Building energy-saving intelligent integrated control system and method

Abstract

The invention belongs to the technical field of intelligent control, and discloses a building energy-saving intelligent integrated control system and method. Comprising the steps of collecting building comprehensive data of a building internal space area; preprocessing the building comprehensive data to obtain standard building data; establishing a building dynamic model according to the standard building data, and predicting by using the dynamic model to obtain an energy consumption change curve and a comfort change curve of the internal space area of the building in a future fixed time period; solving to obtain a series of optimal control strategy sequences by taking the energy consumption change curve and the comfort change curve of the internal space area of the building in the future fixed time period as constraints; the optimal control strategy sequence obtained through solving is distributed to all actuators, and integrated intelligent control over heating equipment, refrigeration equipment, ventilation equipment, illumination equipment and the like is achieved; not only is the energy-saving level of the building improved, but also a comfortable environment is created for the indoor space, and the comprehensive performance of the building is comprehensively enhanced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control technology, and more specifically, to a building energy-saving intelligent integrated control system and method. Background Art

[0002] The patent with application publication number CN118605146A discloses a smart building energy-saving control system, method and electrical equipment. By constructing a three-dimensional topological network based on IP addresses, it can more clearly display the device form, overall structure and topological relationship, which to a certain extent helps to better regulate each controlled device and reduce unnecessary energy consumption. The energy-saving control server determines the adjustment degree of each controlled object according to the energy consumption and actual operating variables of each controlled object, so that each intelligent node control device adjusts the corresponding controlled device to make each controlled object reach the optimal operating state, thereby helping to reduce energy waste as a whole and achieve the purpose of energy saving.

[0003] However, the existing system still lacks the ability to fully perceive the internal environmental status and energy consumption of the building; due to the poor quality of the collected data, it is impossible to accurately grasp the operating status of the building, making it difficult to formulate a scientific and reasonable control strategy; secondly, the existing system has limited modeling and prediction capabilities; they usually use simple empirical models or data-driven models, which cannot accurately describe the complex dynamic processes inside the building, and the predicted energy consumption and comfort changes often deviate from the actual situation, which restricts the foresight and effectiveness of the control strategy; thirdly, when solving the control strategy, the existing system often only considers a single goal and ignores the internal contradictions and balance; this single-goal optimization often cannot truly meet actual needs; in addition, the existing system often uses simple search or rule methods in the solution process, and the optimization ability is limited; due to the high complexity and variability of the building system, it is difficult to obtain the truly optimal control strategy, and the control effect is difficult to achieve an ideal level; it can be seen that the existing building energy-saving control system has obvious deficiencies in many links, which restricts the overall performance of the system and makes it difficult to truly meet actual needs in actual application scenarios.

[0004] In view of this, the present invention proposes a building energy-saving intelligent integrated control system and method to solve the above problems. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solutions: an intelligent integrated control system for building energy saving, comprising: a data acquisition module for collecting comprehensive building data of the internal space area of ​​the building; pre-processing the comprehensive building data to obtain standard building data; Comprehensive prediction module, used to establish a building dynamic model based on standard building data, and use the dynamic model to predict the energy consumption change curve and comfort change curve of the building's internal space area in a fixed time period in the future; The strategy solving module is used to use the energy consumption change curve and comfort change curve of the internal space area of ​​the building in a fixed time period in the future as constraints to solve a series of optimal control strategy sequences; the solved optimal control strategy sequences are distributed to each actuator of the building, and each module is connected by wired and / or wireless means to realize data transmission between modules.

[0006] Furthermore, the comprehensive building data includes environmental parameters and energy consumption data within the building; the environmental parameters within the building include temperature, humidity, carbon dioxide concentration, illumination and fresh air volume; the energy consumption data includes heating capacity, cooling capacity and lighting power.

[0007] Furthermore, the method of preprocessing the building comprehensive data includes: The comprehensive building data is cleaned and the data of different dimensions are dimensionless processed to obtain preliminary building data; the preliminary building data is arranged into corresponding data time series according to the chronological order of data types; the data time series is segmented to obtain m segments of building data; standard building data are extracted from the m segments of building data.

[0008] Furthermore, the data segmentation method includes: For each data time series, regard it as a state sequence of a discrete-time Markov chain, and discretize it into a finite number of states according to the range of the data values ​​of the data time series, that is, determine the state space of the Markov chain; initialize a counting matrix C, and the initial value of all elements in the counting matrix C is 0, traverse the time series data, and for each adjacent time point pair (Y_t, Y_t+1), where Y_t represents the state at time t, and Y_t+1 represents the state at time t+1, determine the state indexes i and j corresponding to Y_t and Y_t+1 respectively; add 1 to the value of the element C(i, j) in the counting matrix C corresponding to the state index i and j, that is, C(i, j) records the number of transitions from state z_i to state z_j; after traversing the entire time series, for each state s_i, calculate its frequency of transitions to other states, that is, C(i, j) divided by the sum of the number of all transitions starting from state z_i; all the frequencies obtained are all the elements of the transition probability matrix P; Using the obtained transition probability matrix, for each time point t in the data time series, calculate the conditional probability P(t, z_i) that the data value at that time point belongs to each state z_i; For each time point t, calculate the maximum value P_max of the conditional probability that the data value at that time point belongs to each state; Based on the maximum value P_max, a probability threshold θ=a×P_max is preset. When P(t, z_i) is greater than or equal to θ, it is determined that the time point t belongs to the state z_i; a is a weighting parameter; Scan the entire data time series and classify adjacent time points belonging to the same state into a data segment, that is, obtain several segments of preliminary segmented data; perform cluster analysis on several segments of preliminary segmented data to obtain m segments of building data.

[0009] Furthermore, the cluster analysis method includes: Calculate the feature vector of each preliminary segmented data, the feature vector is obtained by concatenating the weighted mean vector, variance vector, mean vector, maximum value, minimum value, kurtosis and skewness; The weighted mean vector is calculated by: For Preliminary segmented data, the time range is [t_l _ start,t_l _ end] ,in, For the The starting time point of the preliminary segmented data, For the The end time point of the preliminary segmented data; based on the end time point Define the time weight function; the formula of the time weight function is: ;in, is an attenuation coefficient, is the periodic adjustment coefficient, For time The time weight of is the initial time point of the function, For the cycle; Based on the time weight function, the The weighted mean vector of the preliminary segmented data ; ;in, For data points The confidence weight of the data point For preliminary segmentation of data in time data points; ,in, is the mean of the preliminary segmented data, for The standard deviation of the neighborhood, and is the confidence parameter; Based on the calculated eigenvectors, a distance matrix D is constructed, and the element D(I, J) in the distance matrix D represents the Mahalanobis distance between the I-th preliminary segmented data and the J-th preliminary segmented data; based on the distance matrix D, a merge iteration is performed to obtain m segments of building data.

[0010] Furthermore, the method of performing merging iteration based on the distance matrix D includes: Each preliminary segmented data in the distance matrix D is regarded as a separate cluster, the feature vector of each cluster is calculated, and the comprehensive distance between clusters is calculated based on the feature vector; the calculation formula of the comprehensive distance is: ;in, Cluster and The combined distance between Cluster The characteristic vector of Cluster The characteristic vector of is an adaptive covariance matrix, Cluster The time center, It is a cluster The time center, is the time decay parameter; The merging cost between clusters is calculated based on the comprehensive distance between clusters; the calculation formula of the merging cost is: ;in, is the cluster shape penalty, Cluster and The merger cost between Cluster shape penalty ;in, Cluster The variance of Cluster The variance of Cluster The skewness of Cluster The skewness of and is a learnable parameter; Among all the current pairwise clusters, find the two clusters with the lowest merging cost. and , merge them into new clusters ; New cluster The eigenvector of and The feature vectors of and their sample numbers are weighted and calculated; calculate the new cluster The distance to other clusters and update the distance matrix; When the sum of the squares of the intra-cluster deviations of all clusters is less than the preset merging threshold, the merging is stopped, and finally m clusters corresponding to m segments of building data are obtained.

[0011] Furthermore, the method of establishing the building dynamic model includes: The internal space of the building is discretized into a number of lattice units. Each lattice unit corresponds to a magnetic spin. The magnetic spin can be in two states, up or down, representing the two states of the lattice unit. There are several standard building data in each grid unit; based on the standard building data, define the adjacent grid units and The coupling strength between ;in, For grid cells and grid cells The distance between the centers, For grid cells Any standard building data in For grid cells Any standard building data in For grid cells and grid cells The relative azimuth between is the anisotropy parameter, is the phase parameter; is the exponential parameter; is the intensity adjustment parameter, is the intensity adjustment index parameter; is the distance index parameter; The external field items are defined using environmental parameters and energy consumption data; i.e. ; For grid cells The outfield item, For grid cells The corresponding environmental parameter vector, For grid cells The corresponding energy consumption data vector; is the mapping function; The standard building data of each grid unit is randomly divided into a training set and a test set according to a certain ratio, and the data in the training set is used as the training data; the loss function is to minimize the mean square error between the output and the training data; the standard building data of the pth training data is defined as X_p, and the corresponding real environmental parameters and energy consumption data are obtained; the output is obtained by simulating time evolution using the current coupling parameters; the coupling parameters are initialized to a set of random values, and the training data is iterated several times. The iterative process is: randomly sample a small batch of training data from the training set, calculate the loss function for this small batch of training data, and calculate the gradient of the loss function to the coupling parameters, and use stochastic gradient descent or its variant algorithm to update the coupling parameters according to the gradient; until the preset maximum number of iterations is reached, the optimal coupling parameters are obtained, and the established building dynamic model is obtained.

[0012] Furthermore, the method of using the dynamic model to predict the energy consumption change curve and the comfort change curve of the interior space area of ​​the building within a fixed time period in the future includes: Obtain the current value of the standard building data, and determine the initial state of each grid unit according to the current value of the standard building data; and form the initial states of all grid units into a state vector; In time , according to the current state vector and the optimal coupling parameters, and calculate each grid cell The probability of flipping spin ; ;in, For grid cells In time The spin state of For grid cells In time The spin state of is the random disturbance term; According to the probability of flipping spin , using Monte Carlo sampling, corresponding to the grid unit Generate a uniform random number in the interval [0, 1] ;like , then the grid unit In time The spin state Set to , that is, the spin flips, otherwise the spin state remains unchanged; all lattice units are The spin state of the new state vector is formed; repeat until the time reaches the predicted maximum time step , and obtain a series of state vectors; For each time step, the state vector corresponds to a set of environment parameter vectors and energy consumption data vector , through the reverse mapping function Convert to and ; Environmental parameter vector At the maximum time step All sampling points in the environment parameter change curve constitute the environmental parameter change curve; for each time step, the corresponding comfort index is calculated to form the comfort change curve; the comfort index is the percentage of dissatisfaction ; Using energy consumption data vector At the maximum time step The total energy consumption corresponding to all sampling points in , which constitutes the energy consumption change curve.

[0013] Furthermore, the solution method of the optimal control strategy sequence includes: The optimization objectives are defined as minimizing energy consumption and maximizing comfort, that is, minimizing the total energy consumption in the future and minimizing the comfort index; the optimization objectives are expressed as two objective functions and ; ; ; Define optimization variables as control parameters; define constraints including energy consumption change curve The peak value does not exceed the preset energy consumption limit and the peak and valley ranges of the comfort change curve are all within the preset comfort threshold range; Initialize a strategy population, determine the strategy population size N2 and the dimension n2 of the optimization variable; for each individual in the strategy population, randomly generate a set of optimization variables as the individual's position vector; The two objective functions and The weighted sum of is used as the update function of the strategy population; calculate the individual The value of the update function , that is, the update function is The function value at is recorded as the updated value; Calculate the centroid of the current population As a potential optimal solution; among them, is the position vector of individual d, for each individual , calculate individual With the centroid The Euclidean distance and the unit vector ; For each individual , generates a random step size that follows a normal distribution , update the individual according to the random step size and unit vector The position vector of individual New position vector ; and calculate the new position vector The value of the update function ; ;in, is a constant parameter, is a random variable that follows a specific distribution; Sort the updated values ​​from small to large to obtain the sorted strategy population, and divide the sorted strategy population into two parts, the first part is better individuals, and the second part is worse individuals; The number of better individuals is ;in, and To dynamically adjust the parameters, is the current iteration number; is the rounding function; For each worse individual , randomly select a better individual , calculate the Euclidean distance between the two, recorded as the gap distance ; Update the worse individual according to the gap distance The position vector , get a worse individual New position vector ; Calculate the worse individual New position vector The updated value of ;in, For better individuals and worse individuals The generated random step size follows a normal distribution, For better individuals The position vector of Repeat the iteration to update the position vectors of individuals and worse individuals until the value of the update function converges, and finally output the individual with the smallest update value in the current strategy population, which is recorded as the best individual. The corresponding set of optimization variables is the optimal control strategy sequence.

[0014] A building energy-saving intelligent integrated control method is implemented based on the building energy-saving intelligent integrated control system, comprising: step 1, collecting building comprehensive data of the internal space area of ​​the building; pre-processing the building comprehensive data to obtain standard building data; Step 2: Establish a building dynamic model based on standard building data, and use the dynamic model to predict the energy consumption change curve and comfort change curve of the internal space area of ​​the building in a fixed time period in the future; Step 3: Taking the energy consumption change curve and comfort change curve of the interior space of the building in a fixed time period in the future as constraints, a series of optimal control strategy sequences are solved; and the solved optimal control strategy sequences are distributed to each actuator of the building.

[0015] The technical effects and advantages of the building energy-saving intelligent integrated control system and method of the present invention are as follows: The present invention can fully perceive the environmental status and energy consumption inside the building, laying a solid data foundation for subsequent refined control; by processing the original data, not only the data quality is improved, but also the computational complexity is reduced; secondly, a dynamic modeling method is adopted, which can accurately describe the complex dynamic behavior of the internal space of the building, and accurately predict the energy consumption and comfort changes in the future, which provides important support for the formulation of forward-looking control strategies and helps to actively adjust and optimize the operating status of the building; furthermore, in the process of solving the control strategy, the two goals of energy saving and comfort are organically combined to achieve multi-objective optimization; by weighing the relationship between the two, a balanced optimal control strategy can be obtained, which maximizes the satisfaction of energy saving and comfort needs and avoids the one-sidedness that may be caused by single-objective optimization; in addition, an improved multi-objective optimization algorithm is adopted, which has the advantages of strong global search capability, fast convergence speed, good robustness, etc., and can efficiently solve the optimal control strategy sequence under constraints, thereby improving the control accuracy and performance; finally, the integrated intelligent control of the equipment is realized, which not only improves the energy saving level of the building, but also creates a comfortable environment for the interior, and comprehensively enhances the comprehensive performance of the building. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a building energy-saving intelligent integrated control system according to the present invention; Figure 2 It is a schematic diagram of a building energy-saving intelligent integrated control method of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example 1 See also Figure 1As shown, this embodiment provides an intelligent integrated control system for building energy conservation, including: a data acquisition module for acquiring comprehensive building data of the internal space area of ​​the building; pre-processing the comprehensive building data to obtain standard building data; Comprehensive prediction module, used to establish a building dynamic model based on standard building data, and use the dynamic model to predict the energy consumption change curve and comfort change curve of the building's internal space area in a fixed time period in the future; The strategy solving module is used to take the energy consumption change curve and comfort change curve of the internal space area of ​​the building in a fixed time period in the future as constraints to solve a series of optimal control strategy sequences; the optimal control strategy sequences obtained are distributed to the various actuators of the building to realize integrated intelligent control of heating, cooling, ventilation, lighting and other equipment; the modules are connected by wired and / or wireless means to realize data transmission between modules.

[0019] Furthermore, the comprehensive building data includes environmental parameters and energy consumption data in the building; environmental parameters include temperature, humidity, carbon dioxide concentration, illumination and fresh air volume; energy consumption data includes heating, cooling and lighting power; illumination refers to the luminous flux received per unit area, which is a physical quantity that represents the intensity of light. Illumination reflects the lighting and illumination conditions of the internal space area of ​​the building. The appropriate illumination level is very important for the comfort of the internal environment of the building. Too high or too low illumination will affect people's visual comfort; fresh air volume refers to the amount of fresh air added to the internal space area of ​​the building through the fresh air system; inside the building, due to personnel activities, office equipment operation, etc., heat, carbon dioxide, volatile organic compounds and other pollutants will be generated. If the indoor air is not replaced for a long time, the air quality will gradually deteriorate, affecting human health and comfort. Therefore, it is necessary to continuously deliver fresh air to the room through the fresh air system to dilute and discharge indoor pollutants and maintain good indoor air quality.

[0020] Heating capacity refers to the amount of heat provided by the heating system to the interior space of a building, usually measured in kilowatt-hours; the heating system consumes primary energy (such as natural gas, electricity, etc.) to provide heat for the building, and the amount of heating capacity directly determines the energy consumption level of the heating system; cooling capacity refers to the amount of heat removed from the interior space of a building by the air-conditioning and refrigeration system, usually measured in kilowatt-hours; the air-conditioning compressor needs to consume electricity to complete cooling. The greater the cooling capacity, the higher the power consumption of the air-conditioning system.

[0021] The comprehensive building data is cleaned, that is, outliers and noise data are removed, and data of different dimensions are dimensionless so that they can be calculated and compared at the same order of magnitude; commonly used standardization methods include minimum-maximum standardization and Z-Score standardization to obtain preliminary building data; Arrange the preliminary building data into corresponding data time series according to the time sequence of data types; segment the data time series to obtain m segments of building data; and extract standard building data from the m segments of building data.

[0022] Specifically, for each segment of building data, various environmental parameters and energy consumption data are extracted, such as temperature, humidity, carbon dioxide concentration, illumination, fresh air volume, heating capacity, cooling capacity, lighting power, etc.; for each environmental parameter and energy consumption data, its comprehensive statistical characteristics within the data segment are calculated, and the comprehensive statistical characteristics include mean, variance, maximum value, minimum value, kurtosis and skewness.

[0023] The comprehensive statistical characteristics of each environmental parameter and energy consumption data are spliced ​​into a vector as the standard building data for this segment of building data; for m segments of building data, m vectors are obtained by repeating, namely the standard building data.

[0024] The original time series building data is compressed into a vector of statistically significant standard building data, which can more comprehensively describe the characteristics of each data segment and lay the foundation for subsequent modeling and optimization. At the same time, the dimension of the standard building data vector is relatively low, which is conducive to reducing computational complexity.

[0025] Ways to segment data include: For each data time series (such as temperature time series), it is regarded as a state sequence of a discrete-time Markov chain. According to the range of the data values ​​of the data time series, it is discretized into a finite number of states, that is, the state space of the Markov chain is determined. For example, the temperature data is divided into three states: high temperature, medium temperature, and low temperature.

[0026] Initialize a counting matrix C, where all elements in the counting matrix C are initially 0. Traverse the time series data, and for each adjacent time point pair (Y_t, Y_t+1), where Y_t represents the state at time t and Y_t+1 represents the state at time t+1, determine the state indexes i and j corresponding to Y_t and Y_t+1 respectively; add 1 to the value of the element C(i, j) in the counting matrix C corresponding to the state indexes i and j, that is, C(i, j) records the number of transitions from state z_i to state z_j; after traversing the entire time series, for each state s_i, calculate the frequency of its transitions to other states, that is, C(i, j) divided by the sum of the number of all transitions starting from state z_i; repeat all the frequencies obtained, which are all the elements of the transition probability matrix P.

[0027] Using the transition probability matrix, for each time point t in the data time series, calculate the conditional probability P(t, z_i) that the data value at that time point belongs to each state z_i.

[0028] For each time point t, the maximum value P_max of the conditional probability that the data value at that time point belongs to each state is calculated.

[0029] Based on the maximum value P_max, the probability threshold θ=a×P_max is preset. When P(t, z_i) is greater than or equal to θ, the time point t is determined to belong to the state z_i. a is a weighting parameter, and its value range is a positive real number, usually set to a value in the interval (0, 1]. Different a values ​​will result in different sizes of the probability threshold θ, which in turn affects the difficulty of a time point being classified into a certain state.

[0030] Scan the entire data time series and classify adjacent time points belonging to the same state into a data segment, that is, if time points t and t+1 belong to the same state z_i, they are divided into one data segment; that is, several segments of preliminary segmented data are obtained; cluster analysis is performed on several segments of preliminary segmented data to obtain m segments of building data.

[0031] Specifically, the eigenvector of each preliminary segmented data is calculated. The eigenvector is obtained by concatenating the weighted mean vector, variance vector, mean vector, maximum value, minimum value, kurtosis and skewness. The variance vector is the variance of each variable (data point) in the preliminary segmented data.

[0032] The weighted mean vector is calculated by: For Preliminary segmented data, the time range is [t_l _ start,t_l _ end] ,in, For the The starting time point of the preliminary segmented data, For the The end time point of the preliminary segmented data; based on the end time point Define the time weight function; the formula of the time weight function is: ;in, is an attenuation coefficient, is the periodic adjustment coefficient, which controls the strength of the periodicity. For time The time weight of is the initial time point of the function, controls the phase of the function, and determines the starting position of the entire function in time. For example, a day is 24 hours, a year is 365 days, and so on.

[0033] Based on the time weight function, the The weighted mean vector of the preliminary segmented data ; ;in, For data points The confidence weight of the data point For preliminary segmentation of data in time data points.

[0034] ,in, is the mean of the preliminary segmented data, for The standard deviation of the neighborhood, and is a learnable confidence parameter. Data points that are far from the mean and have a large neighborhood variance have lower confidence. By learning the parameter, the confidence weight can be adaptively determined.

[0035] Based on the calculated eigenvectors, a distance matrix D is constructed, and the element D(I, J) in the distance matrix D represents the Mahalanobis distance between the I-th preliminary segmented data and the J-th preliminary segmented data; based on the distance matrix D, merging iterations are performed. Specifically, each preliminary segmented data in the distance matrix D is regarded as a separate cluster, and the eigenvector of each cluster is calculated. The comprehensive distance between clusters is calculated based on the eigenvectors, and the recent clusters will be assigned a smaller distance.

[0036] The calculation formula of comprehensive distance is: ; in, Cluster and The combined distance between Cluster The characteristic vector of Cluster The characteristic vector of It is an adaptive covariance matrix that automatically adjusts the weights of different feature dimensions according to the data distribution. Cluster The time center, It is a cluster The time center, is the time decay parameter.

[0037] The merging cost between clusters is calculated based on the comprehensive distance between clusters; the calculation formula of the merging cost is: ;in, is the cluster shape penalty, Cluster and The merger cost between .

[0038] Cluster shape penalty ;in, Cluster The variance of Cluster The variance of Cluster The skewness of Cluster The skewness of and It is a learnable parameter. The cluster shape penalty term tends to merge clusters with similar shapes (similar variance and skewness) to avoid deformed clusters.

[0039] Among all the current pairwise clusters, find the two clusters with the lowest merging cost. and , merge them into new clusters ; New cluster The eigenvector of and The feature vectors of and their sample numbers are weighted and calculated; calculate the new cluster The distances to other clusters are calculated and the distance matrix is ​​updated.

[0040] When the sum of the squares of the intra-cluster deviations of all clusters is less than the preset merging threshold, the merging is stopped, and finally m clusters corresponding to m segments of building data are obtained.

[0041] Through the clustering process, the original over-segmented data segments are further merged to obtain a larger and more compact data segment representation, automatically discovering similar patterns within the data and avoiding manual subjective division. At the same time, by setting reasonable termination conditions, the degree of merging can be controlled to obtain the optimal data segmentation result.

[0042] The original time series of comprehensive building data is segmented into multiple data segments, with each data type corresponding to m data segments. The segmentation method can adaptively discover potential pattern transitions from the data, which is beneficial for subsequent modeling and optimization.

[0043] The way to build a dynamic model of a building is as follows: The internal space of the building is discretized into several lattice units. Each lattice unit corresponds to a magnetic spin. The magnetic spin can take values ​​of up or down (-1 or 1), representing the two states of the lattice unit (such as high temperature / low temperature).

[0044] There are several standard building data in each grid unit; based on the standard building data, define the adjacent grid units and The coupling strength between ;in, For grid cells and grid cells the distance between the centers (be it Euclidean distance or other distance metric), For grid cells Any standard building data in For grid cells Any standard building data in For grid cells and grid cells The relative azimuth between them; in a built environment, adjacent lattice units may have different coupling strengths at different azimuths (e.g., different heat conduction in the east-west and north-south directions); is anisotropy parameter, controlling amplitude, with a value range of [0, 1], is the phase parameter, which controls the phase offset and has a value range of [0, 2π); is an exponential parameter that controls the steepness; is the intensity adjustment parameter, which controls the numerical range of the coupling intensity; Adjust the exponential parameter for intensity and control the degree of smoothness; It is the distance exponent parameter, which controls the degree of distance coupling.

[0045] The external field terms are defined using environmental parameters and energy consumption data, and the external field terms reflect the impact of the external environment on the state of the grid unit; that is, ; For grid cells The outfield item, For grid cells The corresponding environmental parameter vector (such as temperature, humidity, illumination, etc.), For grid cells The corresponding energy consumption data vector (such as heating capacity, cooling capacity, lighting power, etc.); is a mapping function, such as a linear function, a multi-layer perceptron (MLP) function, a kernel function, or an attention function.

[0046] The standard building data of each grid unit is randomly divided into a training set and a test set according to a certain ratio (such as 7:3), and the data in the training set is used as the training data; the loss function is to minimize the mean square error between the output and the training data (targeted training grid units).

[0047] Define the standard building data of the p-th training data as X_p, and obtain the corresponding real environmental parameters and energy consumption data; use the current coupling parameters (including coupling intensity and external field terms) to simulate time evolution and get the output Ŷ_p (the model's predicted output for input X_p, including predicted environmental parameters and energy consumption data).

[0048] Initialize the coupling parameters to a set of random values, and iterate the training data several times. The iterative process is: randomly sample a small batch of training data from the training set, calculate the loss function for this small batch of training data, and calculate the gradient of the loss function with respect to the coupling parameters. Use stochastic gradient descent or its variant algorithms (such as Adam, RMSProp, etc.) to update the coupling parameters according to the gradient; until the preset maximum number of iterations is reached, the optimal coupling parameters (coupling strength and external field terms) are obtained, so that the training data can be well fitted, thereby obtaining better prediction and optimization capabilities; and the established building dynamic model is obtained.

[0049] Obtain the current value of the standard building data, and determine the initial state (spin orientation) of each lattice unit according to the current value of the standard building data; and form the initial states of all lattice units into a state vector; In time , according to the current state vector and the optimal coupling parameters (coupling strength and external field terms), calculate each grid unit The probability of flipping spin ; ;in, For grid cells In time The spin state (referring to the value of the magnetic spin corresponding to each lattice unit at a certain moment) is 1 or -1. For grid cells In time The spin state of is 1 or -1. It is a random disturbance term that obeys a certain probability distribution, such as Gaussian distribution or other distribution, in order to simulate the influence of various random factors on the state of the grid unit in actual situations and make it closer to reality.

[0050] According to the probability of flipping spin , using Monte Carlo sampling, determine the grid cells In time Whether to flip the spin; that is, the corresponding grid unit Generate a uniform random number in the interval [0, 1] ;like , then the grid unit In time The spin state Set to , that is, the spin flips, otherwise the spin state remains unchanged; all lattice units are The spin state of the new state vector is formed; repeat until the time reaches the predicted maximum time step , and obtain a series of state vectors.

[0051] For each time step, the state vector corresponds to a set of environment parameter vectors and energy consumption data vector , through the reverse mapping function Convert to and , the reverse mapping function can be a linear function, a nonlinear function (such as MLP) or other functions.

[0052] Environmental parameter vector At the maximum time step All sampling points in the environment parameter change curve constitute the environmental parameter change curve; for each time step, the corresponding comfort index is calculated to form the comfort change curve; the comfort index is the percentage of dissatisfaction , specific, percentage of dissatisfaction It is related to the mean thermal sensation index and is used to predict what proportion of people will be dissatisfied with the current thermal environment given the mean thermal sensation index value.

[0053] ; Mean thermal sensation index It is a widely used thermal comfort index used to evaluate the subjective thermal sensation of the human body to the ambient thermal environment, with a value range of [-3, +3]. The average thermal sensation index is 0, indicating thermal neutrality, that is, the human body feels neither cold nor hot, and is in the most comfortable state. The average thermal sensation index is less than 0, indicating that the human body feels cold. The smaller the average thermal sensation index value, the colder it feels. The average thermal sensation index is greater than 0, indicating that the human body feels hot. The larger the average thermal sensation index value, the hotter it feels.

[0054] The average thermal sensation index needs to take into account multiple environmental parameters and human parameters, including air temperature, average radiant temperature, relative humidity, air flow speed, human activity level, clothing conditions, etc., and is calculated based on the environmental parameter change curve.

[0055] It can be seen that when the average thermal sensation index is 0, the dissatisfaction percentage is equal to 5%, which means that even in an ideal thermal neutral environment, 5% of people will feel dissatisfied, which is caused by individual differences in the human body.

[0056] When the average thermal sensation index deviates from 0, the dissatisfaction percentage will increase rapidly; for example, when the average thermal sensation index is equal to 1, the dissatisfaction percentage is equal to 26%; when the average thermal sensation index is equal to 2, the dissatisfaction percentage is equal to 76%.

[0057] It is hoped that the dissatisfaction percentage is as small as possible to provide an ideal thermal comfort environment. Therefore, the change curve of the dissatisfaction percentage is taken as one of the optimization goals to maximize the satisfaction of human thermal comfort needs while saving energy.

[0058] Using energy consumption data vector At the maximum time step The total energy consumption corresponding to all sampling points in , which constitutes the energy consumption change curve.

[0059] Based on the initial state of the grid and the time evolution process, the state changes of each grid unit in the future are predicted, and the changes in environmental parameters, energy consumption data, comfort indicators and energy consumption are calculated accordingly.

[0060] The optimal control strategy sequence can be solved by: The optimization objectives are defined to include minimizing energy consumption and maximizing comfort, that is, minimizing the total energy consumption in the future and minimizing the comfort index.

[0061] The optimization variables are defined as control parameters that affect the operation of the building environment and energy systems, including heating, cooling, fresh air, and lighting power.

[0062] Define constraints including energy consumption curve The peak value does not exceed the preset energy consumption limit and the peak and valley ranges of the comfort change curve are all within the preset comfort threshold range; Minimizing energy consumption and maximizing comfort are expressed as two objective functions and ; ; ; Initialize a strategy population, determine the strategy population size N2 and the dimension n2 of the optimization variable; for each individual in the strategy population, randomly generate a set of optimization variables as the individual's position vector.

[0063] According to the constraint range of the optimization variable, the position vector is processed to ensure that it is within the feasible region. The two objective functions and The weighted sum of is used as the update function of the strategy population; calculate the individual The value of the update function , that is, the update function is The function value at is recorded as the updated value.

[0064] Calculate the centroid of the current population , as a potential optimal solution; among them, is the position vector of individual d, for each individual , calculate individual With the centroid The Euclidean distance and the unit vector .

[0065] For each individual , generates a random step size that follows a normal distribution , update the individual according to the random step size and unit vector The position vector of individual New position vector ; and calculate the new position vector The value of the update function (Update value).

[0066] ;in, is a constant parameter, It is a random variable that obeys a specific distribution (such as Gaussian distribution or other distribution). It introduces a certain random disturbance in the displacement direction, which helps individuals to jump out of the local optimum and improve the diversity and robustness of the algorithm.

[0067] Sort the updated values ​​from small to large to obtain the sorted strategy population, and divide the sorted strategy population into two parts, the first part is better individuals, and the second part is worse individuals; The number of better individuals is ;in, and To dynamically adjust the parameters, is the current iteration number; is the rounding function; generally speaking The value of should not be too small (to avoid insufficient information utilization) nor too large (to avoid excessive computational overhead).

[0068] For each worse individual , randomly select a better individual , calculate the Euclidean distance between the two, recorded as the gap distance ; Update the worse individual according to the gap distance The position vector , get a worse individual New position vector ; Calculate the worse individual New position vector The updated value of; dynamically adjust the division method of “better” and “worse” individuals according to the actual situation of the population, so as to use the population information more flexibly and efficiently.

[0069] ;in, For better individuals and worse individuals The generated random step size follows a normal distribution, For better individuals The position vector of Repeat the iteration to update the position vectors of individuals and worse individuals until the value of the update function converges, and finally output the individual with the smallest update value in the current strategy population, which is recorded as the best individual. The corresponding set of optimization variables is the optimal control strategy sequence.

[0070] These control strategies are distributed to various actuators to achieve integrated intelligent control of heating, cooling, ventilation, lighting and other equipment.

[0071] This embodiment can fully perceive the environmental status and energy consumption inside the building, laying a solid data foundation for subsequent refined control; by processing the original data, not only the data quality is improved, but also the computational complexity is reduced; secondly, a dynamic modeling method is adopted, which can accurately describe the complex dynamic behavior of the internal space of the building and accurately predict the energy consumption and comfort changes in the future, which provides important support for the formulation of forward-looking control strategies and helps to actively adjust and optimize the operating status of the building; furthermore, in the process of solving the control strategy, the two goals of energy saving and comfort are organically combined to achieve multi-objective optimization; by weighing the relationship between the two, a balanced optimal control strategy can be obtained, which maximizes the satisfaction of energy saving and comfort needs and avoids the one-sidedness that may be caused by single-objective optimization; in addition, an improved multi-objective optimization algorithm is adopted, which has the advantages of strong global search capability, fast convergence speed, good robustness, etc., and can efficiently solve the optimal control strategy sequence under constraint conditions, thereby improving the control accuracy and performance; finally, the integrated intelligent control of the equipment is realized, which not only improves the energy saving level of the building, but also creates a comfortable environment for the interior, and comprehensively enhances the comprehensive performance of the building.

[0072] Example 2 See also Figure 2 As shown, the part not described in detail in this embodiment is described in Example 1, which provides a building energy-saving intelligent integrated control method, including: Step 1: Collect comprehensive building data of the internal space area of ​​the building; pre-process the comprehensive building data to obtain standard building data; Step 2: Establish a building dynamic model based on standard building data, and use the dynamic model to predict the energy consumption change curve and comfort change curve of the internal space area of ​​the building in a fixed time period in the future; Step 3: Take the energy consumption change curve and comfort change curve of the interior space of the building in a fixed time period in the future as constraints, and solve a series of optimal control strategy sequences; distribute the solved optimal control strategy sequences to each actuator of the building to realize integrated intelligent control of heating, cooling, ventilation, lighting and other equipment.

[0073] Example 3 This embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the operation mode of the above-mentioned building energy-saving intelligent integrated control method is implemented.

[0074] Since the electronic device introduced in this embodiment is an electronic device used to implement a building energy-saving intelligent integrated control method in the embodiment of this application, based on the building energy-saving intelligent integrated control method introduced in the embodiment of this application, the technical personnel of this field can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of this application is not described in detail here. As long as the technical personnel of this field implement the electronic device used in the building energy-saving intelligent integrated control method in the embodiment of this application, it belongs to the scope of protection of this application.

[0075] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0076] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A building energy-saving intelligent integrated control system, characterized in that: include: A data acquisition module is used to collect comprehensive building data of the internal space area of ​​the building; Preprocess the comprehensive building data to obtain standard building data; Comprehensive prediction module, used to establish a building dynamic model based on standard building data, and use the dynamic model to predict the energy consumption change curve and comfort change curve of the building's internal space area in a fixed time period in the future; The strategy solving module is used to use the energy consumption change curve and comfort change curve of the internal space area of ​​the building in a fixed time period in the future as constraints to solve a series of optimal control strategy sequences; the solved optimal control strategy sequences are distributed to each actuator of the building, and each module is connected by wired and / or wireless means to realize data transmission between modules.

2. The building energy-saving intelligent integrated control system according to claim 1 is characterized in that: The comprehensive building data includes environmental parameters and energy consumption data within the building; the environmental parameters within the building include temperature, humidity, carbon dioxide concentration, illumination and fresh air volume; the energy consumption data includes heating capacity, cooling capacity and lighting power.

3. The building energy-saving intelligent integrated control system according to claim 2 is characterized in that: The method of preprocessing the building comprehensive data includes: The comprehensive building data is cleaned and the data of different dimensions are dimensionless processed to obtain preliminary building data; the preliminary building data is arranged into corresponding data time series according to the chronological order of data types; the data time series is segmented to obtain m segments of building data; standard building data are extracted from the m segments of building data.

4. The building energy-saving intelligent integrated control system according to claim 3 is characterized in that: The method of performing data segmentation includes: For each data time series, regard it as a state sequence of a discrete-time Markov chain, and discretize it into a finite number of states according to the range of the data values ​​of the data time series, that is, determine the state space of the Markov chain; initialize a counting matrix C, and the initial value of all elements in the counting matrix C is 0, traverse the time series data, and for each adjacent time point pair (Y_t, Y_t+1), where Y_t represents the state at time t, and Y_t+1 represents the state at time t+1, determine the state indexes i and j corresponding to Y_t and Y_t+1 respectively; add 1 to the value of the element C(i, j) in the counting matrix C corresponding to the state index i and j, that is, C(i, j) records the number of transitions from state z_i to state z_j; after traversing the entire time series, for each state s_i, calculate its frequency of transitions to other states, that is, C(i, j) divided by the sum of the number of all transitions starting from state z_i; all the frequencies obtained are all the elements of the transition probability matrix P; Using the obtained transition probability matrix, for each time point t in the data time series, calculate the conditional probability P(t, z_i) that the data value at that time point belongs to each state z_i; For each time point t, calculate the maximum value P_max of the conditional probability that the data value at that time point belongs to each state; Based on the maximum value P_max, a probability threshold θ=a×P_max is preset. When P(t, z_i) is greater than or equal to θ, it is determined that the time point t belongs to the state z_i; a is a weighting parameter; Scan the entire data time series and classify adjacent time points belonging to the same state into a data segment, that is, obtain several segments of preliminary segmented data; perform cluster analysis on several segments of preliminary segmented data to obtain m segments of building data.

5. The building energy-saving intelligent integrated control system according to claim 4 is characterized in that: The cluster analysis method includes: Calculate the feature vector of each preliminary segmented data, the feature vector is obtained by concatenating the weighted mean vector, variance vector, mean vector, maximum value, minimum value, kurtosis and skewness; The weighted mean vector is calculated by: For Preliminary segmented data, the time range is ,in, For the The starting time point of the preliminary segmented data, For the The end time point of the preliminary segmented data; based on the end time point Define the time weight function; the formula of the time weight function is: ;in, is an attenuation coefficient, is the periodic adjustment coefficient, For time The time weight of is the initial time point of the function, For the cycle; Based on the time weight function, the The weighted mean vector of the preliminary segmented data ; ;in, For data points The confidence weight of the data point For preliminary segmentation data at time data points; ,in, is the mean of the preliminary segmented data, for The standard deviation of the neighborhood, and is the confidence parameter; Based on the calculated eigenvectors, a distance matrix D is constructed, and the element D(I, J) in the distance matrix D represents the Mahalanobis distance between the I-th preliminary segmented data and the J-th preliminary segmented data; based on the distance matrix D, a merge iteration is performed to obtain m segments of building data.

6. The building energy-saving intelligent integrated control system according to claim 5 is characterized in that: The method of performing merging iteration based on the distance matrix D includes: Each preliminary segmented data in the distance matrix D is regarded as a separate cluster, the feature vector of each cluster is calculated, and the comprehensive distance between clusters is calculated based on the feature vector; the calculation formula of the comprehensive distance is: ;in, Cluster and The combined distance between Cluster The characteristic vector of Cluster The characteristic vector of is an adaptive covariance matrix, Cluster The time center, It is a cluster The time center, is the time decay parameter; The merging cost between clusters is calculated based on the comprehensive distance between clusters; the calculation formula of the merging cost is: ;in, is the cluster shape penalty, Cluster and The merger cost between Cluster shape penalty ;in, Cluster The variance of Cluster The variance of Cluster The skewness of Cluster The skewness of and is a learnable parameter; Among all the current pairwise clusters, find the two clusters with the lowest merging cost. and , merge them into new clusters ; New cluster The eigenvector of and The feature vectors of and their sample numbers are weighted and calculated; calculate the new cluster The distance to other clusters and update the distance matrix; When the sum of the squares of the intra-cluster deviations of all clusters is less than the preset merging threshold, the merging is stopped, and finally m clusters corresponding to m segments of building data are obtained.

7. The building energy-saving intelligent integrated control system according to claim 6 is characterized in that: The method of establishing the building dynamic model includes: The internal space of the building is discretized into a number of lattice units. Each lattice unit corresponds to a magnetic spin. The magnetic spin can be either up or down, representing the two states of the lattice unit. There are several standard building data in each grid unit; based on the standard building data, define the adjacent grid units and The coupling strength between ;in, For grid cells and grid cells The distance between the centers, Grid cell Any standard building data in For grid cells Any standard building data in Grid cell and grid cells The relative azimuth between is the anisotropy parameter, is the phase parameter; is the exponential parameter; is the intensity adjustment parameter, is the intensity adjustment index parameter; is the distance index parameter; The external field items are defined using environmental parameters and energy consumption data; i.e. ; Grid cell The outfield item, Grid cell The corresponding environmental parameter vector, Grid cell The corresponding energy consumption data vector; is the mapping function; The standard building data of each grid unit is randomly divided into a training set and a test set according to a certain ratio, and the data in the training set is used as the training data; the loss function is to minimize the mean square error between the output and the training data; the standard building data of the pth training data is defined as X_p, and the corresponding real environmental parameters and energy consumption data are obtained; the output is obtained by simulating time evolution using the current coupling parameters; the coupling parameters are initialized to a set of random values, and the training data is iterated several times. The iterative process is: randomly sample a small batch of training data from the training set, calculate the loss function for this small batch of training data, and calculate the gradient of the loss function to the coupling parameters, and use stochastic gradient descent or its variant algorithm to update the coupling parameters according to the gradient; until the preset maximum number of iterations is reached, the optimal coupling parameters are obtained, and the established building dynamic model is obtained.

8. The building energy-saving intelligent integrated control system according to claim 7 is characterized in that: The method of using the dynamic model to predict the energy consumption change curve and the comfort change curve of the interior space area of ​​the building within a fixed time period in the future includes: Obtain the current value of the standard building data, and determine the initial state of each grid unit according to the current value of the standard building data; and form the initial states of all grid units into a state vector; In time , according to the current state vector and the optimal coupling parameters, and calculate each grid cell The probability of flipping spin ; ;in, Grid cell In time The spin state of Grid cell In time The spin state of is the random disturbance term; According to the probability of flipping spin , using Monte Carlo sampling, corresponding to the grid unit Generate a uniform random number in the interval [0, 1] ;like , then the grid unit In time The spin state Set to , that is, the spin flips, otherwise the spin state remains unchanged; all lattice units are The spin state of the new state vector is formed; repeat until the time reaches the predicted maximum time step , and obtain a series of state vectors; For each time step, the state vector corresponds to a set of environment parameter vectors and energy consumption data vector , through the reverse mapping function Convert to and ; Environmental parameter vector At the maximum time step All sampling points in the environment parameter change curve constitute the environmental parameter change curve; for each time step, the corresponding comfort index is calculated to form the comfort change curve; the comfort index is the percentage of dissatisfaction ; Using energy consumption data vector At the maximum time step The total energy consumption corresponding to all sampling points in , which constitutes the energy consumption change curve.

9. The building energy-saving intelligent integrated control system according to claim 8 is characterized in that: The solution method of the optimal control strategy sequence includes: The optimization objectives are defined as minimizing energy consumption and maximizing comfort, that is, minimizing the total energy consumption in the future and minimizing the comfort index; the optimization objectives are expressed as two objective functions and ; ； ； Define optimization variables as control parameters; define constraints including energy consumption change curve The peak value does not exceed the preset energy consumption limit and the peak and valley ranges of the comfort change curve are all within the preset comfort threshold range; Initialize a strategy population, determine the strategy population size N2 and the dimension n2 of the optimization variable; for each individual in the strategy population, randomly generate a set of optimization variables as the individual's position vector; The two objective functions and The weighted sum of is used as the update function of the strategy population; calculate the individual The value of the update function , that is, the update function is The function value at is recorded as the updated value; Calculate the centroid of the current population As a potential optimal solution; among them, is the position vector of individual d, for each individual , calculate individual With the centroid The Euclidean distance and the unit vector ; For each individual , generates a random step size that follows a normal distribution , update the individual according to the random step size and unit vector The position vector of individual New position vector ; and calculate the new position vector The value of the update function ; ;in, is a constant parameter, is a random variable that follows a specific distribution; Sort the updated values ​​from small to large to obtain the sorted strategy population, and divide the sorted strategy population into two parts, the first part is better individuals, and the second part is worse individuals; The number of better individuals is ;in, and To dynamically adjust the parameters, is the current iteration number; is the rounding function; For each worse individual , randomly select a better individual , calculate the Euclidean distance between the two, recorded as the gap distance ; Update the worse individual according to the gap distance The position vector , get a worse individual New position vector ; Calculate the worse individual New position vector The updated value of ;in, For better individuals and worse individuals The generated random step size follows a normal distribution, For better individuals The position vector of Repeat the iteration to update the position vectors of individuals and worse individuals until the value of the update function converges, and finally output the individual with the smallest update value in the current strategy population, which is recorded as the best individual. The corresponding set of optimization variables is the optimal control strategy sequence.

10. A building energy-saving intelligent integrated control method, which is implemented based on the building energy-saving intelligent integrated control system according to any one of claims 1 to 9, characterized in that: include: Step 1: Collect comprehensive building data of the internal space area of ​​the building; Preprocess the comprehensive building data to obtain standard building data; Step 2: Establish a building dynamic model based on standard building data, and use the dynamic model to predict the energy consumption change curve and comfort change curve of the internal space area of ​​the building in a fixed time period in the future; Step 3: Taking the energy consumption change curve and comfort change curve of the interior space of the building in a fixed time period in the future as constraints, a series of optimal control strategy sequences are solved; and the solved optimal control strategy sequences are distributed to each actuator of the building.

Images