Method and system for dynamic monitoring and energy-saving regulation of public institution building energy consumption

Abstract

The present application relates to the technical field of building energy consumption regulation, in particular to a public institution building energy consumption dynamic monitoring and energy saving regulation method and system, the method comprises the following steps: dividing the internal space of the public institution building into m independent functional areas based on the functions of each area of the public institution building; obtaining the spatial geometric dimensions of the m functional areas and their mutual adjacency relationship; randomly selecting one of the areas as a central area and identifying all areas with adjacency relationship and marking them as adjacent areas; and calculating the corresponding energy consumption influence coefficient. The present application obtains energy consumption abnormal regulation areas, can regulate the adjacent areas marked as energy consumption abnormal regulation areas, the regulation includes personnel flow regulation, space closure regulation or air exchange efficiency regulation, reduces the temperature influence of the adjacent areas on the central area, reduces the air conditioning energy consumption of the central area, and significantly enhances the active identification ability and spatial resolution ability of the energy saving regulation system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building energy consumption regulation, in particular to a public institution building energy consumption dynamic monitoring and energy saving regulation method and system. BACKGROUND

[0002] At present, the energy consumption demand of public buildings accounts for a large proportion of global energy consumption demand, public buildings are buildings that develop rapidly, consume a large amount of energy and are difficult to reduce consumption, and the air conditioning energy consumption proportion in public buildings reaches more than 50%, the primary task of public building energy saving is to reduce the energy consumption of the air conditioning system.

[0003] In the existing public institution building, when the air conditioning system is running, for the space that needs to be independently controlled, such as office, conference room, etc., physical closing measures such as setting curtains, keeping doors and windows closed, etc. are usually taken, the core purpose of which is to reduce the direct loss of cold / heat air in the space to the external environment, so as to improve the heat preservation effect of the independent space and reduce energy consumption, however, this conventional method often ignores a key factor, that is, the heat conduction effect between the independent space and its adjacent areas. When the temperature of the adjacent area is continuously unstable or lost, a significant and continuous temperature gradient will be formed on the independent space and these adjacent areas, and this temperature difference will continuously drive heat to conduct through the building envelope, even if the independent space itself takes closing measures, the comfortable temperature reached by the air conditioning adjustment will be continuously lost because of this continuous and high-intensity heat conduction process between the independent space and the adjacent area with poor temperature conditions, and the air conditioning system has to continuously output additional cold or heat to compensate for this part of energy loss through heat conduction, resulting in unnecessary increase of the overall air conditioning energy consumption of the building. SUMMARY

[0004] In order to solve the above problems, the present application provides a public institution building energy consumption dynamic monitoring and energy saving regulation method and system.

[0005] The present application adopts the following technical scheme, a public institution building energy consumption dynamic monitoring and energy saving regulation method, comprising:

[0006] Based on the functions of each area of the public institution building, the internal space of the public institution building is divided into m functional areas, m is a positive integer greater than 1;

[0007] Through the BIM model of the public institution building, the spatial geometric dimensions of the m functional areas and their mutual adjacency relationship are obtained, one of the functional areas is randomly selected as a central area, all functional areas with adjacency relationship are identified and marked as adjacent areas, and the initial temperature of the adjacent areas is obtained;

[0008] According to the abutment relationship between each abutment region and the center region, a corresponding energy consumption influence coefficient is calculated, and whether to retain the abutment region is determined according to the energy consumption influence coefficient;

[0009] A preset unit time length is set.

[0010] The spatial geometric size of the center region and the initial environmental state, the spatial geometric size of the abutment region and the energy consumption influence coefficient, the air conditioner temperature setting value, and the preset unit time length are input into a pre-constructed air conditioner system energy consumption prediction model to obtain a predicted unit time energy consumption.

[0011] A real-time unit time energy consumption curve of the air conditioner system in the center region is obtained.

[0012] The predicted unit time energy consumption is compared and analyzed with the real-time unit time energy consumption curve to generate an energy consumption abnormality regulation instruction.

[0013] As a further description of the above technical solution: the abutment relationship includes wall abutment and corridor connection.

[0014] As a further description of the above technical solution: the method for calculating the corresponding energy consumption influence coefficient of each abutment region and the center region includes:

[0015] When the abutment relationship is wall abutment, the method for obtaining the energy consumption influence coefficient of the center region and the abutment region includes:

[0016] The thermal conductivity of the wall between the center region and the abutment region, and the contact area of the center region and the abutment region and the thickness of the wall are obtained, and the energy consumption influence coefficient is calculated and obtained.

[0017] When the abutment relationship is corridor connection, the method for obtaining the energy consumption influence coefficient of the center region and the abutment region includes:

[0018] The air exchange rate of the corridor between the center region and the abutment region is obtained, the cross-sectional area of the corridor between the center region and the abutment region is obtained, and the energy consumption influence coefficient is calculated and obtained.

[0019] As a further description of the above technical solution: the method for determining whether to retain the abutment region according to the energy consumption influence coefficient includes:

[0020] A preset energy consumption influence coefficient threshold is set, when the obtained energy consumption influence coefficient of the center region and the abutment region is less than the preset energy consumption influence coefficient threshold, the abutment region is removed, when the obtained energy consumption influence coefficient of the center region and the abutment region is greater than or equal to the preset energy consumption influence coefficient threshold, the abutment region is retained, and a one-to-one correspondence between the abutment region and the energy consumption influence coefficient is established.

[0021] As a further description of the above technical solutions: the method for generating the energy consumption anomaly regulation instruction comprises:

[0022] A preset energy consumption difference gradient threshold E1, E2, wherein 0 < E1 < E2; an energy consumption difference AE of a real-time unit time energy consumption and a predicted unit time energy consumption is calculated;

[0023] When AE < E1, no energy consumption anomaly regulation instruction is generated;

[0024] When E1 ≤ AE ≤ E2, a unit time adjustment instruction is generated;

[0025] When AE > E2, an energy consumption anomaly regulation instruction is generated;

[0026] The unit time adjustment instruction is to reduce the unit time length.

[0027] As a further description of the above technical solutions: the training method of the air conditioning system energy consumption prediction model comprises:

[0028] The air conditioning system energy consumption dataset is collected in advance, the air conditioning system energy consumption dataset comprises P groups of spatial geometric dimensions and initial environment states of central regions, spatial geometric dimensions and energy consumption influence coefficients of adjacent regions, air conditioning temperature set values, and preset unit time lengths, and unit time energy consumptions corresponding to the P groups of spatial geometric dimensions and initial environment states of central regions, spatial geometric dimensions and energy consumption influence coefficients of adjacent regions, air conditioning temperature set values, and preset unit time lengths, P is a positive integer greater than 0, and the air conditioning system energy consumption dataset is divided into a training set and a verification set, wherein the training set is used to train the air conditioning system energy consumption prediction model, and the verification set is used to evaluate the generalization performance of the air conditioning system energy consumption prediction model;

[0029] In the training process of the air conditioning system energy consumption prediction model, a cross-entropy loss function is minimized as an optimization target, an early stopping strategy is used to monitor the performance of the verification set, the model performance is optimized by continuously adjusting network parameters, and when the prediction accuracy on the verification set reaches an expected accuracy, it is considered that the air conditioning system energy consumption prediction model has tended to converge, and the training is stopped; the air conditioning system energy consumption prediction model is trained by using a deep neural network based on a multilayer perceptron;

[0030] The spatial geometric dimensions and initial environment states of the central regions, the spatial geometric dimensions and energy consumption influence coefficients of the adjacent regions, the air conditioning temperature set values, and the preset unit time lengths are converted into feature vectors; the input layer of the air conditioning system energy consumption prediction model receives the feature vectors, extracts the nonlinear relationship in the feature vectors through multiple hidden layers, and finally the output layer of the air conditioning system energy consumption prediction model calculates the probability distribution of the unit time energy consumption through a softmax activation function, and outputs the unit time energy consumption corresponding to the maximum probability as the final prediction result.

[0031] As a further description of the above technical solution: further comprises:

[0032] When generating the energy consumption anomaly regulation instruction, the real-time temperature of each adjacent region is collected;

[0033] The characteristic parameters of the adjacent region and the characteristic parameters of the central region are input into a pre-constructed adjacent region temperature prediction model, and the predicted temperature of the adjacent region is output;

[0034] The characteristic parameters of the adjacent region include the initial temperature, the spatial geometric size of the adjacent region, and the energy consumption influence coefficient corresponding to the adjacent region; and the characteristic parameters of the central region include the initial temperature, the spatial geometric size of the central region, and the control temperature set value and the working time of the air conditioning system of the central region.

[0035] The predicted temperature is compared and analyzed with the collected real-time temperature of the adjacent region, and an energy consumption anomaly regulation region is generated.

[0036] As a further description of the above technical solution: the method of comparing and analyzing the predicted temperature with the collected real-time temperature of the adjacent region to generate the energy consumption anomaly regulation region comprises:

[0037] A preset temperature difference threshold Tmax is obtained, and the absolute value of the difference between the predicted temperature and the collected real-time temperature of the adjacent region is denoted as Tzj.

[0038] When Tzj>Tmax, the adjacent region is marked as an energy consumption anomaly regulation region;

[0039] When Tzj≤Tmax, the adjacent region is not marked.

[0040] As a further description of the above technical solution: the training method of the adjacent region temperature prediction model comprises:

[0041] An adjacent region temperature data set is pre-constructed, the construction of the adjacent region temperature data set comprises Y sets of characteristic parameters of the adjacent region and characteristic parameters of the central region, and the real-time temperature of the adjacent region corresponding to the Y sets of characteristic parameters of the adjacent region and characteristic parameters of the central region, Y is a positive integer greater than 0; the adjacent region temperature data set is divided into a training set and a data validation set, wherein the training set is used for parameter learning of the adjacent region temperature prediction model, and the validation set is used for real-time evaluation of the generalization ability of the adjacent region temperature prediction model.

[0042] In the training process of the adjacent area temperature prediction model, a deep neural network structure based on a multilayer perceptron is adopted, the feature parameters of the adjacent area and the feature parameters of the central area are converted into feature vectors as inputs, the nonlinear features in the data are extracted through multiple hidden layers, and finally the probability distribution of the temperature of the adjacent area is generated in the output layer by using a softmax activation function, and the temperature of the adjacent area corresponding to the maximum probability is output as the final prediction result; the training process aims to minimize the cross-entropy loss function, and an early stopping strategy is introduced to monitor the performance of the validation set; when the prediction accuracy on the validation set reaches a preset threshold, it is considered that the adjacent area temperature prediction model has converged, and the training is stopped.

[0043] The public institution building energy consumption dynamic monitoring and energy saving control system is used to realize the public institution building energy consumption dynamic monitoring and energy saving control method, and the system comprises:

[0044] The region division module divides the internal space of the public institution building into m functional regions based on the functions of each region of the public institution building.

[0045] The data acquisition module obtains the spatial geometric dimensions of the m functional regions and their adjacency relationship through the BIM model of the public institution building, randomly selects one of the functional regions as the central region, identifies all the functional regions adjacent to it and marks them as adjacent regions, and obtains the initial temperature of the adjacent regions.

[0046] The region analysis module calculates the energy consumption influence coefficient of each adjacent region according to its adjacency relationship with the central region, and determines whether to retain the adjacent region according to the energy consumption influence coefficient.

[0047] The energy consumption prediction module presets a unit time length, inputs the spatial geometric dimensions of the central region and the initial environmental state, the spatial geometric dimensions of the adjacent region and the energy consumption influence coefficient, the air conditioning temperature set value and the preset unit time length into the pre-constructed air conditioning system energy consumption prediction model, and obtains the predicted unit time energy consumption.

[0048] The energy consumption acquisition module obtains the real-time unit time energy consumption curve of the central region air conditioning system.

[0049] The instruction generation module compares and analyzes the predicted unit time energy consumption and the real-time unit time energy consumption curve to generate an energy consumption abnormality control instruction.

[0050] Advantages:

[0051] The public institution building energy consumption dynamic monitoring and energy saving regulation method provided by the application divides the building space into multiple functional areas based on function division, obtains the space geometric size and physical adjacency relationship of each area by using a BIM model, effectively constructs the initial relationship of energy consumption interference between areas, analyzes two adjacency situations of wall heat conduction and corridor air exchange, calculates the energy consumption influence coefficient, and screens out the adjacency area which has a significant influence on the central area according to the coefficient, so that the energy consumption influence relationship between areas is more accurate.

[0052] Further, by training the air conditioning system energy consumption prediction model, the difference between the predicted value and the actual value can be compared and analyzed in real time, so as to quickly identify operation abnormities, improve the accuracy and sensitivity of energy consumption monitoring, and has the advantages of improving dynamic response and reducing false alarm rate compared with the traditional fixed value threshold alarm mode. And by introducing the temperature prediction and abnormal identification mechanism of the adjacent area, the adjacent area can be analyzed under the condition of energy consumption abnormity in the central area, the energy consumption abnormity regulation area is obtained, the adjacent area marked as the energy consumption abnormity regulation area can be regulated, the regulation includes personnel flow regulation, space sealing regulation or air exchange efficiency regulation, the temperature influence of the adjacent area on the central area is reduced, the air conditioning energy consumption of the central area is reduced, and the active identification ability and spatial resolution ability of the energy saving regulation system are significantly enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0053] The application will be further explained in combination with the drawings and embodiments:

[0054] Figure 1 The flowchart of the public institution building energy consumption dynamic monitoring and energy saving regulation method provided by the embodiment 1 of the application is shown in the figure;

[0055] Figure 2 The flowchart of the public institution building energy consumption dynamic monitoring and energy saving regulation method provided by the embodiment 2 of the application is shown in the figure;

[0056] Figure 3 The timing diagram of the public institution building energy consumption dynamic monitoring and energy saving regulation method provided by the embodiment 1 of the application is shown in the figure;

[0057] Figure 4 The module connection diagram of the public institution building energy consumption dynamic monitoring and energy saving regulation system provided by the embodiment 3 of the application is shown in the figure. DETAILED DESCRIPTION

[0058] In order to make the technical means, creative features, purposes and effects realized by the application easy to understand, the application will be further described in combination with specific drawings. It should be noted that the embodiments in the application and the features in the embodiments can be combined with each other without conflict.

[0059] Embodiment 1

[0060] Referring to Figure 1 and Figure 3 , the embodiment of the present application provides a technical solution: a public institution building energy consumption dynamic monitoring and energy saving control method, comprising:

[0061] Based on the functions of each area of the public institution building, the internal space of the public institution building is divided into m independent functional areas, m is a positive integer greater than 1;

[0062] Through the BIM model of the public institution building, the spatial geometric dimensions of the m functional areas and their mutual adjacency relationship are obtained, one of them is randomly selected as the center area, and all the areas adjacent to it are identified and marked as adjacent areas, and the initial temperature of the adjacent area is obtained;

[0063] The adjacent relationship includes wall adjacency and corridor connection;

[0064] According to the adjacency relationship between each adjacent area and the center area, the corresponding energy consumption influence coefficient is calculated, and whether to retain the adjacent area is determined according to the coefficient;

[0065] According to the adjacency relationship between each adjacent area and the center area, the method for calculating the corresponding energy consumption influence coefficient includes:

[0066] When the adjacent relationship is wall adjacency, the method for obtaining the energy consumption influence coefficient of the center area and the adjacent area includes:

[0067] Obtain the thermal conductivity of the wall between the center area and the adjacent area, and the contact area and the thickness of the wall of the center area and the adjacent area, and obtain the energy consumption influence coefficient by formula calculation;

[0068] Optionally, the formula of the energy consumption influence coefficient is:

[0069] ; In the formula, is the energy consumption influence coefficient of the center area and the adjacent area , is the thermal conductivity of the center area and the adjacent area ; is the contact area of the center area and the adjacent area ; is the wall thickness of the center area and the adjacent area , , and are weight coefficients. , and are all greater than 0, + + =1;

[0070] It should be noted that in the above calculation process, the thermal conductivity, contact area and wall thickness need to be standardized to remove the dimension, and through standardization, different parameters are mapped to the same order of magnitude, ensuring the rationality and physical meaning of the weighted calculation;

[0071] The size of the weight coefficient is a specific value obtained by quantizing each data for subsequent comparison. The size of the weight coefficient depends on the number of comprehensive parameters and the corresponding weight coefficient initially set by the person skilled in the art for each group of comprehensive parameters.

[0072] It should be noted that the thermal conductivity can be obtained by searching the physical performance parameters of a large number of materials including thermal conductivity in the building material manual, engineering design manual or online database.

[0073] When the abutment relationship is a corridor connection, the method for obtaining the energy consumption influence coefficient of the center area and the abutted area includes:

[0074] The air exchange rate of the corridor between the center area and the abutted area is obtained, the cross-sectional area of the corridor between the center area and the abutted area is obtained, and the energy consumption influence coefficient is obtained by formula calculation.

[0075] Alternatively, the formula of the energy consumption influence coefficient is:

[0076] ; in the formula, the energy consumption influence coefficient of the center area and the abutted area , the air exchange rate of the corridor between the center area and the abutted area , the cross-sectional area of the corridor between the center area and the abutted area , and the weight coefficient, and are all greater than 0, + =1;

[0077] The air exchange rate and the cross-sectional area of the corridor are normalized to remove the dimension, and by normalization, different parameters are mapped to the same order of magnitude, ensuring the rationality and physical meaning of the weighted calculation; the size of the weight coefficient is to obtain a specific value by quantizing each data, which is convenient for subsequent comparison, and the size of the weight coefficient depends on the number of comprehensive parameters and the corresponding weight coefficient initially set by the person skilled in the art for each group of comprehensive parameters.

[0078] It should be noted that the air exchange rate of the corridor between the central region and the adjacent region can be obtained by arranging an anemometer or multiple-point measurement in the cross section of the corridor, calculating the average wind speed, and multiplying the average wind speed by the cross-sectional area of the corridor between the central region and the adjacent region to obtain the air exchange rate of the corridor between the central region and the adjacent region .

[0079] A preset energy consumption influence coefficient threshold is set, and when the energy consumption influence coefficient of the central region and the adjacent region is less than the preset energy consumption influence coefficient threshold, the adjacent region is removed, and when the energy consumption influence coefficient of the central region and the adjacent region is greater than or equal to the preset energy consumption influence coefficient threshold, the adjacent region is retained, and a one-to-one correspondence between the adjacent region and the energy consumption influence coefficient is established.

[0080] It should be noted that the energy consumption influence coefficient threshold is determined by the person skilled in the art according to a large number of experiments.

[0081] A preset unit time length is set; optionally, the unit time length is 5 min, 10 min, 30 min, and 60 min.

[0082] It should be noted that different unit times reflect the monitoring accuracy of the air conditioning system energy consumption, but the higher the monitoring accuracy, the higher the monitoring cost, so the monitoring accuracy should be reasonably self-adaptively adjusted.

[0083] The spatial geometric size of the central region, the initial environmental state, the spatial geometric size of the adjacent region, the energy consumption influence coefficient, the air conditioning temperature set value, and the preset unit time length are input into the pre-constructed air conditioning system energy consumption prediction model to obtain the predicted unit time energy consumption.

[0084] The initial environmental state includes the temperature, humidity, and air flow rate of the central region.

[0085] The training method of the air conditioning system energy consumption prediction model comprises:

[0086] A pre-collected air conditioning system energy consumption dataset is collected, the air conditioning system energy consumption dataset includes P sets of space geometry and initial environmental state of the central region, space geometry and energy consumption influence coefficient of the adjacent region, air conditioning temperature setting value and preset unit time length, and unit time energy consumption corresponding to P sets of space geometry and initial environmental state of the central region, space geometry and energy consumption influence coefficient of the adjacent region, air conditioning temperature setting value and preset unit time length, P is a positive integer greater than 0, and the air conditioning system energy consumption dataset is divided into a training set and a validation set, wherein the training set is used to train an air conditioning system energy consumption prediction model, and the validation set is used to evaluate the generalization performance of the air conditioning system energy consumption prediction model;

[0087] It should be noted that for the pre-collected air conditioning system energy consumption dataset, the data needs to be pre-processed, and the discrete features such as the fixed options of the unit time length are one-hot encoded; the continuous features such as temperature, geometry size and energy consumption are normalized, such as standardized to the range of [0, 1], to eliminate the influence of dimension difference on model training.

[0088] During the training process of the air conditioning system energy consumption prediction model, the cross-entropy loss function is minimized as the optimization target, the early stopping strategy is used to monitor the performance of the validation set, the network parameters are continuously adjusted to optimize the model performance; when the prediction accuracy on the validation set reaches the expected accuracy, it is considered that the air conditioning system energy consumption prediction model has converged, and the training is stopped; the air conditioning system energy consumption prediction model is trained using a deep neural network based on a multilayer perceptron;

[0089] The space geometry and initial environmental state of the central region, the space geometry and energy consumption influence coefficient of the adjacent region, the air conditioning temperature setting value and the preset unit time length are converted into a feature vector; the input layer of the air conditioning system energy consumption prediction model receives the feature vector, extracts the nonlinear relationship in the feature vector through multiple hidden layers, and finally the output layer of the air conditioning system energy consumption prediction model calculates the probability distribution of the unit time energy consumption through the softmax activation function, and outputs the unit time energy consumption corresponding to the maximum probability as the final prediction result;

[0090] It should be noted that the classification rule of unit time energy consumption E can be set according to the statistical distribution of historical data, for example, using the quantile method, arranging the unit time energy consumption data of the training set in ascending order, and dividing the data into N intervals according to the percentile, each interval corresponds to an energy consumption level. When the model is predicted, the interval where the output unit time energy consumption value falls is the energy consumption level of the time period.

[0091] It should be noted that the deep neural network of the multilayer perceptron is suitable for capturing high-dimensional and complex coupled relationships, and is convenient for capturing dynamic correlations between long-term environmental changes and energy consumption;

[0092] Obtaining a real-time unit time energy consumption curve of the central area air conditioning system;

[0093] Based on the predicted unit time energy consumption, comparing and analyzing the predicted unit time energy consumption with the real-time unit time energy consumption curve to generate an energy consumption abnormality regulation instruction;

[0094] Based on the energy consumption abnormality regulation instruction, generating a unit time adjustment instruction or an energy consumption regulation instruction;

[0095] The method for generating the energy consumption abnormality regulation instruction comprises:

[0096] Presetting energy consumption difference gradient thresholds E1 and E2, wherein 0

[0097] It should be noted that the energy consumption difference gradient thresholds E1 and E2 are determined by a person skilled in the art according to a large number of experiments;

[0098] When ΔE

[0099] When E1≤ΔE≤E2, a unit time adjustment instruction is generated;

[0100] When ΔE

[0101] The unit time adjustment instruction is to reduce the unit time level, for example, if the current unit time is 60 min, it is reduced to 30 min;

[0102] It should be noted that when ΔE

[0103] When E1≤ΔE≤E2, it indicates that the energy consumption difference between the real-time unit time energy consumption and the predicted unit time energy consumption is slightly abnormal, and at this time, the predicted and real-time obtained unit time needs to be adjusted, that is, the unit time for collection is reduced, the time for obtaining the next unit time energy consumption difference is reduced, the accuracy of judging the energy consumption abnormality is improved, and the energy consumption loss is reduced;

[0104] When ΔE

[0105] In this embodiment, the building space is reasonably refined into multiple functional areas based on functional division, and the spatial geometric dimensions and physical adjacency relationships of each area are obtained by using a BIM model, thereby effectively constructing the initial relationship of energy consumption interference between areas, analyzing two adjacency cases of wall heat conduction and corridor air exchange, calculating the energy consumption influence coefficient, and screening out the adjacent areas that have a significant impact on the central area according to the coefficient, so that the energy consumption influence relationship between areas is more accurate, and the full data collection calculation is avoided, which leads to resource waste and prediction distortion.

[0106] And by training the air conditioning system energy consumption prediction model, the difference between the predicted value and the actual value can be compared and analyzed in real time, so as to quickly identify operation abnormalities, improve the accuracy and sensitivity of energy consumption monitoring, and compared with the traditional fixed value threshold alarm mode, has the advantages of improving dynamic response and reducing false alarm rate, and when there is a slight abnormality, the unit time of prediction and real-time acquisition can be adjusted, that is, the length of the unit time of collection is reduced, the acquisition time of the energy consumption difference value of the next unit time is reduced, the accuracy of judging energy consumption abnormalities is improved, and energy consumption loss is reduced.

[0107] Embodiment 2

[0108] Please refer to Figure 2 , the embodiment of the application provides a technical scheme:

[0109] When the energy consumption abnormality control instruction is generated, the real-time temperature of each adjacent area is collected;

[0110] The characteristic parameters of the adjacent area and the characteristic parameters of the central area are input into the pre-constructed adjacent area temperature prediction model, and the predicted temperature of the adjacent area is output;

[0111] The characteristic parameters of the adjacent area include the initial temperature, the spatial geometric dimension of the adjacent area, and the energy consumption influence coefficient corresponding to the adjacent area;

[0112] The characteristic parameters of the central area include the initial temperature, the spatial geometric dimension of the central area, and the control temperature set value and the working time of the air conditioning system of the central area;

[0113] The training method of the adjacent area temperature prediction model comprises:

[0114] A pre-constructed abutment area temperature dataset includes characteristic parameters of Y groups of abutment areas and characteristic parameters of a central area, and real-time temperatures of the abutment areas corresponding to the characteristic parameters of the Y groups of abutment areas and the characteristic parameters of the central area, Y being a positive integer greater than 0, the characteristic parameters of the abutment areas including initial temperatures, spatial geometric sizes of the abutment areas, and energy consumption influence coefficients corresponding to the abutment areas, and the characteristic parameters of the central area including an initial temperature, a spatial geometric size of the central area, and a control temperature setting value and an air conditioning system working time of the central area; the abutment area temperature dataset is divided into a training set and a data validation set, wherein the training set is used for parameter learning of an abutment area temperature prediction model, and the validation set is used for real-time evaluation of the generalization ability of the abutment area temperature prediction model.

[0115] It should be noted that, for the pre-collection of the abutment area temperature dataset, the data needs to be pre-processed, the discrete features are encoded with one-hot encoding, and the continuous features such as temperature, geometric size, and energy consumption are normalized, such as standardized to the range of [0, 1], to eliminate the influence of dimension difference on model training.

[0116] In the training process of the abutment area temperature prediction model, a deep neural network structure based on a multilayer perceptron is adopted, the characteristic parameters of the abutment areas and the characteristic parameters of the central area are converted into feature vectors as inputs, nonlinear features in the data are extracted through multiple hidden layers, and finally a probability distribution of the abutment area temperature is generated in the output layer using a softmax activation function, and the abutment area temperature corresponding to the maximum probability is output as the final prediction result; the training process aims to minimize the cross-entropy loss function, and an early stopping strategy is introduced to monitor the performance of the validation set; when the prediction accuracy on the validation set reaches a preset threshold, it is considered that the abutment area temperature prediction model has converged, and the training is stopped.

[0117] For example, the temperature distribution rules of the abutment areas are shown in Table 1.

[0118] Table 1 Temperature distribution rules of abutment areas

[0119]

[0120] The predicted temperature is compared with the collected real-time temperature of the abutment area to generate an energy consumption abnormality regulation area.

[0121] The method for comparing the predicted temperature with the collected real-time temperature of the abutment area to generate an energy consumption abnormality regulation area includes:

[0122] A preset temperature difference threshold Tmax is set, and the absolute value of the difference between the predicted temperature and the collected real-time temperature of the abutment area is obtained, denoted as Tzj.

[0123] It should be noted that the temperature difference threshold Tmax is determined by a person skilled in the art according to a large number of experiments.

[0124] When Tzj > Tmax, the adjacent region is marked as an energy consumption abnormal regulation region.

[0125] When Tzj ≤ Tmax, the adjacent region is not marked.

[0126] In this embodiment, after generating the energy consumption abnormal regulation instruction, real-time temperature collection is performed on all adjacent regions, and a model for predicting the temperature of the adjacent region is constructed in combination with the spatial parameters, initial temperature state, air conditioner setting parameters and energy consumption influence coefficient of the central region and the adjacent region, so that the temperature evolution trend of the adjacent region is accurately predicted, and by setting the temperature difference threshold as an evaluation standard, the deviation value between the predicted temperature and the real-time collected temperature is calculated, the adjacent region whose actual temperature rise or temperature drop trend obviously deviates from the model expectation can be quickly identified, and is marked as an energy consumption abnormal regulation region. This way not only can identify the local load transfer caused by abnormal heat conduction or air exchange, but also can identify the adaptability of the air conditioning system control strategy in the spatial distribution, providing decision support for subsequent strategy correction and load redistribution.

[0127] To sum up, by introducing the temperature prediction and abnormal identification mechanism of the adjacent region, the adjacent region is analyzed under the energy consumption abnormality of the central region in this embodiment, the energy consumption abnormal regulation region is obtained, the active identification ability and spatial resolution ability of the energy saving regulation system are significantly enhanced, and the intelligent level and energy efficiency level of the building air conditioning system operation are improved.

[0128] In use, the adjacent region marked as the energy consumption abnormal regulation region can be regulated, including personnel flow regulation, space sealing regulation or air exchange efficiency regulation, reducing the temperature influence of the adjacent region on the central region and reducing the air conditioning energy consumption of the central region.

[0129] Embodiment 3

[0130] Please refer to Figure 4 The embodiment of the present application provides a technical solution: a public institution building energy consumption dynamic monitoring and energy saving regulation system, which is used to realize the public institution building energy consumption dynamic monitoring and energy saving regulation method, and the system comprises:

[0131] A region division module divides the internal space of the public institution building into m independent functional regions based on the functions of each region of the public institution building.

[0132] The data acquisition module acquires the spatial geometric dimensions of the m functional areas and the adjacency relationship therebetween through the BIM model of the public institution building, randomly selects one of the areas as a central area, identifies all areas adjacent to the central area and marks the adjacent areas, and acquires the initial temperature of the adjacent areas;

[0133] The area analysis module calculates the energy consumption influence coefficient of each adjacent area according to the adjacency relationship between the adjacent area and the central area, and determines whether to retain the adjacent area according to the energy consumption influence coefficient.

[0134] The energy consumption prediction module presets a unit time length, inputs the spatial geometric dimensions of the central area and the initial environmental state, the spatial geometric dimensions of the adjacent areas and the energy consumption influence coefficient, the air conditioner temperature setting value and the preset unit time length into a pre-constructed air conditioner system energy consumption prediction model, and obtains the predicted unit time energy consumption.

[0135] The energy consumption acquisition module acquires a real-time unit time energy consumption curve of the central area air conditioner system.

[0136] The instruction generation module compares and analyzes the predicted unit time energy consumption with the real-time unit time energy consumption curve based on the predicted unit time energy consumption, and generates an energy consumption abnormality regulation and control instruction.

[0137] The above shows and describes the basic principles, main features and advantages of the present application. It should be understood by those skilled in the art that the present application is not limited to the above embodiments, and the above embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for dynamic monitoring and energy-saving control of energy consumption of public institution buildings, characterized in that, The method comprises the following steps: Divide the internal space of the public institution building into m functional areas based on the functions of each area of the public institution building, wherein m is a positive integer greater than 1; Obtain the spatial geometric dimensions of the m functional areas and the adjacency relationship therebetween through a BIM model of the public institution building, randomly select one of the functional areas as a central area, identify all the functional areas having the adjacency relationship with the central area and mark them as adjacent areas, and obtain the initial temperature of the adjacent areas; Calculate the energy consumption influence coefficient of each adjacent area according to the adjacency relationship between the adjacent area and the central area, and determine whether to retain the adjacent area according to the energy consumption influence coefficient; Pre-set a unit time length; Input the spatial geometric dimensions of the central area and the initial environmental state, the spatial geometric dimensions of the adjacent areas and the energy consumption influence coefficient, the air conditioning temperature set value, and the pre-set unit time length into a pre-constructed air conditioning system energy consumption prediction model to obtain the predicted unit time energy consumption; Obtain the real-time unit time energy consumption curve of the air conditioning system of the central area; Compare and analyze the predicted unit time energy consumption and the real-time unit time energy consumption curve to generate an energy consumption anomaly regulation and control instruction; When the energy consumption anomaly regulation and control instruction is generated, collect the real-time temperature of each adjacent area; Input the feature parameters of the adjacent areas and the feature parameters of the central area into a pre-constructed adjacent area temperature prediction model to output the predicted temperature of the adjacent areas; The feature parameters of the adjacent areas include the initial temperature, the spatial geometric dimensions of the adjacent areas, and the energy consumption influence coefficient corresponding to the adjacent areas; and the feature parameters of the central area include the initial temperature, the spatial geometric dimensions of the central area, and the control temperature set value and the working time of the air conditioning system of the central area; Compare and analyze the predicted temperature and the collected real-time temperature of the adjacent areas to generate an energy consumption anomaly regulation and control area; Regulate the adjacent areas marked as the energy consumption anomaly regulation and control area, which includes personnel flow regulation, space sealing regulation, or air exchange efficiency regulation, to reduce the temperature influence of the adjacent areas on the central area and reduce the air conditioning energy consumption of the central area.

2. The method for dynamic monitoring and energy-saving control of public institution building energy consumption according to claim 1, characterized in that, The adjacency relationship includes wall adjacency and corridor connection.

3. The method for dynamic monitoring and energy-saving control of public institution building energy consumption according to claim 2, characterized in that, The method for calculating the energy consumption influence coefficient corresponding to each adjacent area according to the adjacency relationship between the adjacent area and the central area comprises the following steps: When the adjacency relationship is wall adjacency, the method for obtaining the energy consumption influence coefficient of the central area and the adjacent area comprises the following steps: Obtain the heat conductivity coefficient of the wall between the central area and the adjacent area, and the contact area of the central area and the adjacent area and the thickness of the wall, and calculate the energy consumption influence coefficient; When the adjacency relationship is corridor connection, the method for obtaining the energy consumption influence coefficient of the central area and the adjacent area comprises the following steps: Obtain the air exchange rate of the corridor between the central area and the adjacent area, obtain the cross-sectional area of the corridor between the central area and the adjacent area, and calculate the energy consumption influence coefficient.

4. The method for dynamic monitoring and energy-saving control of public institution building energy consumption according to claim 1, characterized in that, The method for determining whether to retain the adjacent area according to the energy consumption influence coefficient comprises the following steps: A preset energy consumption influence coefficient threshold is set. When the energy consumption influence coefficient of the center region and the adjacent region obtained is less than the preset energy consumption influence coefficient threshold, the adjacent region is removed. When the energy consumption influence coefficient of the center region and the adjacent region obtained is greater than or equal to the preset energy consumption influence coefficient threshold, the adjacent region is retained, and a one-to-one correspondence between the adjacent region and the energy consumption influence coefficient is established.

5. The method for dynamic monitoring and energy-saving control of public institution building energy consumption according to claim 1, characterized in that, The method for generating the energy consumption anomaly regulation instruction comprises: A preset energy consumption difference gradient threshold E1 and E2 are set, wherein 0 < E1 < E2; the energy consumption difference ΔE of the real-time unit time energy consumption and the predicted unit time energy consumption is calculated; When ΔE < E1, no energy consumption anomaly regulation instruction is generated; When E1 ≤ ΔE ≤ E2, a unit time adjustment instruction is generated; When ΔE > E2, an energy consumption anomaly regulation instruction is generated; The unit time adjustment instruction is to reduce the unit time length.

6. The method for dynamic monitoring and energy-saving control of public institution building energy consumption according to claim 1, characterized in that, The training method of the air conditioning system energy consumption prediction model comprises: An air conditioning system energy consumption dataset is collected in advance, which comprises the spatial geometric size and initial environment state of P groups of center regions, the spatial geometric size and energy consumption influence coefficient of adjacent regions, an air conditioning temperature setting value, and a preset unit time length, and the unit time energy consumption corresponding to the spatial geometric size and initial environment state of the P groups of center regions, the spatial geometric size and energy consumption influence coefficient of adjacent regions, the air conditioning temperature setting value, and the preset unit time length, P is a positive integer greater than 0, and the air conditioning system energy consumption dataset is divided into a training set and a verification set, wherein the training set is used to train the air conditioning system energy consumption prediction model, and the verification set is used to evaluate the generalization performance of the air conditioning system energy consumption prediction model; During the training of the air conditioning system energy consumption prediction model, the minimum cross-entropy loss function is used as the optimization target, the early stopping strategy is used to monitor the performance of the verification set, the model performance is optimized by continuously adjusting the network parameters, and when the prediction accuracy on the verification set reaches the expected accuracy, it is considered that the air conditioning system energy consumption prediction model has tended to converge, and the training is stopped; the air conditioning system energy consumption prediction model is trained using a deep neural network based on a multilayer perceptron; The spatial geometric size and initial environment state of the center region, the spatial geometric size and energy consumption influence coefficient of the adjacent region, the air conditioning temperature setting value, and the preset unit time length are converted into a feature vector; the input layer of the air conditioning system energy consumption prediction model receives the feature vector, extracts the nonlinear relationship in the feature vector through multiple hidden layers, and finally the output layer of the air conditioning system energy consumption prediction model calculates the probability distribution of the unit time energy consumption through a softmax activation function, and outputs the unit time energy consumption corresponding to the maximum probability as the final prediction result.

7. The method for dynamic monitoring and energy-saving control of public institution building energy consumption according to claim 1, characterized in that, The method for comparing and analyzing the predicted temperature with the collected real-time temperature of the adjacent region to generate an energy consumption anomaly regulation region comprises: A preset temperature difference threshold Tmax is set, the absolute value of the difference between the predicted temperature and the collected real-time temperature of the adjacent region is obtained, and is denoted as Tzj; When Tzj > Tmax, the adjacent region is marked as an energy consumption anomaly regulation region; When Tzj ≤ Tmax, the adjacent region is not marked.

8. The method for dynamic monitoring and energy-saving control of public institution building energy consumption according to claim 7, characterized in that, The training method of the abutment area temperature prediction model comprises the following steps: Pre-construct an abutment area temperature dataset, which comprises feature parameters of Y groups of abutment areas and feature parameters of a central area, and real-time temperatures of the abutment areas corresponding to the feature parameters of the Y groups of abutment areas and the feature parameters of the central area, Y being a positive integer greater than 0; divide the abutment area temperature dataset into a training set and a data validation set, wherein the training set is used for parameter learning of the abutment area temperature prediction model, and the validation set is used for real-time evaluation of the generalization ability of the abutment area temperature prediction model; During the training process of the abutment area temperature prediction model, a deep neural network structure based on a multilayer perceptron is adopted to convert the feature parameters of the abutment areas and the feature parameters of the central area into feature vectors as inputs, extract nonlinear features in the data through multiple hidden layers, and finally generate a probability distribution of the abutment area temperature in the output layer using a softmax activation function, output the abutment area temperature corresponding to the maximum probability as the final prediction result; the training process aims to minimize the cross-entropy loss function, and an early stopping strategy is introduced to monitor the performance of the validation set; when the prediction accuracy on the validation set reaches a preset threshold, it is considered that the abutment area temperature prediction model has converged, and the training is stopped.

9. A public institution building energy consumption dynamic monitoring and energy saving control system for implementing the public institution building energy consumption dynamic monitoring and energy saving control method of any one of claims 1-8, characterized in that, The system comprises: A region division module, which divides the internal space of a public institution building into m functional regions based on the functions of each region of the public institution building; A data acquisition module, which acquires the spatial geometric dimensions of the m functional regions and their abutment relationships through a BIM model of the public institution building, randomly selects one of the functional regions as a central region, identifies all the functional regions adjacent to the central region and marks them as abutment regions, and acquires the initial temperatures of the abutment regions; A region analysis module, which calculates the energy consumption influence coefficient of each abutment region according to its abutment relationship with the central region, and determines whether to retain the abutment region according to the energy consumption influence coefficient; An energy consumption prediction module, which presets a unit time length, inputs the spatial geometric dimensions of the central region and the initial environmental state, the spatial geometric dimensions of the abutment regions and the energy consumption influence coefficient, the air conditioning temperature set value, and the preset unit time length into a pre-constructed air conditioning system energy consumption prediction model, and obtains the predicted unit time energy consumption; An energy consumption acquisition module, which acquires the real-time unit time energy consumption curve of the air conditioning system of the central region; An instruction generation module, which compares and analyzes the predicted unit time energy consumption and the real-time unit time energy consumption curve to generate an energy consumption abnormality regulation and control instruction.

Images