Industrial and commercial building temperature control method, device and equipment based on multilayer feedforward neural network and adaptive thermal comfort and medium

Abstract

The invention discloses an industrial and commercial building temperature control method, device and equipment based on a multi-layer feed-forward neural network and self-adaptive thermal comfort and a medium, and the method comprises the steps: training a multi-layer feed-forward neural network prediction model, and obtaining a temperature energy consumption prediction model; determining a comfortable temperature, a temperature range corresponding to the comfortable temperature and a temperature adjustment step length; determining an initial set temperature in the temperature range, and updating the initial set temperature according to a temperature adjustment step length to obtain a plurality of set temperatures; inputting set temperatures into the temperature energy consumption prediction model to obtain indoor predicted temperatures corresponding to the set temperatures and corresponding cooling and heating predicted energy consumption; and the optimal temperature is screened from the set temperatures. The invention belongs to the field of building temperature control. According to the method, the set temperature is dynamically optimized in combination with the dynamic comfort interval calculated in real time.

Description

Technical Field

[0001] The present invention relates to the field of building temperature control, and in particular to a method, device, equipment and medium for controlling industrial and commercial building temperature based on a multi-layer feedforward neural network and adaptive thermal comfort. Background Art

[0002] As global energy demand continues to grow, buildings account for 36% of total energy consumption. Heating, ventilation, and air conditioning (HVAC) systems, as the primary energy consumer, are crucial for energy conservation and emission reduction. Traditional air conditioning systems often use static setpoint temperature control, based on Fanger's Predicted Mean Temperature (PMV) model. These systems treat users as passive heat receptors and ignore the human body's ability to adapt to dynamic thermal environments.

[0003] Existing temperature control strategies for industrial and commercial buildings have significant technical bottlenecks, including: threshold settings based on static thermal comfort models (such as PMV) do not take into account the dynamic thermal adaptability of the human body, resulting in redundant energy consumption; weather forecasts rely on delayed updates of historical data, making it difficult to respond to sudden thermal disturbances; there is a computational delay conflict between high-precision white box models and real-time control systems, while data-driven models lack dynamic comfort interval constraints; at the same time, existing research focuses on low metabolic rate scenarios, and lacks understanding of the thermal comfort mechanisms of high-activity environments such as industrial plants and gymnasiums. The single optimization goal ignores the coordination of multiple factors, exacerbating the trade-off dilemma between energy efficiency and comfort. In view of this, the present invention provides a temperature control method for industrial and commercial buildings based on a multi-layer feedforward neural network and adaptive thermal comfort. Summary of the Invention

[0004] The present invention solves the technical problem of single building temperature control logic in the prior art by providing a temperature control method, device, equipment and medium for industrial and commercial buildings based on a multi-layer feedforward neural network and adaptive thermal comfort, and achieves the technical effect of dynamically optimizing the set temperature by combining a dynamic comfort range with real-time calculation.

[0005] In a first aspect, the present invention provides a temperature control method for industrial and commercial buildings based on a multi-layer feedforward neural network and adaptive thermal comfort, comprising:

[0006] Obtaining indoor condition data and outdoor meteorological condition data of the target building, and based on the indoor condition data and outdoor meteorological condition data of the target building, training a multi-layer feedforward neural network prediction model to obtain a temperature energy consumption prediction model, wherein the temperature energy consumption prediction model is used to predict indoor temperature and cooling and heating energy consumption;

[0007] Determine the comfortable temperature, the temperature range corresponding to the comfortable temperature, and the temperature adjustment step;

[0008] An initial set temperature is determined within the temperature range, and the initial set temperature is updated with a temperature adjustment step to obtain a plurality of set temperatures, wherein the initial set temperature belongs to the set temperature, and the set temperatures are all within the temperature range;

[0009] Input the set temperature into the temperature energy consumption prediction model to obtain the indoor predicted temperature and the corresponding cooling and heating predicted energy consumption corresponding to each set temperature;

[0010] The optimal temperature is selected from the set temperatures based on the temperature range, the predicted indoor temperatures corresponding to several set temperatures, and the corresponding predicted cooling and heating energy consumption.

[0011] Furthermore, based on the temperature range, the predicted indoor temperatures corresponding to the set temperatures, and the corresponding predicted cooling and heating energy consumption, the optimal temperature is screened from the set temperatures, including:

[0012] S151, determining whether the predicted indoor temperature corresponding to the initial set temperature is within the temperature range; if not, discarding the predicted indoor temperature corresponding to the next set temperature, and determining whether the predicted indoor temperature corresponding to the next set temperature is within the temperature range, until the predicted indoor temperature corresponding to the set temperature is within the temperature range; if yes, retaining the predicted indoor temperature;

[0013] S152, for the reserved set temperature, determining whether the indoor predicted temperature of the next set temperature after the reserved set temperature is within the temperature range; if not, discarding the set temperature, and continuing to determine whether the indoor predicted temperature of the next set temperature after the reserved set temperature is within the temperature range, until the indoor predicted temperature of the next set temperature after the reserved set temperature is within the temperature range; if yes, retaining the set temperature;

[0014] S153: If the cooling and heating predicted energy consumption of the reserved set temperature is lower than the cooling and heating predicted energy consumption of the next set temperature after the reserved set temperature, the reserved set temperature is used as the optimal temperature and the next set temperature after the reserved set temperature is deleted; otherwise, the next set temperature after the reserved set temperature is used as the optimal temperature and the reserved set temperature is deleted;

[0015] S154, repeat S152-S153 until all set temperatures are traversed.

[0016] Furthermore, determining the comfortable temperature, the temperature range corresponding to the comfortable temperature, and the temperature adjustment step size includes:

[0017] Determine the average outdoor temperature for the previous day and multiple days, including:

[0018]

[0019] PMOT is the average outdoor temperature of the previous day and several days, Text,d-1 is the average temperature of the day before, T ext,d-2 is the average temperature two days ago, T ext,d-3 is the average temperature three days ago, T ext,d-4 is the average temperature four days ago, T ext,d-5 is the average temperature five days ago, T ext,d-6 is the average temperature six days ago, T ext,d-7 is the average temperature seven days ago;

[0020] Determine the comfortable temperature based on the average outdoor temperature of the previous day and multiple days, including:

[0021] T comfort =0.31PMOT+17.8

[0022] Among them, T comfort For comfortable temperature;

[0023] According to the comfortable temperature, determine the temperature range corresponding to the comfortable temperature, including:

[0024] Limit low =T comfort -3.5

[0025] Limit high =T comfort +3.5

[0026] Among them, Limit low is the lower limit of the temperature range, Limit high is the upper limit of the temperature range;

[0027] The temperature adjustment step was 0.5°C.

[0028] Furthermore, the indoor condition data and outdoor weather condition data of the target building are obtained, including:

[0029] Obtain the target building's indoor condition data and outdoor meteorological condition data to be processed, where the outdoor condition data includes outdoor temperature and solar radiation, and the indoor meteorological condition data includes the time period, the air conditioning operation time within the time period, the indoor temperature of the previous time period, and the air conditioning set temperature;

[0030] The indoor condition data to be processed and the outdoor meteorological condition data to be processed are subjected to denoising, outlier removal, missing value filling and normalization to obtain indoor condition data and outdoor meteorological condition data.

[0031] Furthermore, regarding the multi-layer feedforward neural network prediction model, it includes:

[0032] The input of the multi-layer feedforward neural network prediction model is defined as indoor condition data and outdoor meteorological condition data, and the output is the predicted indoor temperature and cooling and heating energy consumption for the next period of time;

[0033] Determine the structure of the multi-layer feedforward neural network prediction model, including: a dual-branch structure of the multi-layer feedforward neural network prediction model, the branch structure adopts ANN, and the branch structure includes an input shared layer, a hidden layer, and an output layer.

[0034] Furthermore, the loss function of the multi-layer feedforward neural network prediction model includes:

[0035]

[0036] Among them, R 2 is the coefficient of determination, Y i is the true value, is the predicted value, μ is the mean of the true value, df is the degree of freedom, RMSE is the root mean square error, S standard is the standard deviation of the prediction error, and N is the number of samples.

[0037] Furthermore, the initial set temperature is updated with the temperature adjustment step to obtain several set temperatures, including:

[0038] The lower limit of the temperature range is used as the initial set temperature;

[0039] The sum of the initial set temperature and the temperature adjustment step is used as the next set temperature, and iteration is performed until the set temperature exceeds the upper limit of the temperature range.

[0040] In a second aspect, the present invention provides a temperature control device for industrial and commercial buildings based on a multi-layer feedforward neural network and adaptive thermal comfort, comprising:

[0041] An acquisition module is used to obtain indoor condition data and outdoor meteorological condition data of the target building, and based on the indoor condition data and outdoor meteorological condition data of the target building, train a multi-layer feedforward neural network prediction model to obtain a temperature energy consumption prediction model, wherein the temperature energy consumption prediction model is used to predict indoor temperature and cooling and heating energy consumption;

[0042] A determination module, used to determine a comfortable temperature, a temperature range corresponding to the comfortable temperature, and a temperature adjustment step;

[0043] A setting module is used to determine an initial set temperature within the temperature range and update the initial set temperature with a temperature adjustment step to obtain a plurality of set temperatures, wherein the initial set temperature belongs to the set temperature and the set temperatures are all within the temperature range;

[0044] The prediction module is used to input the set temperature into the temperature energy consumption prediction model to obtain the indoor predicted temperature and the corresponding cooling and heating predicted energy consumption corresponding to each set temperature;

[0045] The optimal temperature module is used to filter out the optimal temperature from the set temperatures based on the temperature range, the predicted indoor temperatures corresponding to several set temperatures, and the corresponding predicted cooling and heating energy consumption.

[0046] In a third aspect, the present invention provides an electronic device, comprising:

[0047] processor;

[0048] a memory for storing processor-executable instructions;

[0049] The processor is configured to execute to implement the industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort as provided in the first aspect.

[0050] In a fourth aspect, the present invention provides a non-transitory computer-readable storage medium. When the instructions in the non-transitory computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to implement the industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort as provided in the first aspect.

[0051] One or more technical solutions provided in the present invention have at least the following technical effects or advantages:

[0052] This method uses a multivariable, multi-layer feedforward neural network model to predict indoor temperature and energy consumption within the next hour. This model, combined with real-time calculation of the dynamic comfort range, dynamically optimizes the set temperature. This method overcomes the limitations of traditional static control and single-day predictions, achieving the coordinated optimization of energy efficiency and comfort in industrial and commercial buildings, making it particularly suitable for high-mobility environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0054] Figure 1 A schematic flow chart of the temperature control method for industrial and commercial buildings based on a multi-layer feedforward neural network and adaptive thermal comfort provided by the present invention;

[0055] Figure 2 A schematic diagram of the structure of the multi-layer feedforward neural network prediction model provided by the present invention;

[0056] Figure 3 A schematic diagram of the process for screening the optimal temperature provided by the present invention. DETAILED DESCRIPTION

[0057] The embodiment of the present invention solves the technical problem of single building temperature control logic in the prior art by providing an industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort.

[0058] The technical solution of the present invention is to solve the above technical problems, and the overall idea is as follows:

[0059] A temperature control method for industrial and commercial buildings based on a multi-layer feedforward neural network and adaptive thermal comfort includes: obtaining indoor condition data and outdoor meteorological condition data of a target building, and training a multi-layer feedforward neural network prediction model based on the indoor condition data and outdoor meteorological condition data of the target building to obtain a temperature energy consumption prediction model, wherein the temperature energy consumption prediction model is used to predict indoor temperature and cooling and heating energy consumption; determining a comfortable temperature, a temperature range corresponding to the comfortable temperature, and a temperature adjustment step; determining an initial set temperature within the temperature range, and updating the initial set temperature with the temperature adjustment step to obtain several set temperatures, wherein the initial set temperature belongs to the set temperature, and the set temperatures are all within the temperature range; inputting the set temperature into the temperature energy consumption prediction model to obtain the indoor predicted temperature and the corresponding cooling and heating predicted energy consumption corresponding to each set temperature; screening the optimal temperature from the set temperatures according to the temperature range, the indoor predicted temperatures corresponding to the several set temperatures, and the corresponding cooling and heating predicted energy consumption.

[0060] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0061] First, the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. Furthermore, the character " / " in this document generally indicates an "or" relationship between the associated objects.

[0062] The present invention provides Figure 1 The temperature control method for industrial and commercial buildings based on a multi-layer feedforward neural network and adaptive thermal comfort includes steps S11-S15:

[0063] Step S11, obtain the indoor condition data and outdoor meteorological condition data of the target building, and train the multi-layer feedforward neural network prediction model based on the indoor condition data and outdoor meteorological condition data of the target building to obtain a temperature energy consumption prediction model, wherein the temperature energy consumption prediction model is used to predict the indoor temperature and the cooling and heating energy consumption.

[0064] Specifically, it includes: obtaining the target building's indoor condition data to be processed and the outdoor meteorological condition data to be processed, wherein the outdoor condition data to be processed include outdoor temperature and solar radiation, and the indoor meteorological condition data include time period, air conditioning operation time within the time period, indoor temperature of the previous time period, and air conditioning set temperature; denoising, outlier removal, missing value filling, and normalization are performed on the indoor condition data to be processed and the outdoor meteorological condition data to be processed to obtain indoor condition data and outdoor meteorological condition data.

[0065] In the present invention, the time period is in hours, that is, there are 24 time periods in a day.

[0066] In order to ensure the accuracy of subsequent modeling or analysis, these processed data need to be cleaned and standardized.

[0067] To remove random fluctuations caused by sensor errors, signal interference, etc., filtering algorithms (such as sliding average, Kalman filtering) or wavelet transform methods can be used for denoising.

[0068] Detect and remove obviously unreasonable data points (such as the indoor temperature suddenly jumping to 100°C, or the solar radiation being negative, etc.). Outliers can be eliminated through statistical methods (such as Z-score, IQR) or machine learning detection models.

[0069] When data at certain time points are missing, linear interpolation, forward filling, and backward filling can be used to fill in the gaps.

[0070] Unify data of different dimensions or orders of magnitude to the same scale to facilitate subsequent modeling. Common methods include minimum-maximum normalization (Min-Max) or Z-score normalization.

[0071] About the multi-layer feedforward neural network prediction model, including:

[0072] The input of the multi-layer feedforward neural network prediction model is defined as indoor condition data and outdoor meteorological condition data, and the output is the predicted indoor temperature of the next period and the predicted cooling and heating energy consumption of the next period; the structure of the multi-layer feedforward neural network prediction model is determined, including: a dual-branch structure of the multi-layer feedforward neural network prediction model, the branch structure adopts ANN, and the branch structure includes an input shared layer, a hidden layer and an output layer.

[0073] like Figure 2 As shown, Figure 2 This is a schematic diagram of the structure of the multi-layer feedforward neural network prediction model provided by the present invention.

[0074] Define the inputs and outputs of the model.

[0075] The inputs are: outdoor humidity, outdoor temperature, and solar radiation (representing environmental heat load factors); the current time period (hourly information, used to capture the fluctuation pattern of energy consumption during the day); the air conditioning operation time in the current time period (reflecting the equipment load status); the indoor temperature in the previous hour (reflecting the thermal inertia characteristics of the building); and the set temperature (as the initial input of the control variable).

[0076] The output is: the predicted indoor temperature for the next period; the predicted cooling and heating energy consumption for the next period.

[0077] Determine the structure of a multi-layer feedforward neural network.

[0078] The present invention uses a branch structure ANN, which is divided into two branches, respectively used to predict the indoor temperature and energy consumption in the next period, where:

[0079] Branch 1 includes:

[0080] Input shared layer:

[0081] The shared input layer accepts the normalized 7 feature vectors and passes the information to branch 1 after passing through a fully connected layer.

[0082] Hidden layer:

[0083] Hidden layer 1: Fully connected layer, with 24 neurons and ReLU activation function, mainly used to capture the basic nonlinear relationship in the input features.

[0084] Hidden layer 2: Fully connected layer with 18 neurons, which continues to extract deeper features and model the smoothness of temperature changes.

[0085] Hidden layer 3: Fully connected layer, with 9 neurons, further reduces the dimension and integrates features to provide refined prediction information for the output layer.

[0086] Output layer: Finally, a fully connected layer (single neuron, linear activation) is used to output the predicted indoor temperature value for the next period.

[0087] Branch 2:

[0088] Input shared layer:

[0089] It shares the same input processing part with branch one, and the features obtained from the shared layer serve as the input of branch two.

[0090] Hidden layer:

[0091] Hidden layer 1: Fully connected layer with 45 neurons and ReLU activation function to capture preliminary features related to energy consumption.

[0092] Dropout layer: A dropout layer is added after the first hidden layer with a dropout rate of 20% to prevent overfitting of energy consumption prediction due to data instability.

[0093] Hidden layer 2: Fully connected layer, with 36 neurons and ReLU activation function, is used to further extract preliminary features.

[0094] Hidden layer 3: Fully connected layer, with 28 neurons and ReLU activation function to further extract and integrate nonlinear relationships.

[0095] Hidden layer 4: Fully connected layer, with 18 neurons, also using the ReLU activation function to continue optimizing feature expression.

[0096] Hidden layer 5: Fully connected layer with 9 neurons to generate a more compressed feature representation for the output layer.

[0097] Output layer:

[0098] Finally, a fully connected layer (single neuron, linear activation) is used to output the predicted load energy consumption value for the next period.

[0099] To optimize the number of neurons in the hidden layer, we usually start with 2n+1 neurons, where n is the number of input variables. Model performance improves with the increase in the number of hidden layers and neurons. In this case, 9 neurons produced good performance for training.

[0100] Taking data complexity into account, the branch complexity for temperature and energy consumption prediction differs. Since energy consumption prediction is more complex, a dropout layer is added after the first hidden layer to prevent overfitting. The dropout layer randomly disables 0.2 input neurons in the first hidden layer during each training period, thus preventing any simple dependencies between neurons.

[0101] The loss function of the multi-layer feedforward neural network prediction model includes:

[0102]

[0103] Among them, R 2 is the coefficient of determination, Y i is the true value, is the predicted value, μ is the mean of the true value, df is the degree of freedom, RMSE is the root mean square error, S standard is the standard deviation of the prediction error, and N is the number of samples.

[0104] When the training requirements (number of training times or error loss) are met, the multi-layer feedforward neural network prediction model can be output to obtain a temperature energy consumption prediction model.

[0105] Step S12: Determine the comfortable temperature, the temperature range corresponding to the comfortable temperature, and the temperature adjustment step.

[0106] Based on the ASHRAE Standard 55 comfort criteria, this paper defines adaptive comfort using a weighted calculation of the previous day's and multi-day mean outdoor temperatures (PMOT), and then determines the comfort temperature based on this temperature. PMOT is used to calculate thermal comfort and determine whether the adaptive comfort model is appropriate for a given day.

[0107] Specifically include:

[0108] Determine the average outdoor temperature for the previous day and multiple days, including:

[0109]

[0110] PMOT is the average outdoor temperature of the previous day and several days, T ext,d-1 is the average temperature of the day before, T ext,d-2 is the average temperature two days ago, T ext,d-3 is the average temperature three days ago, T ext,d-4 is the average temperature four days ago, T ext,d-5 is the average temperature five days ago, T ext,d-6 is the average temperature six days ago, T ext,d-7 is the average temperature seven days ago;

[0111] Determine the comfortable temperature based on the average outdoor temperature of the previous day and multiple days, including:

[0112] T comfort =0.31PMOT+17.8

[0113] Among them, T comfort For comfortable temperature;

[0114] According to the comfortable temperature, determine the temperature range corresponding to the comfortable temperature, including:

[0115] Limit low =T comfort -3.5

[0116] Limit high =T comfort +3.5

[0117] Among them, Limit low is the lower limit of the temperature range, Limit high is the upper limit of the temperature range;

[0118] The temperature adjustment step was 0.5°C.

[0119] It is understood that the acceptable level given in the present invention is 80%, that is:

[0120] Limit low =T comfort -3.5

[0121] Limit high =T comfort +3.5

[0122] In addition, a temperature range can be determined with a 90% acceptance level, namely:

[0123] Limit low =T comfort -2.5

[0124] Limit high =T comfort +2.5

[0125] In step S13, an initial set temperature is determined within the temperature range, and the initial set temperature is updated with a temperature adjustment step to obtain a plurality of set temperatures, wherein the initial set temperature belongs to the set temperatures, and the set temperatures are all within the temperature range.

[0126] The initial set temperature is updated with the temperature adjustment step to obtain several set temperatures, including: taking the lower limit of the temperature range as the initial set temperature; taking the sum of the initial set temperature and the temperature adjustment step as the next set temperature, and iterating until the set temperature exceeds the upper limit of the temperature range.

[0127] For example, if the temperature range is 21.5-28.5, 21.5 can be used as the initial set temperature. By updating the initial set temperature using the temperature adjustment step, the values can be 22.0, 22.5, 23.0, ... 28.5. Temperatures outside the temperature range are no longer included. Temperatures from 21.5, 22.0, 22.5, 23.0 ... 28.5 are all considered set temperatures.

[0128] Step S14: inputting the set temperature into the temperature energy consumption prediction model to obtain the predicted indoor temperature and the corresponding cooling and heating predicted energy consumption corresponding to each set temperature.

[0129] After inputting the set temperature into the temperature energy consumption prediction model, the indoor predicted temperature for the next period and the corresponding cooling and heating predicted energy consumption for the next period corresponding to each set temperature are obtained.

[0130] Step S15 , selecting the optimal temperature from the set temperatures according to the temperature range, the predicted indoor temperatures corresponding to the set temperatures, and the corresponding predicted cooling and heating energy consumptions.

[0131] The present invention provides Figure 3 The flowchart of selecting the optimal temperature shown in FIG. 1 shows a process of selecting the optimal temperature, which selects the optimal temperature from the set temperatures according to the temperature range, the predicted indoor temperatures corresponding to the set temperatures, and the corresponding predicted cooling and heating energy consumption, specifically including S151-S153:

[0132] S151, determine whether the indoor predicted temperature corresponding to the initial set temperature is included in the temperature range. If not, discard it, and determine whether the indoor predicted temperature of the next set temperature is included in the temperature range, until the indoor predicted temperature of the set temperature is included in the temperature range; if included, retain it.

[0133] For example, still taking the above example, the set temperatures are 21.5, 22.0, 22.5, 23.0…28.5, and the temperature range is 21.5-28.5.

[0134] If the predicted indoor temperature corresponding to 21.5 is not within the range of 21.5-28.5, then discard 21.5 and determine whether the predicted indoor temperature corresponding to 22.0 is within the range of 21.5-28.5. If so, retain it; if not, continue to determine 22.5.

[0135] S152, for the reserved set temperature, determine whether the indoor predicted temperature of the next set temperature of the reserved set temperature is included in the temperature range; if not, discard it, and continue to determine whether the indoor predicted temperature of the next set temperature of the reserved set temperature is included in the temperature range, until the indoor predicted temperature of the next set temperature of the reserved set temperature is included in the temperature range; if included, retain it.

[0136] For example, if 22.5 is the first retained temperature, then continue to determine whether the predicted indoor temperature corresponding to 23.0 is between 21.5 and 28.5. If not, continue to determine 23.5.

[0137] S153, if the cooling and heating predicted energy consumption of the reserved set temperature is lower than the cooling and heating predicted energy consumption of the next set temperature of the reserved set temperature, the reserved set temperature is used as the optimal temperature and the next set temperature of the reserved set temperature is deleted; otherwise, the next set temperature of the reserved set temperature is used as the optimal temperature and the reserved set temperature is deleted.

[0138] For example, if 22.5 is the first retained temperature and 23.5 is the second retained temperature, the cooling and heating predicted energy consumption corresponding to 22.5 is compared with the cooling and heating predicted energy consumption corresponding to 23.5. If the cooling and heating predicted energy consumption corresponding to 22.5 is lower, 22.5 is retained; otherwise, 23.5 is retained and the unreserved temperature is deleted.

[0139] S154, repeat S152-S153 until all set temperatures are traversed.

[0140] For example, if the retained value is 22.5, then continue to determine whether the indoor predicted temperature corresponding to 24.0 is included in the temperature range. If so, compare the cooling and heating predicted energy consumption corresponding to 22.5 with the cooling and heating predicted energy consumption corresponding to 24.0. If not, continue to determine 24.5 until all temperatures are traversed.

[0141] In summary, the present invention provides a temperature control method for industrial and commercial buildings based on a multi-layer feedforward neural network and adaptive thermal comfort, comprising: obtaining indoor condition data and outdoor meteorological condition data of a target building, and training a multi-layer feedforward neural network prediction model based on the indoor condition data and outdoor meteorological condition data of the target building to obtain a temperature energy consumption prediction model, wherein the temperature energy consumption prediction model is used to predict indoor temperature and cooling and heating energy consumption; determining a comfort temperature, a temperature range corresponding to the comfort temperature, and a temperature adjustment step size; determining an initial set temperature within the temperature range, and updating the initial set temperature with the temperature adjustment step size to obtain a plurality of set temperatures, wherein the initial set temperature belongs to the set temperature and the set temperature is included in the temperature range; inputting the set temperature into the temperature energy consumption prediction model to obtain a corresponding indoor predicted temperature and corresponding cooling and heating predicted energy consumption for each set temperature; and screening the set temperature to obtain an optimal temperature based on the temperature range, the corresponding indoor predicted temperature of the plurality of set temperatures, and the corresponding cooling and heating predicted energy consumption. The present invention predicts indoor temperature and energy consumption in the next hour by constructing a multivariable multi-layer feedforward neural network model, and dynamically optimizes the set temperature by combining the dynamic comfort range calculated in real time. This method breaks through the limitations of traditional static control and single-day prediction, achieving coordinated optimization of energy efficiency and comfort in industrial and commercial buildings, and is particularly suitable for scenarios with high personnel mobility.

[0142] Based on the same inventive concept, the present invention provides an industrial and commercial building temperature control device based on a multi-layer feedforward neural network and adaptive thermal comfort, comprising:

[0143] An acquisition module is used to obtain indoor condition data and outdoor meteorological condition data of the target building, and based on the indoor condition data and outdoor meteorological condition data of the target building, train a multi-layer feedforward neural network prediction model to obtain a temperature energy consumption prediction model, wherein the temperature energy consumption prediction model is used to predict indoor temperature and cooling and heating energy consumption;

[0144] A determination module, used to determine a comfortable temperature, a temperature range corresponding to the comfortable temperature, and a temperature adjustment step;

[0145] A setting module is used to determine an initial set temperature within the temperature range and update the initial set temperature with a temperature adjustment step to obtain a plurality of set temperatures, wherein the initial set temperature belongs to the set temperature and the set temperatures are all within the temperature range;

[0146] The prediction module is used to input the set temperature into the temperature energy consumption prediction model to obtain the indoor predicted temperature and the corresponding cooling and heating predicted energy consumption corresponding to each set temperature;

[0147] The optimal temperature module is used to filter out the optimal temperature from the set temperatures based on the temperature range, the predicted indoor temperatures corresponding to several set temperatures, and the corresponding predicted cooling and heating energy consumption.

[0148] Based on the same inventive concept, the present invention further provides an electronic device, comprising:

[0149] processor;

[0150] a memory for storing processor-executable instructions;

[0151] The processor is configured to execute to implement the industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort as provided above.

[0152] Based on the same inventive concept, the present invention also provides a non-temporary computer-readable storage medium. When the instructions in the storage medium are executed by the processor of an electronic device, the electronic device is enabled to implement the industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort as provided above.

[0153] Since the electronic device described in this embodiment is an electronic device used to implement the information processing method in the embodiment of the present invention, based on the information processing method described in the embodiment of the present invention, those skilled in the art will be able to understand the specific implementation of the electronic device of this embodiment and its various variations. Therefore, how the electronic device implements the method in the embodiment of the present invention will not be described in detail here. As long as the electronic device used by those skilled in the art to implement the information processing method in the embodiment of the present invention falls within the scope of protection of the present invention.

[0154] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0155] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0156] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0157] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0158] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0159] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A temperature control method for industrial and commercial buildings based on a multi-layer feedforward neural network and adaptive thermal comfort, characterized in that: include: Obtaining indoor condition data and outdoor meteorological condition data of the target building, and based on the indoor condition data and outdoor meteorological condition data of the target building, training a multi-layer feedforward neural network prediction model to obtain a temperature energy consumption prediction model, wherein the temperature energy consumption prediction model is used to predict indoor temperature and cooling and heating energy consumption; Determine the comfortable temperature, the temperature range corresponding to the comfortable temperature, and the temperature adjustment step; An initial set temperature is determined within the temperature range, and the initial set temperature is updated with a temperature adjustment step to obtain a plurality of set temperatures, wherein the initial set temperature belongs to the set temperature, and the set temperatures are all within the temperature range; Input the set temperature into the temperature energy consumption prediction model to obtain the indoor predicted temperature and the corresponding cooling and heating predicted energy consumption corresponding to each set temperature; The optimal temperature is selected from the set temperatures based on the temperature range, the predicted indoor temperatures corresponding to several set temperatures, and the corresponding predicted cooling and heating energy consumption.

2. The industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort according to claim 1, characterized in that: Based on the temperature range, the predicted indoor temperature corresponding to several set temperatures, and the corresponding predicted cooling and heating energy consumption, the optimal temperature is selected from the set temperatures, including: S151, determining whether the predicted indoor temperature corresponding to the initial set temperature is within the temperature range; if not, discarding the predicted indoor temperature corresponding to the next set temperature, and determining whether the predicted indoor temperature corresponding to the next set temperature is within the temperature range, until the predicted indoor temperature corresponding to the set temperature is within the temperature range; if yes, retaining the predicted indoor temperature; S152, for the reserved set temperature, determining whether the indoor predicted temperature of the next set temperature after the reserved set temperature is within the temperature range; if not, discarding the set temperature; and continuing to determine whether the indoor predicted temperature of the next set temperature after the reserved set temperature is within the temperature range, until the indoor predicted temperature of the next set temperature after the reserved set temperature is within the temperature range; if yes, retaining the set temperature; S153: If the cooling and heating predicted energy consumption of the reserved set temperature is lower than the cooling and heating predicted energy consumption of the next set temperature after the reserved set temperature, the reserved set temperature is used as the optimal temperature and the next set temperature after the reserved set temperature is deleted; otherwise, the next set temperature after the reserved set temperature is used as the optimal temperature and the reserved set temperature is deleted; S154, repeat S152-S153 until all set temperatures are traversed.

3. The industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort according to claim 1, characterized in that: Determine the comfortable temperature, the temperature range corresponding to the comfortable temperature, and the temperature adjustment step, including: Determine the average outdoor temperature for the previous day and multiple days, including: PMOT is the average outdoor temperature of the previous day and several days, T ext,d-1 is the average temperature of the day before, T ext,d-2 is the average temperature two days ago, T ext,d-3 is the average temperature three days ago, T ext,d-4 is the average temperature four days ago, T ext,d-5 is the average temperature five days ago, T ext,d-6 is the average temperature six days ago, T ext,d-7 is the average temperature seven days ago; Determine the comfortable temperature based on the average outdoor temperature of the previous day and multiple days, including: T comfort =0.31PMOT+17.8 Among them, T comfort For comfortable temperature; According to the comfortable temperature, determine the temperature range corresponding to the comfortable temperature, including: Limit low =T comforr -3.5 Limit high =T comfort +3.5 Among them, Limit low is the lower limit of the temperature range, Limit high is the upper limit of the temperature range; The temperature adjustment step was 0.5°C.

4. The industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort according to claim 1, characterized in that: Obtain indoor condition data and outdoor weather condition data of the target building, including: Obtaining the target building's indoor condition data and outdoor meteorological condition data to be processed, wherein the outdoor condition data to be processed includes outdoor temperature and solar radiation, and the indoor meteorological condition data includes time period, air conditioning operation time within the time period, indoor temperature in the previous time period, and air conditioning set temperature; The indoor condition data to be processed and the outdoor meteorological condition data to be processed are subjected to denoising, outlier removal, missing value filling and normalization to obtain indoor condition data and outdoor meteorological condition data.

5. The industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort as claimed in claim 1, characterized in that: About the multi-layer feedforward neural network prediction model, including: The input of the multi-layer feedforward neural network prediction model is defined as indoor condition data and outdoor meteorological condition data, and the output is the predicted indoor temperature and cooling and heating energy consumption for the next period of time; Determine the structure of the multi-layer feedforward neural network prediction model, including: a dual-branch structure of the multi-layer feedforward neural network prediction model, the branch structure adopts ANN, and the branch structure includes an input shared layer, a hidden layer, and an output layer.

6. The industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort as claimed in claim 1, characterized in that: The loss function of the multi-layer feedforward neural network prediction model includes: Among them, R 2 is the coefficient of determination, Y i is the true value, is the predicted value, μ is the mean of the true value, df is the degree of freedom, RMSE is the root mean square error, S standard is the standard deviation of the prediction error, and N is the number of samples.

7. The industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort as claimed in claim 3, characterized in that: The initial set temperature is updated with the temperature adjustment step to obtain several set temperatures, including: The lower limit of the temperature range is used as the initial set temperature; The sum of the initial set temperature and the temperature adjustment step is used as the next set temperature, and iteration is performed until the set temperature exceeds the upper limit of the temperature range.

8. Industrial and commercial building temperature control device based on multi-layer feedforward neural network and adaptive thermal comfort, characterized in that: include: An acquisition module is used to obtain indoor condition data and outdoor meteorological condition data of the target building, and based on the indoor condition data and outdoor meteorological condition data of the target building, train a multi-layer feedforward neural network prediction model to obtain a temperature energy consumption prediction model, wherein the temperature energy consumption prediction model is used to predict indoor temperature and cooling and heating energy consumption; A determination module, used to determine a comfortable temperature, a temperature range corresponding to the comfortable temperature, and a temperature adjustment step; A setting module is used to determine an initial set temperature within the temperature range and update the initial set temperature with a temperature adjustment step to obtain a plurality of set temperatures, wherein the initial set temperature belongs to the set temperature and the set temperatures are all within the temperature range; The prediction module is used to input the set temperature into the temperature energy consumption prediction model to obtain the indoor predicted temperature and the corresponding cooling and heating predicted energy consumption corresponding to each set temperature; The optimal temperature module is used to filter out the optimal temperature from the set temperatures based on the temperature range, the predicted indoor temperatures corresponding to several set temperatures, and the corresponding predicted cooling and heating energy consumption.

9. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute to implement the industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that When the instructions in the non-transitory computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to implement the industrial and commercial building temperature control method based on a multi-layer feedforward neural network and adaptive thermal comfort as described in any one of claims 1 to 7.

Images