Building performance prediction method, device, equipment and medium

Abstract

The present invention relates to the field of artificial intelligence technology, and provides a building performance prediction method, device, equipment and medium. The method can be based on a multi-task learning model with a dual-head input and four-task output architecture, and a multi-task building energy prediction model trained with existing building data within a specified longitude and latitude range, so as to achieve simultaneous prediction of multiple building performance indicators, thereby improving prediction efficiency. The target building characteristics and target meteorological parameters of the target building are input into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted rooftop photovoltaic power generation, predicted south facade photovoltaic power generation and predicted east and west facade photovoltaic power generation of the target building, and further generate a target building performance prediction result for the target building, so as to achieve rapid and accurate simulation of building performance by means of artificial intelligence.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a building performance prediction method, device, equipment and medium. Background Art

[0002] The urban heat island effect has a significant impact on the energy consumption of urban buildings. The urban heat island effect refers to the phenomenon that the temperature in the city is significantly higher than that in the surrounding suburbs due to factors such as large amounts of artificial heating, high heat storage bodies such as buildings and roads, and reduced green space.

[0003] Therefore, in order to reduce the impact of the urban heat island effect, it is necessary to simulate building performance to make urban buildings more rational.

[0004] However, the building performance simulation method in the prior art is complex, time-consuming, and has low accuracy. In addition, each prediction can only be performed for a single task and a single indicator, that is, the single-task agent model needs to be trained and loaded for each target separately, which significantly increases the computing cost and time overhead, so the prediction efficiency is also low. For example: when it is necessary to predict the building energy consumption intensity, roof photovoltaic power generation, south facade photovoltaic power generation, and east and west facade photovoltaic power generation, it is necessary to train an independent prediction model for the building energy consumption intensity, roof photovoltaic power generation, south facade photovoltaic power generation, and east and west facade photovoltaic power generation, which causes the calculation time to increase exponentially. Summary of the invention

[0005] In view of the above, it is necessary to provide a building performance prediction method, device, equipment and medium to solve the problems of low efficiency and low accuracy of building performance prediction.

[0006] A building performance prediction method, the building performance prediction method comprising:

[0007] Construct an initial prediction model based on a multi-task learning model with a dual-head input and four-task output architecture;

[0008] Acquire existing building data within a specified longitude and latitude range, and use the existing building data to train the initial prediction model to obtain a multi-task building energy prediction model;

[0009] In response to a building energy prediction instruction for a target building, obtaining a target building feature of the target building and a target meteorological parameter of a location where the target building is located;

[0010] Inputting the target building characteristics and the target meteorological parameters into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted roof photovoltaic power generation, predicted south facade photovoltaic power generation, and predicted east and west facade photovoltaic power generation of the target building;

[0011] A target building performance prediction result of the target building is generated according to the predicted building energy consumption intensity, the predicted rooftop photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation.

[0012] According to a preferred embodiment of the present invention, the multi-task learning model based on the dual-head input and four-task output architecture constructs an initial prediction model, including:

[0013] Constructing an architectural feature path with architectural features as input and a meteorological parameter path with meteorological parameters as input; wherein the architectural features include at least one indicator and architectural function features under architectural morphological parameters, and the meteorological parameters include at least one meteorological indicator; wherein the architectural feature path includes two layers of one-dimensional convolutional neural network 1D CNN, a batch normalization layer Batch Normalization, a random dropout layer Dropout, and a global average pooling layer Global Average Pooling connected end to end; the meteorological parameter path includes a positional encoding layer Positional Encoding, two long short-term memory networks LSTM, and a multi-head attention mechanism Multi-Head Attention connected end to end; wherein the input feature dimension of the architectural feature path is (X, 1), X represents the total number of indicators of the architectural features; the input feature dimension of the meteorological parameter path is (8760, Y), Y represents the total number of indicators of the meteorological parameters, and 8760 represents 8760 hours in a year;

[0014] A feature splicing layer is constructed to splice the output of the building feature path and the output of the meteorological parameter path, and the spliced ​​features are used as the input of the multi-task output layer to obtain the initial prediction model; wherein the output of the multi-task output layer includes building energy consumption intensity, rooftop photovoltaic power generation, south facade photovoltaic power generation, and east and west facade photovoltaic power generation.

[0015] According to a preferred embodiment of the present invention, the using of the existing building data to train the initial prediction model to obtain a multi-task building energy prediction model comprises:

[0016] The existing building data is split according to the configuration ratio to obtain a training set and a validation set;

[0017] Construct loss function;

[0018] The initial prediction model is trained by back propagation using the training set and the loss function; wherein the learning rate of the model is adaptively optimized during each round of training using the Adam optimizer;

[0019] During the training process, when it is detected that the loss function reaches convergence, the training is stopped, and the current model is verified using the verification set;

[0020] When the current model passes the verification, the current model is determined as the multi-task building energy prediction model.

[0021] According to a preferred embodiment of the present invention, constructing a loss function includes:

[0022] The loss function is constructed using a peak constraint module and a mean square error;

[0023] The peak constraint module uses the PEAK function to increase the penalty for prediction errors between peaks and valleys; the mean square error is used to capture the deviation between the predicted value and the actual value of the model.

[0024] According to a preferred embodiment of the present invention, the verifying the current model using the verification set includes:

[0025] Inputting the data in the validation set into the current model in sequence for prediction to obtain a prediction result;

[0026] Calculate the root mean square error coefficient of variation and the coefficient of determination of the current model according to the prediction results;

[0027] When the root mean square error coefficient of variation is less than a first threshold value, and the determination coefficient is greater than a second threshold value, it is determined that the current model passes the verification; or

[0028] When the root mean square error coefficient of variation is greater than or equal to the second threshold, and / or the determination coefficient is less than or equal to the second threshold, it is determined that the current model has not passed the verification.

[0029] According to a preferred embodiment of the present invention, generating a target building performance prediction result of the target building according to the predicted building energy consumption intensity, the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation includes:

[0030] Determining the estimated energy consumption value of the target building according to the predicted building energy consumption intensity;

[0031] Determine the estimated energy saving of the target building after installing photovoltaic power generation according to the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation;

[0032] The target building performance prediction result is generated according to the predicted energy consumption value and the predicted energy saving amount.

[0033] According to a preferred embodiment of the present invention, after generating the target building performance prediction result of the target building according to the predicted building energy consumption intensity, the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation, the method further includes:

[0034] When the predicted building energy consumption intensity is greater than or equal to a third threshold, sending a renovation prompt for the target building to a designated user;

[0035] When receiving a remodeled building of the target building based on the remodeling prompt feedback, predicting the remodeled building energy consumption intensity of the remodeled building using the multi-task building energy prediction model;

[0036] When the energy consumption intensity of the renovated building is less than the predicted building energy consumption intensity, calculating the difference between the predicted building energy consumption intensity and the renovated building energy consumption intensity as the energy consumption optimization value of the renovated building;

[0037] The energy consumption optimization value is sent to the designated user.

[0038] A building performance prediction device, comprising:

[0039] A construction unit for constructing an initial prediction model based on a multi-task learning model with a dual-head input and four-task output architecture;

[0040] A training unit, used to obtain existing building data within a specified longitude and latitude range, and use the existing building data to train the initial prediction model to obtain a multi-task building energy prediction model;

[0041] an acquisition unit, configured to acquire, in response to a building energy prediction instruction for a target building, a target building feature of the target building and a target meteorological parameter of a location where the target building is located;

[0042] An input unit, used to input the target building characteristics and the target meteorological parameters into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted roof photovoltaic power generation, predicted south facade photovoltaic power generation, and predicted east and west facade photovoltaic power generation of the target building;

[0043] A generating unit is used to generate a target building performance prediction result of the target building according to the predicted building energy consumption intensity, the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation.

[0044] A computer device, comprising:

[0045] a memory storing at least one instruction; and

[0046] A processor executes the instructions stored in the memory to implement the building performance prediction method.

[0047] A computer-readable storage medium stores at least one instruction, and the at least one instruction is executed by a processor in a computer device to implement the building performance prediction method.

[0048] It can be seen from the above technical solutions that the present invention can obtain a multi-task building energy prediction model based on a multi-task learning model with a dual-head input and four-task output architecture, and training of existing building data within a specified longitude and latitude range, so as to realize simultaneous prediction of multiple building performance indicators and improve prediction efficiency; the target building characteristics and target meteorological parameters of the target building are input into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted rooftop photovoltaic power generation, predicted south facade photovoltaic power generation and predicted east and west facade photovoltaic power generation of the target building, and further generate the target building performance prediction result of the target building, so as to realize fast and accurate simulation of building performance based on artificial intelligence means. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a flow chart of a preferred embodiment of the building performance prediction method of the present invention;

[0050] Figure 2 is a schematic diagram of the architectural feature path of the present invention;

[0051] Figure 3 is a schematic diagram of the meteorological parameter path of the present invention;

[0052] Figure 4 is a schematic diagram of the initial prediction model of the present invention;

[0053] Figure 5 It is the comparison result of the performance of the common model of the present invention and the multi-task building energy prediction model;

[0054] Figure 6 It is a functional module diagram of a preferred embodiment of the building performance prediction device of the present invention;

[0055] Figure 7 It is a schematic diagram of the structure of a computer device of a preferred embodiment of the method for realizing building performance prediction of the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0057] like Figure 1FIG. 1 is a flow chart of a preferred embodiment of the method for predicting building performance according to the present invention. According to different requirements, the order of the steps in the flow chart can be changed, and some steps can be omitted.

[0058] The building performance prediction method is applied to one or more computer devices, which are devices that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and their hardware includes but is not limited to microprocessors, application specific integrated circuits (ASICs), programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0059] The computer device may also include a server, which may be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0060] Among them, Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0061] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0062] The network where the computer device is located includes but is not limited to the Internet, a wide area network, a metropolitan area network, a local area network, a virtual private network (VPN), etc.

[0063] S10, build the initial prediction model based on the multi-task learning model with dual-head input and four-task output architecture.

[0064] In this embodiment, the multi-task learning model based on the dual-head input and four-task output architecture constructs the initial prediction model, including:

[0065] Constructing an architectural feature path with architectural features as input and a meteorological parameter path with meteorological parameters as input; wherein the architectural features include at least one indicator and architectural function characteristics under architectural form parameters, and the meteorological parameters include at least one meteorological indicator; wherein the architectural feature path includes two layers of one-dimensional convolutional neural networks 1D (One-dimensional) CNN (Convolutional Neural Networks), a batch normalization layer Batch Normalization (BN), a random dropout layer Dropout, and a global average pooling layer Global Average Pooling connected end to end; the meteorological parameter path includes a positional embedding layer PositionalEncoding, two long short-term memory networks LSTM (Long Short-Term Memory), and a multi-head attention mechanism Multi-Head Attention connected end to end; wherein the input feature dimension of the architectural feature path is (X, 1), X represents the total number of indicators of the architectural features; the input feature dimension of the meteorological parameter path is (8760, Y), Y represents the total number of indicators of the meteorological parameters, and 8760 represents 8760 hours in a year;

[0066] A feature splicing layer is constructed to splice the output of the building feature path and the output of the meteorological parameter path, and the spliced ​​features are used as the input of the multi-task output layer to obtain the initial prediction model; wherein the output of the multi-task output layer includes building energy consumption intensity, rooftop photovoltaic power generation, south facade photovoltaic power generation, and east and west facade photovoltaic power generation.

[0067] Among them, at least one indicator under the building form parameters may include, but is not limited to, a combination of one or more of the following indicators: total building area, building volume, building height, window area, shape coefficient, number of surrounding buildings, building density, distance between buildings, sky visibility factor, perimeter area ratio, etc.

[0068] The building functional characteristics may be used to characterize the functional attributes of the building, such as exhibition function, office function, residential function, hospital function, etc.

[0069] The meteorological index may include, but is not limited to, a combination of one or more of the following indexes: temperature, humidity, wind speed and wind direction, etc.

[0070] The building features are unique features among various building samples, and the meteorological parameters are features shared among all samples.

[0071] See also Figure 2-4,in, Figure 2 is a schematic diagram of the architectural feature path of the present invention, Figure 3 is a schematic diagram of the meteorological parameter path of the present invention, Figure 4 It is a schematic diagram of the initial prediction model of the present invention. Specifically, Building features represent building feature paths, Weather features represent meteorological parameter paths, Building features can be used to perform vector feature extraction on building features, and Weather features can be used to perform time series feature extraction on meteorological parameters; Feature concatenation represents the feature concatenation layer, Multi-task output represents the multi-task output layer, EUI (Energy Use Intensity) represents the building energy consumption intensity (the dimension of EUI is (1, 8760), and the dimensions of Rooftop PV, R+S PV, and R+SEW PV are also (1, 8760)), Rooftop PV represents the rooftop photovoltaic power generation, R+S PV represents the south facade photovoltaic power generation, R+SEW PV represents the photovoltaic power generation of the east and west facades; the input feature dimension of the building feature path is (129, 1), 129 represents the indicators of 129 building features such as total building area, building volume, building height, window area, and shape coefficient; the input feature dimension of the meteorological parameter path is (8760, 4), 8760 represents 8760 hours in a year, and 4 represents the indicators of four meteorological parameters: temperature, humidity, wind speed, and wind direction, that is, the prediction step is in hours. Compared with the existing solution that can only simulate the energy consumption throughout the year, the prediction accuracy is higher, so the prediction is more accurate. Kernel represents the convolution kernel of the one-dimensional convolutional neural network, sigmoid represents the activation function, and tanh represents the hyperbolic tangent function. Specifically, in the forward propagation process of the model, Buildingfeatures and Weather features are used to extract features of building features and meteorological parameters, respectively, and then the extracted features are spliced ​​using the feature splicing layer. The spliced ​​comprehensive feature representation is then passed through the multi-task output layer to simultaneously predict four variables: building energy consumption intensity, rooftop photovoltaic power generation, south facade photovoltaic power generation, and east and west facade photovoltaic power generation.

[0072] certainly, Figure 2-4 The dimensions of the data are only examples. In actual use, they can be configured according to the actual properties of the data.

[0073] In the above embodiment, the local patterns in the architectural features can be extracted through the constructed architectural feature paths, so as to further improve the richness and accuracy of the feature expression.

[0074] In the above embodiment, the two long short-term memory networks LSTM in the meteorological parameter path can capture the long-term dependencies in the time series, and the multi-head attention mechanism can find the global dependency pattern in the entire sequence. Specifically, first, the input meteorological parameters pass through the position embedding layer, then the time series features are extracted through two LSTMs, and finally the multi-head attention mechanism is added, so that the model can calculate attention in different subspaces, thereby enhancing the diversity and expression ability of feature representation.

[0075] S11, obtaining existing building data within a specified longitude and latitude range, and using the existing building data to train the initial prediction model to obtain a multi-task building energy prediction model.

[0076] In this embodiment, the specified latitude and longitude range affects the lighting characteristics of the area where the building is located. For example: for buildings in the southern hemisphere, the lighting pattern is similar, so the specified longitude and latitude range can include the longitude and latitude range of the southern hemisphere, and accordingly, the output of the multi-task building energy prediction model includes the building energy consumption intensity, roof photovoltaic power generation, north facade photovoltaic power generation, and east and west facade photovoltaic power generation; and for buildings in the northern hemisphere, the lighting pattern is similar, so the specified longitude and latitude range can include the longitude and latitude range of the northern hemisphere, and accordingly, the output of the multi-task building energy prediction model includes the building energy consumption intensity, roof photovoltaic power generation, south facade photovoltaic power generation, and east and west facade photovoltaic power generation.

[0077] In this embodiment, the existing building data may include building features, meteorological parameters, and other data that will be used in subsequent model training.

[0078] In this embodiment, the using of the existing building data to train the initial prediction model to obtain a multi-task building energy prediction model includes:

[0079] The existing building data is split according to the configuration ratio to obtain a training set and a validation set;

[0080] Construct loss function;

[0081] The initial prediction model is trained by back propagation using the training set and the loss function; wherein the learning rate of the model is adaptively optimized during each round of training using the Adam optimizer;

[0082] During the training process, when it is detected that the loss function reaches convergence, the training is stopped, and the current model is verified using the verification set;

[0083] When the current model passes the verification, the current model is determined as the multi-task building energy prediction model.

[0084] Among them, the configuration ratio can be configured according to actual needs, such as 80% / 20%.

[0085] Among them, the Adam optimizer can effectively handle optimization problems in high-dimensional parameter space.

[0086] Specifically, the construction of the loss function includes:

[0087] The loss function is constructed using a peak constraint module and a mean square error (MSE);

[0088] The peak constraint module uses the PEAK function to increase the penalty for prediction errors between peaks and valleys; the mean square error is used to capture the deviation between the predicted value and the actual value of the model.

[0089] In the above embodiment, the peak error in prediction is focused on based on the peak constraint module, that is, during the training process, the penalty for peak and trough prediction errors is increased to reduce the deviation of the model at these key points, thereby improving the performance of the model in peak prediction.

[0090] In the above embodiment, the deviation between the model prediction value and the actual value can be captured based on the mean square error, and the updating of the model parameters can be guided.

[0091] Specifically, the verifying the current model using the verification set includes:

[0092] Inputting the data in the validation set into the current model in sequence for prediction to obtain a prediction result;

[0093] Calculate the coefficient of variation of the root mean squared error (CVRMSE) and the coefficient of determination (R-squared, R²) of the current model according to the prediction results;

[0094] When the root mean square error coefficient of variation is less than a first threshold value, and the determination coefficient is greater than a second threshold value, it is determined that the current model passes the verification; or

[0095] When the root mean square error coefficient of variation is greater than or equal to the second threshold, and / or the determination coefficient is less than or equal to the second threshold, it is determined that the current model has not passed the verification.

[0096] The first threshold and the second threshold can be configured according to actual needs. For example, the first threshold can be configured as 1, and the second threshold can be configured as 0.75.

[0097] In the above embodiment, the root mean square error coefficient of variation is a dimensionless error metric that can effectively measure the ratio of the prediction error to the actual data average, so that the errors on different data sets can be directly compared. A lower root mean square error coefficient of variation value indicates that the model's error in the prediction is less variable relative to the actual value, which means that the model has better prediction accuracy.

[0098] In the above embodiment, the coefficient of determination provides a measure of the degree of fit between the model prediction results and the actual values.

[0099] Therefore, combining the root mean square error coefficient of variation and the determination coefficient can effectively evaluate the performance of the model to verify the generalization ability and prediction effect of the model.

[0100] Of course, the model may also be verified using the verification set at the end of each training cycle, which is not limited in the present invention.

[0101] In order to verify the model performance, the most basic model and the model in this case were compared. The comparison results can be found in Figure 5 , Figure 5 It is the comparison result of the performance of the common model of the present invention and the multi-task building energy prediction model. It can be seen from the above performance comparison results that after adding the peak constraint module to the model, the model performance can reach the best, and the peak constraint significantly reduces the prediction error of the model at the peak and trough points, so that the CVRMSE is further reduced and the R² value reaches the highest. This shows that the peak constraint module effectively improves the prediction accuracy of the model at extreme values ​​and enhances the reliability of the overall prediction. In addition, with the complexity of the model structure and the enhancement of the feature processing capability, the prediction performance of the model is gradually improved. From the most basic model to the final model containing CNN, LSTM, multi-head attention and peak constraint, each step of improvement significantly improves the accuracy and fitting ability of the model. The multi-task learning framework performs well in the prediction of building energy consumption and photovoltaic power generation. Through reasonable design and optimization, it can provide more accurate and reliable predictions, providing strong support for the building design and optimization process. This series of improvements verifies the importance and effectiveness of using complex model structures and advanced feature processing mechanisms in multi-objective optimization problems.

[0102] S12, in response to a building energy prediction instruction for a target building, obtaining target building characteristics of the target building and target meteorological parameters of a location where the target building is located.

[0103] In this embodiment, the building energy prediction instruction may be triggered by the designer of the target building.

[0104] In this embodiment, the target building characteristics of the target building may include the total building area, building volume, building height, window area, shape coefficient, number of surrounding buildings, building density, distance between buildings, sky visibility factor, perimeter area ratio and other characteristics of the target building.

[0105] In this embodiment, the target meteorological parameters may include parameters such as temperature, humidity, wind speed and wind direction at the location of the target building.

[0106] S13, inputting the target building characteristics and the target meteorological parameters into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted rooftop photovoltaic power generation, predicted south facade photovoltaic power generation and predicted east and west facade photovoltaic power generation of the target building.

[0107] Through the above embodiments, the architectural characteristics and meteorological parameters of the target building can be processed based on the trained multi-task building energy prediction model to simultaneously predict the building energy consumption intensity, rooftop photovoltaic power generation, south facade photovoltaic power generation, and east and west facade photovoltaic power generation of the target building. Artificial intelligence-based means can not only efficiently utilize shared information, but also reduce system complexity and improve computing efficiency.

[0108] S14, generating a target building performance prediction result of the target building according to the predicted building energy consumption intensity, the predicted rooftop photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation.

[0109] In this embodiment, the target building performance prediction result of the target building is generated according to the predicted building energy consumption intensity, the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation, and the predicted east and west facade photovoltaic power generation, including:

[0110] Determining the estimated energy consumption value of the target building according to the predicted building energy consumption intensity;

[0111] Determine the estimated energy saving of the target building after installing photovoltaic power generation according to the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation;

[0112] The target building performance prediction result is generated according to the predicted energy consumption value and the predicted energy saving amount.

[0113] Through the above embodiment, not only the energy consumption value of the building can be simulated, but also the estimated energy saving after installing photovoltaics can be given to assist in judging the effect of installing photovoltaic panels for reference by designers.

[0114] In this embodiment, after generating the target building performance prediction result of the target building according to the predicted building energy consumption intensity, the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation, the method further includes:

[0115] When the predicted building energy consumption intensity is greater than or equal to a third threshold, sending a renovation prompt for the target building to a designated user;

[0116] When receiving a remodeled building of the target building based on the remodeling prompt feedback, predicting the remodeled building energy consumption intensity of the remodeled building using the multi-task building energy prediction model;

[0117] When the energy consumption intensity of the renovated building is less than the predicted building energy consumption intensity, calculating the difference between the predicted building energy consumption intensity and the renovated building energy consumption intensity as the energy consumption optimization value of the renovated building;

[0118] The energy consumption optimization value is sent to the designated user.

[0119] The third threshold is used to reflect whether the energy consumption of the building exceeds the standard.

[0120] The designated user may be an architectural designer or the like.

[0121] For example: when the building energy consumption intensity M of the initial building exceeds the third threshold, it means that the energy consumption of the initial building exceeds the standard. At this time, the designer can be prompted to redesign and renovate the initial building, and simulate the energy consumption intensity N of the renovated building. If N is less than M, it means that the energy consumption value can be reduced after the renovation. The reduced energy consumption NM can be further fed back to the designer for reference whether to renovate the building again or whether to implement the renovation.

[0122] It can be seen from the above technical solutions that the present invention can obtain a multi-task building energy prediction model based on a multi-task learning model with a dual-head input and four-task output architecture, and training of existing building data within a specified longitude and latitude range, so as to realize simultaneous prediction of multiple building performance indicators and improve prediction efficiency; the target building characteristics and target meteorological parameters of the target building are input into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted rooftop photovoltaic power generation, predicted south facade photovoltaic power generation and predicted east and west facade photovoltaic power generation of the target building, and further generate the target building performance prediction result of the target building, so as to realize fast and accurate simulation of building performance based on artificial intelligence means.

[0123] like Figure 6, which is a functional module diagram of a preferred embodiment of the building performance prediction device of the present invention. The building performance prediction device 11 includes a construction unit 110, a training unit 111, an acquisition unit 112, an input unit 113, and a generation unit 114. The module / unit referred to in the present invention refers to a series of computer program segments that can be executed by a processor and can complete fixed functions, which are stored in a memory. In this embodiment, the functions of each module / unit will be described in detail in subsequent embodiments.

[0124] The construction unit 110 is used to construct an initial prediction model based on a multi-task learning model with a dual-head input and four-task output architecture;

[0125] The training unit 111 is used to obtain existing building data within a specified longitude and latitude range, and use the existing building data to train the initial prediction model to obtain a multi-task building energy prediction model;

[0126] The acquisition unit 112 is used to acquire the target building characteristics of the target building and the target meteorological parameters of the location of the target building in response to the building energy prediction instruction for the target building;

[0127] The input unit 113 is used to input the target building characteristics and the target meteorological parameters into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted roof photovoltaic power generation, predicted south facade photovoltaic power generation, and predicted east and west facade photovoltaic power generation of the target building;

[0128] The generating unit 114 is used to generate a target building performance prediction result of the target building according to the predicted building energy consumption intensity, the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation.

[0129] It can be seen from the above technical solutions that the present invention can obtain a multi-task building energy prediction model based on a multi-task learning model with a dual-head input and four-task output architecture, and training of existing building data within a specified longitude and latitude range, so as to realize simultaneous prediction of multiple building performance indicators and improve prediction efficiency; the target building characteristics and target meteorological parameters of the target building are input into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted rooftop photovoltaic power generation, predicted south facade photovoltaic power generation and predicted east and west facade photovoltaic power generation of the target building, and further generate the target building performance prediction result of the target building, so as to realize fast and accurate simulation of building performance based on artificial intelligence means.

[0130] like Figure 7 FIG. 1 is a schematic diagram of the structure of a computer device of a preferred embodiment of the method for predicting building performance according to the present invention.

[0131] The computer device 1 may include a memory 12 , a processor 13 , and a bus, and may also include a computer program stored in the memory 12 and executable on the processor 13 , such as a building performance prediction program.

[0132] Those skilled in the art will appreciate that the schematic diagram is merely an example of the computer device 1 and does not constitute a limitation on the computer device 1. The computer device 1 may be a bus-type structure or a star-type structure. The computer device 1 may also include more or less other hardware or software than shown in the diagram, or a different arrangement of components. For example, the computer device 1 may also include input and output devices, network access devices, etc.

[0133] It should be noted that the computer device 1 is only an example, and other existing or future electronic products that are suitable for the present invention should also be included in the protection scope of the present invention and included here by reference.

[0134] The memory 12 includes at least one type of readable storage medium, which includes flash memory, mobile hard disk, multimedia card, card-type memory (such as SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 12 can be an internal storage unit of the computer device 1, such as a mobile hard disk of the computer device 1. In other embodiments, the memory 12 can also be an external storage device of the computer device 1, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the computer device 1. Further, the memory 12 can also include both an internal storage unit of the computer device 1 and an external storage device. The memory 12 can not only be used to store application software and various types of data installed in the computer device 1, such as the code of the building performance prediction program, but also can be used to temporarily store data that has been output or is to be output.

[0135] In some embodiments, the processor 13 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips. The processor 13 is the control core (Control Unit) of the computer device 1, and uses various interfaces and lines to connect various components of the entire computer device 1, and executes various functions and processes data of the computer device 1 by running or executing programs or modules (such as executing building performance prediction programs, etc.) stored in the memory 12, and calling data stored in the memory 12.

[0136] The processor 13 executes the operating system of the computer device 1 and various installed applications. The processor 13 executes the applications to implement the steps in the above-mentioned various building performance prediction method embodiments, for example Figure 1 Steps shown.

[0137] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory 12 and executed by the processor 13 to complete the present invention. The one or more modules / units may be a series of computer-readable instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program in the computer device 1. For example, the computer program may be divided into a construction unit 110, a training unit 111, an acquisition unit 112, an input unit 113, and a generation unit 114.

[0138] The above-mentioned integrated unit implemented in the form of software function modules can be stored in a computer-readable storage medium. The above-mentioned software function modules are stored in a storage medium and include several instructions for enabling a computer device (which can be a personal computer, a computer device, or a network device, etc.) or a processor to execute the part of the building performance prediction method described in each embodiment of the present invention.

[0139] If the module / unit integrated in the computer device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware devices through a computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of each of the above-mentioned method embodiments can be implemented.

[0140] The computer program includes computer program code, which may be in source code form, object code form, executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory, etc.

[0141] Furthermore, the computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the blockchain node, etc.

[0142] The blockchain referred to in this invention is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the blockchain underlying platform, platform product service layer, and application service layer.

[0143] The bus can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 The bus is represented by only one straight line, but it does not mean that there is only one bus or one type of bus. The bus is configured to realize the connection and communication between the memory 12 and at least one processor 13, etc.

[0144] Although not shown, the computer device 1 may also include a power source (such as a battery) for supplying power to various components. Preferably, the power source may be logically connected to the at least one processor 13 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include any components such as one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, and power status indicators. The computer device 1 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0145] Furthermore, the computer device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is usually used to establish a communication connection between the computer device 1 and other computer devices.

[0146] Optionally, the computer device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode) touch device. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the computer device 1 and to display a visual user interface.

[0147] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0148] Figure 7 Only the computer device 1 having components 12-13 is shown, and it can be understood by those skilled in the art that Figure 7 The structure shown does not constitute a limitation on the computer device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0149] Combination Figure 1 , the memory 12 in the computer device 1 stores a plurality of instructions to implement a building performance prediction method, and the processor 13 can execute the plurality of instructions to implement:

[0150] Construct an initial prediction model based on a multi-task learning model with a dual-head input and four-task output architecture;

[0151] Acquire existing building data within a specified longitude and latitude range, and use the existing building data to train the initial prediction model to obtain a multi-task building energy prediction model;

[0152] In response to a building energy prediction instruction for a target building, obtaining a target building feature of the target building and a target meteorological parameter of a location where the target building is located;

[0153] Inputting the target building characteristics and the target meteorological parameters into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted roof photovoltaic power generation, predicted south facade photovoltaic power generation, and predicted east and west facade photovoltaic power generation of the target building;

[0154] A target building performance prediction result of the target building is generated according to the predicted building energy consumption intensity, the predicted rooftop photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation.

[0155] Specifically, the specific implementation method of the processor 13 for the above instructions can refer to Figure 1 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0156] It should be noted that the data involved in this case were all obtained legally.

[0157] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0158] The present invention can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present invention can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present invention can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0159] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0160] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0161] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0162] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0163] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the present invention can also be implemented by one unit or device through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any particular order.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A building performance prediction method, characterized in that: The building performance prediction method comprises: An initial prediction model is constructed based on a multi-task learning model with a dual-head input and four-task output architecture, including: constructing an architectural feature path with architectural features as input, and a meteorological parameter path with meteorological parameters as input; wherein the architectural features include at least one indicator and architectural functional features under architectural morphological parameters, and the meteorological parameters include at least one meteorological indicator; wherein the architectural feature path includes two layers of one-dimensional convolutional neural networks 1D CNN, a batch normalization layer Batch Normalization, a random dropout layer Dropout, and a global average pooling layer GlobalAverage Pooling connected end to end in sequence; the meteorological parameter path includes a positional embedding layer PositionalEncoding, two long short-term memory networks LSTM, and a multi-head attention mechanism Multi-Head Attention; wherein, the input feature dimension of the building feature path is (X, 1), X represents the total number of indicators of the building feature; the input feature dimension of the meteorological parameter path is (8760, Y), Y represents the total number of indicators of the meteorological parameter, and 8760 represents 8760 hours in a year; construct a feature splicing layer to splice the output of the building feature path and the output of the meteorological parameter path, and use the spliced ​​features as the input of the multi-task output layer to obtain the initial prediction model; wherein, the output of the multi-task output layer includes building energy consumption intensity, roof photovoltaic power generation, south facade photovoltaic power generation, and east and west facade photovoltaic power generation; Acquire existing building data within a specified longitude and latitude range, and use the existing building data to train the initial prediction model to obtain a multi-task building energy prediction model; In response to a building energy forecast instruction for a target building, obtaining a target building feature of the target building and a target meteorological parameter of a location where the target building is located; Inputting the target building characteristics and the target meteorological parameters into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted roof photovoltaic power generation, predicted south facade photovoltaic power generation, and predicted east and west facade photovoltaic power generation of the target building; A target building performance prediction result of the target building is generated according to the predicted building energy consumption intensity, the predicted rooftop photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation.

2. The building performance prediction method according to claim 1, characterized in that: The using the existing building data to train the initial prediction model to obtain a multi-task building energy prediction model comprises: The existing building data is split according to the configuration ratio to obtain a training set and a validation set; Construct loss function; The initial prediction model is trained by back propagation using the training set and the loss function; wherein the learning rate of the model is adaptively optimized during each round of training using the Adam optimizer; During the training process, when it is detected that the loss function reaches convergence, the training is stopped, and the current model is verified using the verification set; When the current model passes the verification, the current model is determined as the multi-task building energy prediction model.

3. The building performance prediction method according to claim 2, characterized in that: The construction of the loss function includes: The loss function is constructed using a peak constraint module and a mean square error; The peak constraint module uses the PEAK function to increase the penalty for prediction errors between peaks and valleys; the mean square error is used to capture the deviation between the predicted value and the actual value of the model.

4. The building performance prediction method according to claim 2, characterized in that: The using the verification set to verify the current model includes: Inputting the data in the validation set into the current model in sequence for prediction to obtain a prediction result; Calculate the root mean square error coefficient of variation and the coefficient of determination of the current model according to the prediction results; When the root mean square error coefficient of variation is less than a first threshold value, and the determination coefficient is greater than a second threshold value, it is determined that the current model passes the verification; or When the root mean square error coefficient of variation is greater than or equal to the second threshold, and / or the determination coefficient is less than or equal to the second threshold, it is determined that the current model has not passed the verification.

5. The building performance prediction method according to claim 1, characterized in that: The target building performance prediction result of the target building is generated according to the predicted building energy consumption intensity, the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation, including: Determining the estimated energy consumption value of the target building according to the predicted building energy consumption intensity; Determine the estimated energy saving of the target building after installing photovoltaic power generation according to the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation; The target building performance prediction result is generated according to the predicted energy consumption value and the predicted energy saving amount.

6. The building performance prediction method according to claim 1, characterized in that: After generating the target building performance prediction result of the target building according to the predicted building energy consumption intensity, the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation, the method further includes: When the predicted building energy consumption intensity is greater than or equal to a third threshold, sending a renovation prompt for the target building to a designated user; When receiving a remodeled building of the target building based on the remodeling prompt feedback, predicting the remodeled building energy consumption intensity of the remodeled building using the multi-task building energy prediction model; When the energy consumption intensity of the renovated building is less than the predicted building energy consumption intensity, calculating the difference between the predicted building energy consumption intensity and the renovated building energy consumption intensity as the energy consumption optimization value of the renovated building; The energy consumption optimization value is sent to the designated user.

7. A building performance prediction device, characterized in that: The building performance prediction device comprises: A construction unit is used to construct an initial prediction model based on a multi-task learning model with a dual-head input and four-task output architecture, including: constructing an architectural feature path with architectural features as input, and a meteorological parameter path with meteorological parameters as input; wherein the architectural features include at least one indicator and architectural functional features under architectural morphological parameters, and the meteorological parameters include at least one meteorological indicator; wherein the architectural feature path includes two layers of one-dimensional convolutional neural network 1D CNN, batch normalization layer Batch Normalization, random dropout layer Dropout and global average pooling layer Global Average Pooling connected end to end; the meteorological parameter path includes position embedding layers Positional Encoding, two long short-term memory networks LSTM and a multi-head attention mechanism Multi-HeadAttention; wherein, the input feature dimension of the building feature path is (X, 1), X represents the total number of indicators of the building feature; the input feature dimension of the meteorological parameter path is (8760, Y), Y represents the total number of indicators of the meteorological parameter, and 8760 represents 8760 hours in a year; construct a feature splicing layer to splice the output of the building feature path and the output of the meteorological parameter path, and use the spliced ​​features as the input of the multi-task output layer to obtain the initial prediction model; wherein, the output of the multi-task output layer includes building energy consumption intensity, roof photovoltaic power generation, south facade photovoltaic power generation, and east and west facade photovoltaic power generation; A training unit, used to obtain existing building data within a specified longitude and latitude range, and use the existing building data to train the initial prediction model to obtain a multi-task building energy prediction model; an acquisition unit, configured to acquire, in response to a building energy prediction instruction for a target building, a target building feature of the target building and a target meteorological parameter of a location where the target building is located; An input unit, used to input the target building characteristics and the target meteorological parameters into the multi-task building energy prediction model to obtain the predicted building energy consumption intensity, predicted roof photovoltaic power generation, predicted south facade photovoltaic power generation, and predicted east and west facade photovoltaic power generation of the target building; A generating unit is used to generate a target building performance prediction result of the target building according to the predicted building energy consumption intensity, the predicted roof photovoltaic power generation, the predicted south facade photovoltaic power generation and the predicted east and west facade photovoltaic power generation.

8. A computer device, characterized in that: The computer device comprises: a memory storing at least one instruction; and A processor executes instructions stored in the memory to implement the building performance prediction method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one instruction, and the at least one instruction is executed by a processor in a computer device to implement the building performance prediction method according to any one of claims 1 to 6.

Images