Energy management method and system
By constructing the building's energy usage data and environmental parameter change images, combined with the LSTM neural network and fuzzy logic rules, the problem of low prediction accuracy of building energy usage data is solved, high-precision energy prediction and control are achieved, and the level of building energy management is improved. .
Patent Information
- Application Number
- CN202410699544.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-31
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-05-31
AI Technical Summary
The existing technology cannot effectively monitor and predict building energy use data, resulting in low energy use accuracy and inability to effectively grasp building energy consumption levels and tap energy saving potential.
By constructing the changing images of the energy use data, environmental parameters and human behavior information of the target building, extracting features and combining them, using LSTM neural network to establish an energy prediction model, and combining fuzzy logic rules for energy control.
It improves the prediction accuracy of future energy use data, improves the level of building energy control, and provides accurate grasp of building energy consumption and taps energy saving potential.
Smart Images

Figure CN118470430B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field related to energy management, and more specifically, to an energy management and control method and system. Background Art
[0002] It is reported that China's per capita energy possession is only 40% of the world average, and building energy consumption accounts for about 40% of the total social energy consumption, but the energy efficiency is currently only 33%. Therefore, building energy management is crucial to achieving energy conservation and emission reduction goals. At present, the monitoring technology for various building indicators has achieved rapid development, but it has not been effectively applied. The prediction accuracy of future energy usage data is low, and it is impossible to grasp the building energy consumption level, tap the energy-saving potential, and improve the building energy management level. Therefore, it is necessary to design a technical solution that can overcome the above defects to a certain extent. Summary of the invention
[0003] One object of the present invention is to provide an energy management and control method and system, which can accurately predict future energy usage data and improve the level of energy management and control.
[0004] In order to achieve these purposes and other advantages of the present invention, according to one aspect of the present invention, the present invention provides an energy management and control method, including: S1: constructing an energy usage data change image of a target building; S2: constructing an environmental parameter change image of the target building; S3: constructing an incoming pedestrian flow change image, an outgoing pedestrian flow change image, an equipment noise change image and a human noise change image of the target building; S4: extracting features of the energy usage data change image, the environmental parameter change image, the incoming pedestrian flow change image, the outgoing pedestrian flow change image, the equipment noise change image and the human noise change image, and combining them to obtain combined features; S5: inputting the combined features into a pre-established energy prediction model to obtain future energy usage data of the target building so as to perform energy management and control on the target building.
[0005] Furthermore, in S1, the energy usage data includes electricity consumption data, water consumption data and gas consumption data in the target building, and the energy usage data change image is constructed according to the energy usage data.
[0006] Further, in S2, the environmental parameters include an internal temperature value, an external temperature value, an internal humidity value, and an external humidity value of the target building, and the environmental parameter change image is constructed according to the environmental parameters.
[0007] Furthermore, in S3, the incoming crowd flow change image is constructed according to the crowd flow entering the target building, the outgoing crowd flow change image is constructed according to the crowd flow leaving the target building, the equipment noise change image is constructed according to the identified equipment noise and noise value, and the human noise change image is constructed according to the identified human noise and noise value.
[0008] Furthermore, in S4, the energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images and the human noise change images of multiple time periods during the historical period are obtained, features are extracted and combined to obtain combined features, a training set is constructed, and an LSTM neural network is input to obtain an energy prediction model through training; wherein, during the training process, combination coefficients are randomly assigned to the features extracted from the energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images and the human noise change images, and the optimal combination coefficient corresponding to the energy prediction model is selected to predict the future energy usage data.
[0009] Furthermore, the energy usage data change image, the environmental parameter change image, the incoming human flow change image, the outgoing human flow change image, the equipment noise change image and the human noise change image are bar graphs or radar graphs.
[0010] Furthermore, in S5, fuzzy logic rules are established based on energy management and control knowledge, and the future energy usage data is compared with the fuzzy logic rules to perform energy management and control on the target building.
[0011] According to another aspect of the present invention, an energy management and control system is also provided, including: an energy data construction module, used to construct an energy usage data change image of a target building, the energy usage data including electricity consumption data, water consumption data and gas consumption data; an environmental parameter construction module, used to construct an environmental parameter change image of the target building, the environmental parameters including internal temperature value, external temperature value, internal humidity value and external humidity value; a human behavior information construction module, used to construct an incoming pedestrian flow change image, an outgoing pedestrian flow change image, an equipment noise change image and a human noise change image of the target building; a combination module, used to extract features of the energy usage data change image, the environmental parameter change image, the incoming pedestrian flow change image, the outgoing pedestrian flow change image, the equipment noise change image and the human noise change image, and combine them to obtain combined features; a management and control module, used to input the combined features into a pre-established energy prediction model to obtain future energy usage data of the target building so as to perform energy management and control on the target building.
[0012] Furthermore, the energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images and the human noise change images of multiple time periods during the historical period are obtained, features are extracted and combined to obtain combined features, a training set is constructed, and the LSTM neural network is input to obtain the energy prediction model through training; wherein, during the training process, combination coefficients are randomly assigned to the features extracted from the energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images and the human noise change images, and the optimal combination coefficient corresponding to the energy prediction model is selected to predict the future energy usage data.
[0013] Furthermore, the management and control module inputs the combined features into the pre-established energy prediction model to obtain the future energy usage data of the target building to perform energy management and control on the target building.
[0014] The present invention has at least the following beneficial effects:
[0015] The present invention obtains the energy usage data, environmental parameters, and human behavior information (human flow, noise) of the target building, establishes an energy prediction model, and uses the energy prediction model to predict the future energy usage data of the target building to perform energy management and control on the target building; the present invention not only takes into account the energy usage data, but also takes into account the environmental parameters and human behavior information. The established energy prediction model has a high prediction accuracy for future energy usage data, providing a reference for building energy management and control.
[0016] Other advantages, objectives and features of the present invention will be embodied in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 This is a flow chart of an embodiment of the present application. DETAILED DESCRIPTION
[0018] The present invention is further described in detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.
[0019] It should be understood that the terms such as “having”, “including” and “comprising” used herein do not exclude the existence or addition of one or more other elements or combinations thereof.
[0020] like Figure 1 As shown, the embodiment of the present application provides an energy management and control method, including:
[0021] S1: constructing an energy usage data change image of a target building; optionally, the energy usage data is derived from various monitoring instruments in the target building, such as an electric meter, a water meter and a gas meter; preferably, the energy usage data includes electricity consumption data, water consumption data and gas consumption data in the target building, and the energy usage data change image is constructed based on the energy usage data; optionally, the energy usage data change image is a time-varying graph, with a time interval of 3 hours, 6 hours, 12 hours or 24 hours;
[0022] S2: Construct an environmental parameter change image of the target building; optionally, the environmental data includes wind force, weather, air pressure, temperature, humidity, etc.; preferably, the environmental data includes the internal temperature value, external temperature value, internal humidity value, and external humidity value of the target building, and the environmental parameter change image is constructed according to the environmental parameters; optionally, the environmental parameter change image is a time-varying graph, with a time interval of 6 hours, 12 hours, or 24 hours; optionally, multiple temperature sensors and humidity sensors are set inside and outside the target building, and the internal temperature value, external temperature value, internal humidity value, and external humidity value are calculated using the average value or weighted average value of each internal and external temperature sensor;
[0023] S3: construct the target building's incoming traffic flow change image, outgoing traffic flow change image, equipment noise change image and human noise change image; S3 reflects the human behavior of the target building, and the human behavior information includes the incoming traffic flow, outgoing traffic flow and various noise values of the target building. The incoming traffic flow and outgoing traffic flow can give the quantitative characteristics of humans in the target building, and the noise value can reflect the human activities inside the target building and the intensity of human use of equipment; the incoming traffic flow change image is constructed according to the traffic flow entering the target building, the outgoing traffic flow change image is constructed according to the traffic flow leaving the target building, the equipment noise change image is constructed according to the identified equipment noise and noise value, and the human noise change image is constructed according to the identified human noise and noise value; optionally, various noises in the target building are collected in advance, and features are extracted as a judgment Standard, multiple sound sensors are set in the target building to collect the noise in the target building, and the judgment criteria are used to distinguish and save the noise values separately; optionally, the flow change image, the outgoing flow change image, the equipment noise change image and the human noise change image are all time change images, and the time interval is 6 hours, 12 hours or 24 hours; optionally, cameras can be set at the entrances and exits of the target building to count the incoming and outgoing flow; optionally, the incoming and outgoing flow are read from the access control system of the target building; optionally, multiple sound sensors can be set in the target building to collect and identify the sound signals in the target building, identify various types of sound signals, equipment noise such as elevator noise signals, ventilation equipment noise signals, refrigeration and heating equipment noise signals, stove noise signals, etc., human noise includes talking noise, etc., and record the noise values separately;
[0024] S4: extracting features of the energy usage data change image, the environmental parameter change image, the incoming human flow change image, the outgoing human flow change image, the equipment noise change image, and the human noise change image, and combining them to obtain combined features; optionally, using a CNN neural network for feature extraction; optionally, extracting feature values of each image; optionally, the combination may be a linear combination or a nonlinear combination;
[0025] S5: Input the combined features into a pre-established energy prediction model to obtain the future energy usage data of the target building so as to perform energy management and control on the target building; optionally, since the energy usage data is time-sensitive, a long short-term memory network (LSTM) is used to extract time series features, establish an energy prediction model, and improve the prediction accuracy of the energy prediction model; further, input the features contained in the current energy usage data, environmental parameters, and human behavior information into the energy prediction model to obtain energy usage data for a period of time in the future, and manage the energy usage data of the target building based on the future energy usage data; optionally, compare the future energy usage data with a predetermined threshold, and adjust the energy usage equipment of the target building based on the comparison result; this embodiment not only takes into account the energy usage data, but also takes into account the environmental data and human behavior information, thus overcoming the defect of only considering the energy usage data. The established energy prediction model has a high prediction accuracy for the future energy usage data, thus providing a reference for mastering the building energy consumption level, tapping the energy-saving potential, and improving the building energy management and control level.
[0026] Preferably, in S4, the energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images and the human noise change images of multiple time periods during the historical period are obtained, features are extracted and combined to obtain combined features, a training set is constructed, and an LSTM neural network is input to obtain an energy prediction model through training; wherein, during the training process, combination coefficients are randomly assigned to the features extracted from the energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images and the human noise change images, and the combination coefficient corresponding to the optimal energy prediction model is selected to predict the future energy usage data;
[0027] That is to say, the historical period is divided into multiple time periods, each of which is 5 to 10 days. The energy usage data, environmental parameters, incoming traffic, outgoing traffic, equipment noise value, and human noise value are counted every 6 hours, 12 hours, or 24 hours. The traffic flow change image, outgoing traffic flow change image, equipment noise change image, and human noise change image are drawn with the data values and time points as coordinates.
[0028] Optionally, the energy usage data change image, the environmental parameter change image, the incoming human flow change image, the outgoing human flow change image, the equipment noise change image and the human noise change image are bar graphs or radar graphs, preferably radar graphs, and the combined features obtained by extracting features can reflect the changes in energy usage data, environmental parameters and human behavior information to a greater extent;
[0029] CNN is a neural network model for image processing that can automatically learn features in images, while LSTM is a neural network model for time series data processing that can remember previous information and use it in subsequent processing;
[0030] This embodiment first uses CNN to fully extract image features, fully utilizes the advantages of CNN in image processing, especially extracts small changes in environmental parameters and human behavior information, and then uses the combined features of a time period as input, inputs the LSTM neural network, and performs training to obtain an energy prediction model. The energy prediction model includes 50-200 LSTM units, and the last layer is a fully connected layer, which is used to output predicted future energy usage data, including electricity consumption data, water consumption data, and gas consumption data, which are used as the basis for energy management and control; during the training process, different coefficients are randomly assigned to each feature, and a variety of combined features are obtained by combining them as inputs, and the prediction accuracy of the various coefficients is compared, and the combination coefficient of the energy prediction model with the highest prediction accuracy is selected. When predicting future energy usage data, the optimal combination coefficient is used to combine the current features; that is, this embodiment is not only the training of LSTM neural network parameters, but also the training of combination coefficients, so as to fully explore the influence of energy usage data, environmental parameters, incoming traffic, outgoing traffic, equipment noise value, and human noise value on prediction accuracy, and improve prediction accuracy;
[0031] After testing, the prediction accuracy of the model established using only energy usage data is 85.6%, the prediction accuracy of the model established using only energy usage data change images is 86.7%, the prediction accuracy of the model established using only energy usage data change images and environmental parameter change images is 89.2%, and the prediction accuracy of the energy prediction model established in this embodiment can reach 94.2%.
[0032] In another embodiment, in S5, fuzzy logic rules are established based on energy management knowledge, and the future energy usage data is compared with the fuzzy logic rules to perform energy management on the target building; optionally, fuzzy logic rules are established using expert knowledge, energy management specifications and other knowledge; optionally, a predetermined threshold is determined by an expert based on the energy usage data of the target building, and the fuzzy rule may be that the expert constructs a management strategy based on multiple different ratios of the energy usage data to the predetermined threshold; after obtaining the future energy usage data using the energy prediction model, the ratio to the predetermined threshold is calculated, and then the fuzzy rules are used to obtain the corresponding management strategy to perform energy management on the target building.
[0033] The embodiment of the present application also provides an energy management and control system, including:
[0034] An energy data construction module is used to construct an energy usage data change image of a target building, wherein the energy usage data includes electricity consumption data, water consumption data and gas consumption data; optionally, the energy usage data is derived from various monitoring instruments in the target building, such as an electricity meter, a water meter and a gas meter; preferably, the energy usage data includes electricity consumption data, water consumption data and gas consumption data in the target building, and the energy usage data change image is constructed based on the energy usage data; optionally, the energy usage data change image is a change graph over time, with a time interval of 3 hours, 6 hours, 12 hours or 24 hours;
[0035] An environmental parameter construction module is used to construct an environmental parameter change image of the target building, wherein the environmental parameters include an internal temperature value, an external temperature value, an internal humidity value, and an external humidity value; optionally, the environmental data include wind force, weather, air pressure, air temperature, humidity, etc.; preferably, the environmental data includes the internal temperature value, external temperature value, internal humidity value, and external humidity value of the target building, and the environmental parameter change image is constructed according to the environmental parameters; optionally, the environmental parameter change image is a change graph over time, with a time interval of 6 hours, 12 hours, or 24 hours;
[0036] A human behavior information construction module is used to construct an image of the change in the flow of people entering the target building, an image of the change in the flow of people going out, an image of the change in equipment noise, and an image of the change in human noise; the behavior information includes the flow of people entering the target building, the flow of people going out, and various noise values; the flow of people entering and the flow of people going out can give the quantitative characteristics of humans in the target building, and the noise value can reflect the intensity of human activities and human use of equipment inside the target building; the image of the change in the flow of people entering the target building is constructed according to the flow of people going out of the target building, the image of the change in equipment noise is constructed according to the identified equipment noise and noise value, and the image of the change in human noise is constructed according to the identified human noise and noise value; optionally, the pre- First, various noises in the target building are collected, and features are extracted as judgment criteria. Multiple sound sensors are set in the target building to collect the noise in the target building, and the noise values are distinguished by using the judgment criteria, and the noise values are saved separately; optionally, the flow change image, the outgoing flow change image, the equipment noise change image and the human noise change image are all time-varying graphs, with a time interval of 6 hours, 12 hours or 24 hours; optionally, cameras can be set at the entrances and exits of the target building to count the incoming and outgoing flow; optionally, the incoming and outgoing flow are read from the access control system of the target building; optionally, multiple sound sensors can be set in the target building to collect and identify sound signals in the target building, identify various types of sound signals, and record the noise values separately;
[0037] A combination module is used to extract features of the energy usage data change image, the environmental parameter change image, the incoming human flow change image, the outgoing human flow change image, the equipment noise change image and the human noise change image, and combine them to obtain a combined feature; optionally, a CNN neural network is used for feature extraction; optionally, a feature value of each image is extracted; optionally, the combination may be a linear combination or a nonlinear combination;
[0038] A management and control module, used for inputting the combined features into a pre-established energy prediction model to obtain future energy usage data of the target building, so as to perform energy management and control on the target building;
[0039] This embodiment inputs the features contained in the current energy usage data, environmental parameters and human behavior information into the energy prediction model to obtain energy usage data for a period of time in the future, and manages the energy usage data of the target building based on the future energy usage data; optionally, the future energy usage data is compared with a predetermined threshold, and the energy usage equipment of the target building is adjusted according to the comparison result; this embodiment not only considers the energy usage data, but also considers the environmental data and human behavior information, overcomes the defects of only considering the energy usage data, and the established energy prediction model has a high prediction accuracy for the future energy usage data, which provides a reference for mastering the building energy consumption level, exploring the energy-saving potential, and improving the building energy management and control level.
[0040] In another embodiment, the energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images and the human noise change images of multiple time periods in the historical period are obtained, features are extracted and combined to obtain combined features, a training set is constructed, and an LSTM neural network is input to obtain the energy prediction model through training; wherein, during the training process, combination coefficients are randomly assigned to the features extracted from the energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images and the human noise change images, and the combination coefficient corresponding to the optimal energy prediction model is selected to predict the future energy usage data; this embodiment first uses CNN to fully extract image features, fully utilizes the advantages of CNN in image processing, especially extracts small changes in environmental parameters and human behavior information, and then inputs the combined features into LSTM for training to obtain an energy prediction model, the last layer of the energy prediction model is a fully connected layer, which is used to output the predicted future energy usage data, and the prediction accuracy of the obtained energy prediction model can reach 94.2%.
[0041] In another embodiment, the management and control module inputs the combined features into the pre-established energy prediction model to obtain the future energy usage data of the target building to perform energy management on the target building; optionally, fuzzy logic rules are established using expert knowledge, energy management specifications and other knowledge; optionally, a predetermined threshold is determined by an expert based on the energy usage data of the target building, and the fuzzy rule may be that the expert constructs a management and control strategy based on multiple different ratios of the energy usage data to the predetermined threshold; after obtaining the future energy usage data using the energy prediction model, the ratio to the predetermined threshold is calculated, and then the fuzzy rules are used to obtain the corresponding management and control strategy to perform energy management on the target building.
[0042] The number of devices and processing scales described here are used to simplify the description of the present invention. Applications, modifications and variations of the energy management method and system of the present invention are obvious to those skilled in the art.
[0043] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and the implementation modes, and they can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.
Claims
1. An energy management and control method, characterized in that: include: S1: Construct an image of the energy usage data changes of the target building; S2: constructing an environmental parameter change image of the target building; S3: constructing an image of the change in the flow of people entering the target building, an image of the change in the flow of people leaving the target building, an image of the change in the noise of equipment, and an image of the change in the noise of humans; S4: extracting features of the energy usage data change image, the environmental parameter change image, the incoming human flow change image, the outgoing human flow change image, the equipment noise change image, and the human noise change image, and combining them to obtain a combined feature; S5: inputting the combined features into a pre-established energy prediction model to obtain future energy usage data of the target building, so as to perform energy management and control on the target building; In S1, the energy usage data includes electricity consumption data, water consumption data and gas consumption data in the target building, and the energy usage data change image is constructed according to the energy usage data; In S2, the environmental parameters include an internal temperature value, an external temperature value, an internal humidity value, and an external humidity value of the target building, and the environmental parameter change image is constructed according to the environmental parameters; In S3, the incoming traffic flow change image is constructed according to the traffic flow entering the target building, the outgoing traffic flow change image is constructed according to the traffic flow leaving the target building, the equipment noise change image is constructed according to the identified equipment noise and noise value, and the human noise change image is constructed according to the identified human noise and noise value; In S4, the energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images, and the human noise change images in multiple time periods during the historical period are obtained, features are extracted and combined to obtain combined features, a training set is constructed, and the LSTM neural network is input to obtain an energy prediction model through training; During the training process, combination coefficients are randomly assigned to features extracted from the energy usage data change image, the environmental parameter change image, the incoming human flow change image, the outgoing human flow change image, the equipment noise change image and the human noise change image, and the optimal combination coefficient corresponding to the energy prediction model is selected to predict the future energy usage data.
2. The energy management and control method according to claim 1, characterized in that: The energy usage data change image, the environmental parameter change image, the incoming human flow change image, the outgoing human flow change image, the equipment noise change image and the human noise change image are bar graphs or radar graphs.
3. The energy management and control method according to claim 1, characterized in that: In S5, fuzzy logic rules are established based on energy management and control knowledge, and the future energy usage data is compared with the fuzzy logic rules to perform energy management and control on the target building.
4. Energy management and control system, characterized in that: include: An energy data construction module, used to construct an image of energy usage data changes of a target building, wherein the energy usage data includes electricity consumption data, water consumption data and gas consumption data; An environmental parameter construction module, used to construct an environmental parameter change image of the target building, wherein the environmental parameters include an internal temperature value, an external temperature value, an internal humidity value, and an external humidity value; A human behavior information construction module is used to construct an image of the change in the flow of people entering the target building, an image of the change in the flow of people leaving the target building, an image of the change in the noise of equipment, and an image of the change in the noise of humans; A combination module is used to extract features of the energy usage data change image, the environmental parameter change image, the incoming human flow change image, the outgoing human flow change image, the equipment noise change image and the human noise change image, and combine them to obtain a combined feature; A management and control module, used for inputting the combined features into a pre-established energy prediction model to obtain future energy usage data of the target building, so as to perform energy management and control on the target building; The energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images and the human noise change images of multiple time periods during the historical period are obtained, features are extracted and combined to obtain combined features, a training set is constructed, and the LSTM neural network is input to obtain the energy prediction model through training; wherein, during the training process, combination coefficients are randomly assigned to the features extracted from the energy usage data change images, the environmental parameter change images, the incoming human flow change images, the outgoing human flow change images, the equipment noise change images and the human noise change images, and the optimal combination coefficient corresponding to the energy prediction model is selected to predict the future energy usage data.
