

[0053] where, is the position of the ith individual in the t+1th generation, is the position of the ith individual in the tth generation, Rnis a random number following a normal distribution, the function ones(1, dim) generates a row vector containing dim elements, each of which is set to 1, R is a random number ranging from 0 to 1, representing the complexity of the situation related to the leader, A is the situation evaluation value, used to determine whether the leader changes its search strategy, ω is the knowledge acquisition difficulty coefficient, which gradually decreases over time, and the expression is:

[0054]

[0055] The top 40% to 80% of pre-adapted individuals in the population are designated as explorers, and the position is updated using the explorer strategy with a Gaussian mutation strategy, as follows:

[0056]

[0057] where, is the position of the ith individual in the t+1th generation, is the position of the ith individual in the tth generation, is the least adapted individual in the tth generation population, represents a Gaussian random variable with mean 0 and variance ;

[0058] The top 80% to 90% of pre-adapted individuals in the population are designated as followers, and the Gaussian mutation strategy is applied, as follows:

[0059]

[0060] where, represents the position of the most adapted individual in the human population at the tth iteration, Rd represents a random number in the range of [1, dim], represents a Gaussian random variable with mean 0 and variance ;

[0061] The last 10% of pre-adapted individuals in the population are designated as failures, and the population replenishment equation is improved after the Gaussian mutation strategy:

[0062]

[0063] wherein, X i+1 is the position of the newly generated individual of the population, X best is the position of the individual with the highest fitness in the population, X i is the position of the loser individual in the population, Rn represents a random number following a normal distribution, and σ is a parameter controlling the amplitude of Gaussian variation, referred to as a standard deviation, is a standard normal distribution.

[0064] Optionally, the optimization of the parameters of the SOFTS model by the IHEOA algorithm comprises:

[0065] A target function f(θ) is defined according to the mean square error of the SOFTS model, wherein θ is a set of parameters to be optimized, including an embedding dimension d, a core representation dimension d', an encoder layer number N, a prediction window length H, and a learning rate L; and a formula of the target function f(θ) is as follows:

[0066]

[0067] wherein N is a sample number, Y i is an actual value, is a model prediction value;

[0068] In initialization, an initial parameter set θ={d, d', N, H, L} is generated by a population initialization method of the IHEOA algorithm;

[0069] A prediction model is defined, and a formula is as follows:

[0070]

[0071] wherein X i is an input feature, and SOFTS(θ, X i ) is a prediction result of the SOFTS model based on the parameter set θ;

[0072] In each iteration, an exploration and development stage of the IHEOA algorithm is applied to update the parameter set θ;

[0073] The performance of the SOFTS model corresponding to each parameter set is calculated by calculating a current fitness, and an optimal parameter set is updated;

[0074] When a termination condition is reached, an optimal parameter set is output, and is applied to the SOFTS model.

[0075] Optionally, the prediction process of the IHEOA-SOFTS ventilation system model comprises:

[0076] A preprocessed data set X t ={A t , T is input into the ventilation system model.ext,t ,H ext,t},wherein X t is the total input of the model at time step t, A t is the opening of each fan at time step t, T ext,t is the external temperature at time step t, H ext,t is the external humidity at time step t;

[0077] In the ventilation system model, the input data is embedded into a high-dimensional space by an embedding layer, and the formula is:

[0078] S0=Embedding(X t )

[0079] wherein S0 is the output feature representation of the embedding layer;

[0080] The output feature representation of the embedding layer is input into the STAR module, and the output feature representation of the i-th layer is obtained by time series processing, and the formula is:

[0081] S i =STAR(S i-1 )

[0082] wherein S i represents the feature representation output by the i-th layer in the STAR module, and S i-1 represents the output feature representation of the i-1-th layer in the STAR module;

[0083] The output feature representation of the i-1-th layer is processed by a multi-layer perception to extract a core representation, and the formula is as follows:

[0084] o i =MLP(S i-1 )

[0085] wherein o i is the core representation of the i-th layer;

[0086] The core representation and the feature representation are repeated and spliced to generate a combined feature representation, and the formula is:

[0087] F i =Repeat_Concat(S i-1 ,o i )

[0088] wherein F i is the combined feature of the feature representation input into the i-th multi-layer perception and the core representation output;

[0089] The combined feature representation is processed by an encoder layer, and the original feature and the processed feature representation are combined through a residual, and the formula is as follows:

[0090] S i = MLP(F i ) + S i-1

[0091] where S i is the feature representation output by the encoder layer;

[0092] The feature representation output by the encoder layer is input to a linear layer, and the output prediction is represented as:

[0093]

[0094] where represents the prediction result, S L represents the final feature representation obtained by layer-by-layer processing through a multi-layer neural network, represents the temperature at each fan at the prediction time step t, E total,t represents the total power consumption at the prediction time step t.

[0095] Optionally, the comprehensive objective function considering indoor temperature control accuracy and total power consumption is established, and the opening of each fan in the IHEOA-SOFTS ventilation system model is continuously optimized and adjusted using the IHEOA algorithm to determine the optimal relationship between fan opening and indoor temperature and total power consumption, including:

[0096] A comprehensive objective function f(θ) is established, which considers indoor temperature control accuracy and total power consumption, and is defined as follows:

[0097]

[0098] where T target is the target temperature, is the mean square error of temperature control, is the total power consumption, and w1 and w2 are weight coefficients for balancing the influence of different objectives;

[0099] The IHEOA intelligent algorithm is used to initialize the population to generate an initial fan opening solution set A t Each solution is represented as a combination of fan openings, and the population initialization formula is as follows:

[0100]

[0101] For each solution in the population its fitness value is calculated, which is determined by the comprehensive objective function The formula is:

[0102]

[0103] According to the fitness value, the population is screened, and better solutions are selected to form a new candidate population, and the screened population is represented as:

[0104] P' = Select (P)

[0105] On the basis of the screened population, a crossover operation is performed to generate new candidate solutions, and the crossover formula is:

[0106] P'' = Crossover (P')

[0107] The mutation operation is performed on the new solution generated by the crossover, and each candidate solution is slightly adjusted, and the update formula is:

[0108] P''' = Mutate (P'')

[0109] The mutated solution is fused with the current population to generate an updated population, and the update formula is:

[0110] P = Update (P, P''')

[0111] When the predetermined number of iterations or the target function value converges, the algorithm is terminated and the final optimized fan opening parameter set is output, represented as:

[0112]

[0113] Use the optimized fan opening Control the ventilation system, adjust the relationship between the fan opening and the temperature and total power consumption of each place in the room, and realize the optimization control of the fan parameters and the energy saving optimization of the system.

[0114] To achieve the above purpose, the second aspect of the present application proposes a ventilation system energy saving optimization system based on neural network, comprising:

[0115] A preprocessing module is configured to collect historical operation data of the fan, and perform mean interpolation and wavelet transform filtering preprocessing on the historical operation data; the historical operation data includes the opening of each fan, external temperature, humidity, temperature at the fan and total energy consumption;

[0116] A first optimization module is configured to optimize the human evolution optimization algorithm HEOA by using the Gaussian mutation strategy to obtain the improved human evolution optimization algorithm IHEOA;

[0117] A second optimization module is configured to optimize the parameters of the SOFTS model by the IHEOA algorithm; the parameters include embedding dimension, core representation dimension, encoder layer number, prediction window length and learning rate;

[0118] The model construction module is configured to construct an IHEOA-SOFTS ventilation system model to establish a relationship between the air volume opening degree and the indoor temperature and energy consumption; the IHEOA-SOFTS ventilation system model takes the opening degree of each fan, the external temperature, and the humidity as input and takes the temperature at each fan and the total energy consumption as output;

[0119] The optimal relationship output module is configured to establish a comprehensive target function considering the indoor temperature control accuracy and the total power consumption, to minimize the comprehensive target function, and to continuously optimize and adjust the opening degree of each fan in the IHEOA-SOFTS ventilation system model by using the IHEOA algorithm, to determine the optimal relationship between the fan opening degree and the indoor temperature and the total power consumption.

[0120] To achieve the above object, the third aspect of the present application provides an electronic device, comprising: a processor, and a memory connected with the processor in communication;

[0121] The memory stores computer execution instructions;

[0122] The processor executes the computer execution instructions stored in the memory to implement the method according to any one of the first aspect.

[0123] To achieve the above object, the fourth aspect of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores computer execution instructions, and the computer execution instructions are executed by a processor to implement the method according to any one of the first aspect.

[0124] To achieve the above object, the fifth aspect of the present application provides a computer program product, wherein the computer program is executed by a processor to implement the method according to any one of the first aspect.

[0125] The technical scheme provided by the embodiments of the present application at least brings the following beneficial effects:

[0126] By using the IHEOA-SOFTS model, the ventilation system energy saving optimization method based on the neural network can significantly reduce the total energy consumption of the ventilation system while taking into account the indoor temperature control accuracy; by optimizing the fan opening degree parameter, the optimal relationship between the fan opening degree and the indoor temperature and the total energy consumption is established, the accurate control of the fan parameter is realized, and the energy saving and operation efficiency of the system are improved. The present application can effectively balance the comfort and energy consumption targets, and provides an efficient solution for the intelligent optimization of the ventilation system.

[0127] Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter. BRIEF DESCRIPTION OF DRAWINGS

[0128] The above and / or additional aspects and advantages of the present application will become apparent and more readily appreciated from the following description, taken in conjunction with the following drawings of exemplary embodiments of the present application, wherein:

[0129] Figure 1 A simple flowchart of a neural network-based ventilation system energy-saving optimization method provided by an embodiment of the present application;

[0130] Figure 2 A detailed flowchart of a neural network-based ventilation system energy-saving optimization method provided by an embodiment of the present application;

[0131] Figure 3 A structural diagram of a neural network-based ventilation system energy-saving optimization system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0132] The embodiments of the present application are described in detail below with reference to the accompanying drawings, in which the same or similar components have the same or similar designations and functions throughout. The embodiments described below are exemplary and are intended to explain the present application, and are not to be understood as limiting the present application.

[0133] In view of the problems of high energy consumption, low control precision and inability to accurately respond to environmental changes in the existing ventilation system energy-saving optimization method, an embodiment of the present application provides a neural network-based ventilation system energy-saving optimization method, Figure 1 and Figure 2 A flowchart of a neural network-based ventilation system energy-saving optimization method provided by an embodiment of the present application. Referring to Figure 1 and Figure 2 , the method comprises the following steps:

[0134] Step S1, collecting historical running data of the fan, and performing mean interpolation and wavelet transform filtering preprocessing on the historical running data.

[0135] In the embodiment of the present application, the historical running data includes the opening of each fan, the external temperature, the humidity, the temperature at the fan and the total energy consumption. By preprocessing the collected data, noise can be removed and missing values can be filled, providing high-quality data input for subsequent model construction.

[0136] Specifically, the following sub-steps are included:

[0137] Step S11, performing mean interpolation on the historical running data, identifying missing values in the data set, calculating the mean value of all non-missing values in the data set, and filling the calculated mean value to the missing value position, the calculation formula is as follows:

[0138]

[0139] where x i is the non-missing value, n is the total number of non-missing values, and mean is the mean value at the position where the missing value is filled. Mean imputation can effectively reduce the impact of data missing on model accuracy and provide complete data input for subsequent steps.

[0140] Step S12, select Daubechies wavelet as the wavelet function, and perform wavelet transform filtering on the imputed data. Wavelet transform can decompose the signal into different scales, thereby extracting the feature information in the data and removing high-frequency noise, providing a smooth data representation for subsequent analysis.

[0141] Step S13, perform discrete wavelet transform on the filtered data to obtain wavelet coefficients, and the calculation formula is as follows:

[0142]

[0143] ψ j,k (m)=2 -j / 2 ψ(2 -j ·m-k)

[0144] where W j,k is the wavelet coefficient at scale j and position k, j is the scale, k is the position parameter, m is the position index of the original data, and ψ j,k (m) is the scaled and translated wavelet function, scaling and translation are achieved by scaling the original wavelet function ψ by 2 -j / 2 and translating it by 2 -j ·m-k, and f(m) is the value of the original data at position m. Through discrete wavelet transform, multi-scale features in the data can be extracted to support subsequent processing.

[0145] Step S14, in order to remove noise and unimportant information, hard thresholding is performed on the wavelet coefficients, and the formula is as follows:

[0146]

[0147] where λ is the set threshold value, if the absolute value of the wavelet coefficient W j,k is greater than the threshold value λ, the coefficient is retained, otherwise it is set to 0, and those smaller coefficients are removed, which usually correspond to noise or unimportant signal features.

[0148] This process can effectively remove the noise component in the data.

[0149] Step S15, based on the wavelet coefficients after thresholding, reconstruct the smooth data signal, and the formula is as follows:

[0150]

[0151] where f'(m) is the value of the reconstructed data at position m, obtained by multiplying all the reconstructed wavelet coefficients W th,j,k with the corresponding wavelet function ψ j,k (m) and summing over all scales j and positions k.

[0152] The overall reconstructed data is denoted as:

[0153] f' = {f'(m) | m = 1, 2,..., M}

[0154] where M is the length of the signal and f' is the overall reconstructed data.

[0155] The reconstructed data is smooth and noise-free, providing high-quality input data for subsequent steps.

[0156] Step S2, the Gaussian mutation strategy is used to optimize the human evolution optimization algorithm HEOA, and the improved human evolution optimization algorithm IHEOA is obtained.

[0157] In the embodiments of the present application, in order to improve the global optimization ability of the algorithm, the human evolution optimization algorithm HEOA is improved by introducing a Gaussian mutation strategy to obtain the improved human evolution optimization algorithm IHEOA. The IHEOA algorithm generates a population in the initial stage and adjusts the positions of the population individuals through multiple iterations to optimize the objective function.

[0158] Specifically, the following sub-steps are included:

[0159] Step S21, the population is initialized by a logical chaotic mapping method to generate an initial population with randomness and uniform distribution. The calculation formula of the logical chaotic mapping is as follows:

[0160] x i = α · x i-1 · (1 - x i-1 ), 0 ≤ x0≤ 1, i = 1, 2,..., N, α = 4

[0161] where x i represents the ith iteration value, x i-1 represents the previous iteration value, N is the population size, and α is the control parameter of the chaotic mapping. The chaotic value x i is mapped to the search space to obtain the position of the initial population, and the calculation formula is:

[0162]

[0163] where x new represents the initial position of the ith individual, lb and ub are the lower and upper limits of the search space, respectively.

[0164] By initializing the population, the uniformity of the initial solution distribution is ensured, providing a good foundation for subsequent optimization.

[0165] Step S22, in each iteration of the human evolution optimization algorithm, the individual fitness of each individual in the population is calculated, and a normally distributed random number is generated by applying the Gaussian mutation strategy to perturb the new solution near the original solution, and the update formula is as follows:

[0166] X new = X old + Gaussian(0, σ)

[0167] Where X new is the new solution after mutation, X old is the original solution before mutation, Gaussian(0, σ) represents a Gaussian distributed random number with mean 0 and variance σ 2 .

[0168] At the same time, in the human exploration stage, in the face of unexplored areas and limited knowledge, individuals tend to use a unified search strategy. In the embodiments of the present application, the individual updates the position by combining the Gaussian mutation strategy with the adaptive search method, and the update formula is as follows:

[0169]

[0170] Where β is the adaptive function, dim is the dimension of the problem, X i and X best i+1 represent the current position and the updated next position of the i-th individual, respectively, X is the best position explored so far, is the average position of the current population, Levy represents the levy distribution, f jump is the jump coefficient, and rand is a random number in the range [0, 1].

[0171] The calculation formula of the average position

[0172]

[0173] The adaptive function β function is used to adjust the parameters according to the iteration number and the current position, reflecting the gradual difficulty of knowledge exploration and the group characteristics of humans, and the calculation formula is as follows:

[0174]

[0175] In order to simulate the acquisition and spiral development of knowledge in the exploration stage, Levy distribution is used. The expression of Levy distribution is as follows:

[0176]

[0177] where μ ~ N(0, D), v ~ N(0, D), γ is 1.5.

[0178] In order to improve the dispersion of the search position, a jump strategy is introduced in the exploration stage, and the jump coefficient f jump The expression is as follows:

[0179]

[0180] Step S22, in the human development stage, the population is divided into leaders, explorers, followers and failures, and a Gaussian mutation strategy is applied to each role, and the position is updated using the corresponding strategy.

[0181] (1) Leader: for each leader i in the population, a Gaussian mutation strategy is applied, and the formula is as follows:

[0182]

[0183] where X i+1 is the new solution after mutation, X i is the original solution before mutation, represents a Gaussian random variable with mean 0 and variance .

[0184] The leader has rich knowledge and is usually located in the optimal area. In the embodiment of the application, the top 40% of individuals in the pre-adaptation population are designated as leaders, and the position is updated using the leader strategy and the Gaussian mutation strategy, and the formula is as follows:

[0185]

[0186] where, is the position of the i-th individual in the t+1 generation, is the position of the i-th individual in the t generation, Rn represents a random number following a normal distribution, the function ones(1, dim) represents generating a row vector containing dim elements, where each element is set to 1, R is a random number ranging from 0 to 1, representing the complexity of the situation related to the leader, A is a situation evaluation value, used to determine whether the leader changes its search strategy, in the embodiment of the application, the value of A can be 0.6, according to the complexity of the specific position, the leader will select the appropriate search strategy.

[0187] ω is the knowledge acquisition difficulty coefficient, which gradually decreases over time, and the expression is as follows:

[0188]

[0189] When R < A, the position update of the individual is influenced by the current generation t, a random number rand and the maximum number of iterations Max iter The exponential decay term ensures that the magnitude of the position update gradually decreases over time, which helps to converge to a local optimal solution.

[0190] When R ≥ A, the position update of the individual is determined by the weight coefficient ω and an offset based on a normally distributed random number Rn, which is a vector of all ones, indicating that the explorer has made a certain magnitude of movement in each dimension. This strategy increases the diversity of the search space and helps to jump out of the local optimum.

[0191] (2) Explorer: Explorers play a crucial role in venturing into unknown territories to discover global optimal solutions. In the embodiments of the present application, the top 40% to 80% of the pre-adapted individuals in the population are designated as explorers, and the explorer strategy is used to update the position and add a Gaussian mutation strategy. Explorers are brave to enter unknown territories, and the formula is as follows:

[0192]

[0193] Where, is the position of the ith individual in the t+1th generation, is the position of the ith individual in the tth generation, is the least adapted individual in the tth generation population, represents a Gaussian random variable with mean 0 and variance .

[0194] When the position X i of the individual is significantly different from the position X worst of the worst individual, the update magnitude is large, encouraging the individual to move away from the current poor solution and move towards a better solution.

[0195] (3) Follower: In the embodiments of the present application, the top 80% to 90% of the pre-adapted individuals in the population are designated as followers, and the Gaussian mutation strategy is applied, and the formula is as follows:

[0196]

[0197] Where, represents the position of the most adapted individual in the human population at the tth iteration, Rd represents a random number in the range of [1, dim], and the introduction of the random number Rd ensures that each update is not completely deterministic, thereby avoiding the concentration of all individuals in one position and maintaining the diversity of the population, represents a Gaussian random variable with mean 0 and variance .

[0198] (4) Losers: In this embodiment, the last 10% of individuals in the pre-adapted population are designated as losers. For each loser i in the population, those individuals with poor adaptability who remain in the population are called losers. These losers with poor adaptability who cannot integrate into society will be eliminated, and the population will be replenished by reproduction in areas suitable for human development. The population replenishment equation after the Gaussian mutation strategy is improved as follows:

[0199]

[0200] Among them, X i+1 X represents the location of newly generated individuals in the population. best X represents the location of the individual with the highest fitness in the population. i Let Rn be the position of a loser individual in the population, Rn be a random number following a normal distribution, and σ be a parameter controlling the amplitude of Gaussian variation, called the standard deviation. It follows a standard normal distribution.

[0201] Step S24: After each iteration, check if the termination condition is met, such as reaching the maximum number of iterations or the objective function value converging. If the condition is met, stop the algorithm and output the optimal solution.

[0202] Step S3: Optimize the parameters of the SOFTS model using the IHEOA algorithm.

[0203] In this embodiment, the key parameters of the SOFTS model are optimized using the IHEOA algorithm to improve the model's predictive performance and generalization ability. The optimized parameters include the embedding dimension d, the core representation dimension d′, the number of encoder layers N, the prediction window length H, and the learning rate L.

[0204] Specifically, it includes the following sub-steps:

[0205] Step S31: Define the objective function f(θ) based on the mean squared error of the SOFTS model, where θ is the set of parameters to be optimized, including the embedding dimension d, the core representation dimension d′, the number of encoder layers N, the prediction window length H, and the learning rate L. The formula for the objective function f(θ) is as follows:

[0206]

[0207] Where N is the number of samples, Y i This is the actual value. These are model predictions.

[0208] It is understandable that minimizing the objective function f(θ) ensures that the SOFTS model minimizes its prediction error on the training data, thereby improving model performance.

[0209] At initialization, an initial parameter set θ = {d, d', N, H, L} is generated by the population initialization method of the IHEOA algorithm. The initial parameter set provides an initial solution for subsequent optimization iterations.

[0210] At step S32, the input features are predicted by the SOFTS model, and the prediction formula is as follows:

[0211]

[0212] where X i is the input feature, and SOFTS(θ, X i ) is the prediction result of the SOFTS model based on the parameter set θ. This prediction model provides the basis for subsequent fitness calculation and parameter optimization.

[0213] At step S33, in each iteration, the exploration phase and development phase of the IHEOA algorithm are applied to update the parameter set θ. The exploration phase is used to search for the global optimal solution, and the development phase is used for local optimization. In each iteration, the algorithm adjusts the parameter set through a Gaussian mutation strategy, and adjusts the position of the solution combined with fitness calculation, thereby improving the optimization effect.

[0214] At step S34, the performance of the SOFTS model corresponding to each parameter set is evaluated by calculating the current fitness. The fitness calculation is based on the objective function f(θ), i.e., the prediction error of the current parameter set on the training data. According to the fitness value, the optimal parameter set is updated to ensure that the optimal parameter solution is always retained during the optimization process.

[0215] At step S35, when the preset termination condition is reached (e.g., the maximum number of iterations is reached or the objective function value converges), the algorithm terminates and outputs the optimal parameter set, and applies the optimized parameter set to the SOFTS model for the final prediction task of the model, to improve the prediction accuracy and generalization ability.

[0216] At step S4, the IHEOA-SOFTS ventilation system model is constructed, and the relationship between the air volume opening and the indoor temperature and energy consumption is established.

[0217] In the embodiments of the present application, the IHEOA-SOFTS ventilation system model is constructed to establish the mapping relationship between the opening of each fan and the indoor temperature and total energy consumption. The model takes the opening of each fan, the external temperature and humidity as input, and the temperature at each fan and the total energy consumption as output, providing an accurate prediction basis for fan parameter optimization and system energy saving.

[0218] Specifically, the following sub-steps are included:

[0219] At step S41, the preprocessed historical operation data is input as input, and the input data set is represented as Xt = {A t , T ext,t , H ext,t}, where X t is the total input of the model at time step t, A t is the opening of each fan at time step t, T ext,t is the external temperature at time step t, H ext,t is the external humidity at time step t.

[0220] Step S42, in the ventilation system model, the input data X t = {A t , T ext,t , H ext,t} is embedded into a high-dimensional space by an embedding layer for subsequent processing. The output feature representation of the embedding layer is represented as:

[0221] S0 = Embedding(X t )

[0222] where S0 is the output feature representation of the embedding layer.

[0223] The purpose of the embedding layer is to convert the original input data into a high-dimensional vector representation so that the model can better capture the complex relationships between input features.

[0224] Step S43, input the output feature representation of the embedding layer into the STAR module to capture the long and short-term dependencies of time series data layer by layer. The output of the STAR module is represented as:

[0225] S i = STAR(S i-1 )

[0226] where S i represents the feature representation of the i-th layer output in the STAR module, and S i-1 represents the output feature representation of the i-1-th layer in the STAR module.

[0227] Step S44, the output feature representation of the i-1-th layer of the STAR module is processed by a multi-layer perception to extract the core representation, which is as follows:

[0228] o i = MLP(S i-1 )

[0229] where o i is the core representation of the i-th layer, which is used to further capture the core information of the features.

[0230] Step S45, repeat and concatenate the core representation and the corresponding feature representation to generate a combined feature representation, and the specific formula is:

[0231] F i = Repeat_Concat(S i-1 , o i )

[0232] Wherein, F i is the combined feature of the feature representation input to the i-th layer of the multi-layer perception and the output core representation.

[0233] Step S46, the combined feature representation is processed through the encoder layer, and the original feature is combined with the processed feature through the residual connection to enhance the expression ability of the model. The encoder layer output is represented as:

[0234] S i = MLP(F i ) + S i-1

[0235] Wherein, S i is the feature representation output by the encoder layer.

[0236] Step S47, the final output feature representation S L output by the encoder layer is input to the linear layer, and the prediction result is output, represented as:

[0237]

[0238] Wherein, represents the prediction result, S L represents the final feature representation obtained by layer-by-layer processing through the multi-layer neural network, represents the temperature at each fan at the prediction time step t, E total,t represents the total power consumption at the prediction time step t.

[0239] Through the above steps, the IHEOA-SOFTS ventilation system model can accurately predict the relationship between each fan opening and indoor temperature and total power consumption, providing efficient support for energy-saving optimization of the ventilation system.

[0240] In application, by inputting the air volume opening, external temperature and humidity data through the trained model, the temperature at each fan and the total power consumption can be predicted, and the formula is:

[0241]

[0242] Wherein, represents the predicted temperature at each fan at the t+k step, θ represents the total power consumption of the prediction of the t+kth step * θ represents the total power consumption of the prediction of the t+kth step

[0243] Step S5, a comprehensive objective function considering indoor temperature control accuracy and total power consumption is established, and the opening of each fan in the IHEOA-SOFTS ventilation system model is continuously optimized and adjusted by using the IHEOA algorithm to minimize the comprehensive objective function, so as to determine the optimal relationship between the fan opening and the temperature in each place in the room and the total power consumption.

[0244] In the embodiments of the present application, the key parameters (fan opening) of the IHEOA-SOFTS ventilation system model are optimized by the IHEOA intelligent algorithm to achieve the optimal energy saving and control effect.

[0245] The process mainly includes the following sub-steps:

[0246] Step S51, first, the comprehensive objective function f(θ) is defined by combining the indoor temperature control accuracy and the total power consumption. The expression of the objective function is as follows:

[0247]

[0248] Wherein, T target is the target temperature, is the mean square error of temperature control, is the total power consumption, w1 and w2 are weight coefficients, which represent the weights of temperature control and energy consumption respectively, and the specific values can be adjusted according to actual requirements.

[0249] Step S52, the initial population is initialized by using the IHEOA intelligent algorithm to generate the initial fan opening solution set. Each solution is represented as a combination of fan openings, and the population initialization formula is as follows:

[0250]

[0251] Wherein, N is the population size, and the population serves as the initial input of the optimization process, providing a basis for subsequent fitness calculation.

[0252] Step S53, after initializing the population, the fitness value of each solution in the population is calculated The fitness value is determined by the comprehensive objective function The specific steps are as follows: the temperature and total power consumption at each fan are predicted by using the ventilation system model according to the input parameters, and the prediction formula is as follows:

[0253]

[0254] After the calculation of the fitness value is completed, the input for the next step of the screening process is provided.

[0255] Step S54, after calculating the fitness value, the population is screened according to the fitness value, and the better solutions are selected to form a new candidate population. The screened population is represented as:

[0256] P' = Select (P)

[0257] The screened solutions have higher fitness values, and have a greater possibility of generating new solutions that meet the optimization target. Based on the screened population, a crossover operation is performed to generate new candidate solutions, and the crossover formula is:

[0258] P" = Crossover (P')

[0259] This operation generates new solutions by combining the characteristics of better solutions, providing diversity for the subsequent optimization process.

[0260] Step S55, in order to further increase the diversity of the population, a mutation operation is performed on the new solutions generated by the crossover. The mutation operation adjusts each candidate solution slightly to avoid the population falling into a local optimal solution. The mutated population is represented as:

[0261] P'" = Mutate (P")

[0262] This step ensures that the optimization process can explore a larger search space, thereby improving the global search ability of the algorithm.

[0263] Step S56, the mutated solutions are fused with the current population to generate an updated population, and the update formula is:

[0264] P = Update (P, P"')

[0265] Step S57, when the predetermined number of iterations or the target function value converges, the algorithm is terminated and the final optimized fan opening parameter set is output, represented as:

[0266]

[0267] Step S58, finally, the optimized fan opening control the ventilation system, adjust the relationship between the fan opening and the temperature in each place and the total power consumption, realize the optimization control of the fan parameters and the energy saving optimization of the system, and achieve the goal of system energy saving and precise control.

[0268] In order to realize the above embodiment, the application also provides a ventilation system energy saving optimization system based on neural network. Figure 3 A structure diagram of a ventilation system energy saving optimization system 10 based on neural network provided by the embodiment of the application is shown in FIG. 1. As shown in the figure, the system includes: Figure 3

[0269] ​The preprocessing module 100 is used for collecting historical operation data of the fan, and performing mean value interpolation and wavelet transform filtering preprocessing on the historical operation data; the historical operation data includes the opening degree, external temperature, humidity, temperature at the fan and total energy consumption of each fan;

[0270] The first optimization module 200 is used for optimizing the human evolution optimization algorithm HEOA by adopting a Gaussian variation strategy to obtain an improved human evolution optimization algorithm IHEOA;

[0271] The second optimization module 300 is used for optimizing the parameters of the SOFTS model by the IHEOA algorithm; the parameters include embedding dimension, core representation dimension, number of encoder layers, prediction window length and learning rate;

[0272] The model construction module 400 is used for constructing an IHEOA-SOFTS ventilation system model to establish the relationship between the air volume opening degree and indoor temperature and energy consumption; the IHEOA-SOFTS ventilation system model takes the opening degree, external temperature and humidity of each fan as input, and takes the temperature at each fan and total energy consumption as output;

[0273] The optimal relationship output module 500 is used for establishing a comprehensive objective function considering indoor temperature control accuracy and total power consumption, and continuously optimizing and adjusting the opening degree of each fan in the IHEOA-SOFTS ventilation system model by the IHEOA algorithm to minimize the comprehensive objective function, so as to determine the optimal relationship between the fan opening degree and indoor temperature and total energy consumption.

[0274] As to the system in the above embodiment, the specific manner in which the various modules perform operations has been described in detail in the embodiment related to the method, and will not be described in detail here.

[0275] In order to realize the above embodiment, the present application further provides an electronic device, comprising a processor and a memory connected with the processor in communication; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to realize the method provided by the foregoing embodiment.

[0276] In order to realize the above embodiment, the present application further provides a computer readable storage medium, the computer readable storage medium stores computer execution instructions, and the computer execution instructions are executed by the processor to realize the method provided by the foregoing embodiment.

[0277] In order to realize the above embodiment, the present application further provides a computer program product, comprising a computer program, and the computer program is executed by the processor to realize the method provided by the foregoing embodiment.

[0278] It should be understood that the various forms of flow shown above can be used to reorder, add, or remove steps. For example, the steps recited in the present application can be performed in parallel, in series, or in a different order, as long as the desired results of the present application are achieved, which is not limited herein.

[0279] The foregoing detailed description has not described the limitations of the scope of protection of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A method for energy-saving optimization of ventilation systems based on neural networks, characterized in that, Includes the following steps: Collect historical operating data of the wind turbine, and perform mean interpolation and wavelet transform filtering preprocessing on the historical operating data; The historical operating data includes the opening degree of each fan, external temperature, humidity, temperature at the fan, and total energy consumption; The human evolutionary optimization algorithm HEOA is optimized using a Gaussian mutation strategy, resulting in an improved human evolutionary optimization algorithm IHEOA. The parameters of the SOFTS model are optimized using the IHEOA algorithm; these parameters include embedding dimension, core representation dimension, number of encoder layers, prediction window length, and learning rate. An IHEOA-SOFTS ventilation system model was constructed to establish the relationship between air volume opening degree, indoor temperature, and energy consumption. The IHEOA-SOFTS ventilation system model takes the opening degree of each fan, external temperature, and humidity as inputs, and the temperature at each fan and total energy consumption as outputs. A comprehensive objective function considering indoor temperature control accuracy and total power consumption is established. With the goal of minimizing the comprehensive objective function, the opening degree of each fan in the IHEOA-SOFTS ventilation system model is continuously optimized and adjusted using the IHEOA algorithm to determine the optimal relationship between the fan opening degree and the indoor temperature and total power consumption. The improved human evolutionary optimization algorithm IHEOA, obtained by optimizing the HEOA algorithm using a Gaussian mutation strategy, includes: To simulate the chaotic stage of early human evolution, a logical chaotic mapping method is used to initialize the population. The calculation formula is as follows: in, Representing the The value of the next iteration. Represents the value of the previous iteration. It refers to population size. For the control parameters of the chaotic mapping; to set the chaotic values Mapping to the search space, we obtain the location of the initial population, calculated using the following formula: in, Indicates the first The initial position of the individual, and These are the lower and upper limits of the search space, respectively; In each iteration of the human evolutionary optimization algorithm, the fitness of each individual in the population is calculated, and a Gaussian mutation strategy is applied to generate a normally distributed random number perturbation, so that the new solution is fine-tuned in the vicinity of the original solution. The update formula is as follows: in, It is a new solution after mutation. It is the original solution before the mutation. This indicates that the mean is 0 and the variance is . The Gaussian distributed random numbers; during the human exploration phase, individuals search using a Gaussian mutation strategy, and the formula for updating positions is as follows: in, It is an adaptive function. It is the dimension of the problem. and They represent the first The current position and the updated next position of each individual. It is the best location explored to date. It is the average position of the current population. Represents the Levy distribution. It is the jump coefficient. It is a random number in the range [0,1], where t is the current algebra. Maximum number of iterations; In the stages of human development, the population is divided into leaders, explorers, followers and losers. A Gaussian mutation strategy is applied to each role and the position is updated using the corresponding strategy. Check the termination condition. If the termination condition is met, stop the iteration and output the optimal solution. The optimization of the SOFTS model parameters using the IHEOA algorithm includes: The objective function is defined based on the mean squared error of the SOFTS model. ,in, It is the set of parameters to be optimized, including the embedding dimension. Core representation dimension Number of encoder layers Prediction window length Learning rate The objective function The formula is as follows: in, It is the sample size. This is the actual value. These are model predictions; During initialization, the initial parameter set is generated using the population initialization method of the IHEOA algorithm. ; Define the prediction model using the following formula: in, These are input features. The SOFTS model is based on a set of parameters. The prediction results; In each iteration, during the exploration and development phase of the IHEOA algorithm, the parameter set is updated. ; By calculating the current fitness, the performance of the SOFTS model corresponding to each parameter set is calculated, and the optimal parameter set is updated. Once the termination condition is met, the optimal set of parameters is output and applied to the SOFTS model.

2. The method according to claim 1, characterized in that, The preprocessing of the historical running data by mean interpolation and wavelet transform filtering includes: The historical data is imputed by mean, missing values ​​are identified, the mean of all non-missing values ​​is calculated, and the calculated mean is used to fill in the missing value positions. The calculation formula is as follows: in, For non-missing values, It is the total number of non-missing values. The mean of the positions used to fill in missing values; Wavelet transform filtering was performed on the mean-interpolated dataset, and the Daubechies wavelet was selected as the wavelet function. Perform a discrete wavelet transform on the filtered data to obtain the wavelet coefficients. The calculation formula is as follows: in, In scale and location Wavelet coefficients on It is a scale. It is a position parameter. It is the location index of the original data. It is the scaled and translated wavelet function. Scale and translation are achieved by scaling the original wavelet function... Scaling Double and translate To achieve this, For the original data in location The value on; The wavelet coefficients are subjected to hard thresholding to remove noise and unimportant information, as shown in the following formula: in, It is a set threshold, if the wavelet coefficients The absolute value is greater than the threshold If the coefficient is positive, retain it; otherwise, set it to 0. The data is reconstructed using the wavelet coefficients after thresholding, as shown in the following formula: in, For the reconstructed data in location The value on, by using all the reconstructed wavelet coefficients With the corresponding wavelet function Multiply, and for all scales and location We obtain the result by summing. The reconstructed data is represented as follows: in It is the length of the signal. This is the data after overall reconstruction.

3. The method according to claim 1, characterized in that, In the stages of human development, the population is divided into leaders, explorers, followers, and losers. A Gaussian mutation strategy is applied to each role, and positions are updated using the corresponding strategy, including: For each leader in the population Applying the Gaussian mutation strategy, the formula is as follows: in, It is a new solution after mutation. It is the original solution before the mutation. This indicates that the mean is 0 and the variance is . Gaussian random variables; The top 40% of individuals in pre-adaptation within the population are designated as leaders. A leader strategy is used to update positions, and a Gaussian mutation strategy is applied, as shown in the following formula: in, For the first The generation The location of each individual For the first The generation The location of each individual Represents random numbers that follow a normal distribution, the function This generates a row vector containing the element dim, where each element is set to 1. This is a random number in the range [0,1], representing the complexity of the situation related to the leader. This is a situation assessment value used to determine whether the leader should change their search strategy. The difficulty coefficient of knowledge acquisition gradually decreases over time, and its expression is: Individuals ranking in the top 40% to 80% of prefitting within the population are designated as explorers. The explorer strategy is used to update their positions, and a Gaussian mutation strategy is applied, as shown in the following formula: in, For the first The generation The individual's position, For the first The generation The individual's position, It is in the The least adapted individual in the population, This indicates that the mean is 0 and the variance is . Gaussian random variables; Individuals ranking in the top 80% to 90% of prefitting within the population are designated as followers, and a Gaussian mutation strategy is applied, as shown in the following formula: in, Representing the The position of the most adaptable individual in the human population at the next iteration. express Random numbers within a range This indicates that the mean is 0 and the variance is . Gaussian random variables; By designating the bottom 10% of prefitted individuals in the population as losers, the population replenishment equation, after Gaussian mutation strategy, becomes: in, The location of newly generated individuals in the population. This represents the position of the individual with the highest fitness in the population. The position of the loser individual in the population. Represents a random number that follows a normal distribution. The parameter that controls the magnitude of Gaussian variation is called the standard deviation. It follows a standard normal distribution.

4. The method according to claim 1, characterized in that, The prediction process of the IHEOA-SOFTS ventilation system model includes: Input the preprocessed dataset into the ventilation system model ,in, In order to time step Total input of the time model In order to time step The opening degree of each fan at that time. In order to time step external temperature at that time In order to time step External humidity at that time; In the ventilation system model, the input data is embedded into a high-dimensional space through an embedding layer, as shown in the formula: in, This represents the output features of the embedding layer. The output feature representation of the embedding layer is input into the STAR module, and the first feature is obtained through time series processing. The output feature representation of the layer is given by the formula: in, Indicates the first in the STAR module Feature representation of the layer output Indicates the first in the STAR module The output feature representation of the layer; Through multilayer perceptron, the first The output feature representation of the layer is processed to extract the core representation, as shown in the following formula: in, For the first The core representation of a layer; By repeating and concatenating the core representation and feature representation, a combined feature representation is generated, as shown in the formula: in, For input to the first The combined features of the feature representation and the core representation of the output of a multilayer perceptron; The combined feature representation is processed by an encoder layer, and the original features and the processed feature representation are combined by residuals, as shown in the following formula: in, The feature representation of the encoder layer output; The feature representation output from the encoder layer is input into the linear layer, and the output prediction result is represented as follows: in, Indicates the prediction result. This represents the final feature representation obtained after processing layer by layer through a multi-layer neural network. Indicates the prediction time step The temperature at each fan. Indicates the prediction time step Total power consumption during operation.

5. The method according to claim 4, characterized in that, The process involves establishing a comprehensive objective function that considers indoor temperature control accuracy and total power consumption. To minimize this comprehensive objective function, the IHEOA algorithm is used to continuously optimize and adjust the opening degree of each fan in the IHEOA-SOFTS ventilation system model, determining the optimal relationship between fan opening degree and indoor temperature and total power consumption. This includes: Establish a comprehensive objective function The comprehensive objective function Taking into account both indoor temperature control accuracy and total power consumption, the following definition is provided: in, For the target temperature, This represents the mean square error of temperature control. That is the total power consumption. These are weighting coefficients used to balance the influence of different objectives; The population is initialized using the IHEOA intelligent algorithm to generate an initial set of wind turbine opening solutions. Each solution is represented as a combination of wind turbine opening degrees, and the population initialization formula is as follows: For each solution in the population Calculate its fitness value, which is determined by the comprehensive objective function. The decision is made using the following formula: The population is screened based on fitness values, and better solutions are selected to form a new candidate population. The screened population is represented as follows: Based on the selected population, a crossover operation is performed to generate new candidate solutions. The crossover formula is as follows: Perform a mutation operation on the new solutions generated by the crossover, make minor adjustments to each candidate solution, and update the formula as follows: The mutated solution is merged with the current population to generate an updated population. The update formula is: When the predetermined number of iterations is reached or the objective function value converges, the algorithm terminates and outputs the final optimized set of wind turbine opening parameters, represented as: Use the optimized fan opening Control the ventilation system, adjust the relationship between fan opening and indoor temperature and total power consumption, and achieve optimized control of fan parameters and energy-saving optimization of the system.

6. A neural network-based energy-saving optimization system for ventilation systems, characterized in that, include: The preprocessing module is used to collect historical operating data of the wind turbine and perform mean interpolation and wavelet transform filtering preprocessing on the historical operating data; The historical operating data includes the opening degree of each fan, external temperature, humidity, temperature at the fan, and total energy consumption; The first optimization module is used to optimize the human evolutionary optimization algorithm HEOA using a Gaussian mutation strategy, resulting in an improved human evolutionary optimization algorithm IHEOA. The second optimization module is used to optimize the parameters of the SOFTS model using the IHEOA algorithm; the parameters include embedding dimension, core representation dimension, number of encoder layers, prediction window length, and learning rate. The model building module is used to build the IHEOA-SOFTS ventilation system model and establish the relationship between air volume opening degree and indoor temperature and energy consumption. The IHEOA-SOFTS ventilation system model takes the opening degree of each fan, external temperature and humidity as inputs, and the temperature at each fan and total energy consumption as outputs. The optimal relationship output module is used to establish a comprehensive objective function that considers indoor temperature control accuracy and total power consumption. With the goal of minimizing the comprehensive objective function, the IHEOA algorithm is used to continuously optimize and adjust the opening degree of each fan in the IHEOA-SOFTS ventilation system model to determine the optimal relationship between the fan opening degree and the indoor temperature and total power consumption. The improved human evolutionary optimization algorithm IHEOA, obtained by optimizing the HEOA algorithm using a Gaussian mutation strategy, includes: To simulate the chaotic stage of early human evolution, a logical chaotic mapping method is used to initialize the population. The calculation formula is as follows: in, Representing the The value of the next iteration. Represents the value of the previous iteration. It refers to population size. For the control parameters of the chaotic mapping; to set the chaotic values Mapping to the search space, we obtain the location of the initial population, calculated using the following formula: in, Indicates the first The initial position of the individual, and These are the lower and upper limits of the search space, respectively; In each iteration of the human evolutionary optimization algorithm, the fitness of each individual in the population is calculated, and a Gaussian mutation strategy is applied to generate a normally distributed random number perturbation, so that the new solution is fine-tuned in the vicinity of the original solution. The update formula is as follows: in, It is a new solution after mutation. It is the original solution before the mutation. This indicates that the mean is 0 and the variance is . The Gaussian distributed random numbers; during the human exploration phase, individuals search using a Gaussian mutation strategy, and the formula for updating positions is as follows: in, It is an adaptive function. It is the dimension of the problem. and They represent the first The current position and the updated next position of each individual. It is the best location explored to date. It is the average position of the current population. Represents the Levy distribution. It is the jump coefficient. It is a random number in the range [0,1], where t is the current algebra. Maximum number of iterations; In the stages of human development, the population is divided into leaders, explorers, followers and losers. A Gaussian mutation strategy is applied to each role and the position is updated using the corresponding strategy. Check the termination condition. If the termination condition is met, stop the iteration and output the optimal solution. The optimization of the SOFTS model parameters using the IHEOA algorithm includes: The objective function is defined based on the mean squared error of the SOFTS model. ,in, It is the set of parameters to be optimized, including the embedding dimension. Core representation dimension Number of encoder layers Prediction window length Learning rate The objective function The formula is as follows: in, It is the sample size. This is the actual value. These are model predictions; During initialization, the initial parameter set is generated using the population initialization method of the IHEOA algorithm. ; Define the prediction model using the following formula: in, These are input features. The SOFTS model is based on a set of parameters. The prediction results; In each iteration, during the exploration and development phase of the IHEOA algorithm, the parameter set is updated. ; By calculating the current fitness, the performance of the SOFTS model corresponding to each parameter set is calculated, and the optimal parameter set is updated. Once the termination condition is met, the optimal set of parameters is output and applied to the SOFTS model.

7. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-5.

Images