Building energy management method and device based on embedded industrial computer

Abstract

The invention discloses a building energy management method and device based on an embedded industrial computer, and relates to the technical field of energy management. Acquiring power consumption of each target sub-region to obtain first power consumption data as an actual energy consumption coefficient; periodically acquiring the video stream of each target sub-region, and calculating to obtain an activeness coefficient; obtaining second electricity consumption and inputting the second electricity consumption into the electricity consumption prediction model to obtain a predicted energy consumption coefficient; obtaining environmental parameters of each target sub-region, and calculating an energy consumption threshold according to the predicted energy consumption coefficient, the activity coefficient and the environmental parameters; if the actual energy consumption coefficient is greater than an energy consumption threshold, determining an energy management scheme for each target sub-region in the target region; according to the method, the predicted energy consumption coefficient, the activity coefficient and the environmental parameters are comprehensively considered, the energy consumption threshold value can be dynamically set, and the targeted energy management scheme is formulated in time when the actual energy consumption exceeds the threshold value, so that the scheme formulation is more efficient and more accurate.

Description

Technical Field

[0001] The present invention belongs to the technical field, and in particular relates to a building energy management method and device based on an embedded industrial computer. Background Art

[0002] Faced with the tense situation of traditional energy supply and the negative impact on the environment caused by its mining and use, building energy and various parameters are monitored through embedded industrial computers to achieve precise energy management.

[0003] Patent No. CN118368791A discloses a building energy management method and system based on video surveillance; a host control system, a video surveillance acquisition unit and a lighting system, the host control system is arranged in the building control room; the host control system is equipped with a pre-trained lighting information preprocessing unit and a lighting control strategy generation unit; the video surveillance acquisition unit is arranged in the building, and its field of view covers at least the corridor and part of the working area; the lighting system includes a lighting unit, a power switch connected in series with the lighting unit, a lower control unit connected to the power switch, the lower control unit is connected to the host control system, and the lower control unit is connected to the memory. The present invention obtains the illuminance and the personnel position distribution and movement path information through the video surveillance acquisition unit, changes the working state of the lighting system, and ensures that the lighting resources are allocated on demand.

[0004] Although the above-mentioned technologies have solved some problems, some problems still exist. For example, the above-mentioned technologies monitor too few parameters, which makes the judgment in the energy management process inaccurate, resulting in poor results of the management plans formulated subsequently. Summary of the invention

[0005] The purpose of the present invention is to solve the problem that there are too few monitored parameters, which makes the judgment in the energy management process inaccurate and leads to poor effect of the management plan formulated subsequently, and proposes a building energy management method and device based on an embedded industrial computer.

[0006] In a first aspect of the present invention, a building energy management method based on an embedded industrial computer is first proposed, the method comprising: Acquire the power consumption of each target sub-area in the target area to obtain a first power consumption data set, and use the first power consumption as the actual energy consumption coefficient; the first power consumption data set includes the first power consumption data of each target sub-area; Periodically obtain the video stream of each target sub-area, identify the flow of people in the video stream and calculate the activity coefficient; Obtaining a second power consumption, and inputting the second power consumption into a power consumption prediction model to obtain a predicted energy consumption coefficient; the second power consumption is the power consumption of a period before the first power consumption data; Acquire environmental parameters of each target sub-area, and calculate an energy consumption threshold value according to the predicted energy consumption coefficient, activity coefficient and environmental parameters; the environmental parameters are the absolute temperature difference, absolute humidity difference and air quality in the target sub-area; If the actual energy consumption coefficient is greater than the energy consumption threshold, an energy management plan is determined for each target sub-area in the target area.

[0007] Optionally, identifying the flow of people in the video stream and calculating the activity coefficient includes: Determine the working state of the embedded industrial computer, and determine the frame extraction frequency of the video stream according to the working state; Extracting frames from the video stream at the frame extraction frequency to obtain an image set, performing background recognition on images in the image set, and removing background features in the images to obtain foreground features; Performing two-dimensional contour recognition on the foreground features and determining key nodes, and constructing a human frame diagram according to the key nodes; the key nodes are determined by the activity areas of human joints; Extracting a gait cycle from the image set according to the human frame diagram, and performing gait recognition according to the gait cycle to obtain gait features; the gait cycle is the time from when the heel of one side touches the ground to when the heel of the same side touches the ground again during the movement of the human body; Pixel features in the foreground features are extracted, and the flow of people in the target sub-area is determined according to the gait features and the pixel features, and the activity coefficient is calculated.

[0008] Optionally, inputting the second power consumption into a power consumption prediction model to obtain a predicted energy consumption coefficient includes: Obtain historical power consumption data and historical temperature data in the target area, and input the historical power consumption data and the historical temperature data into a long short-term memory network model for training; the historical power consumption data and the historical temperature data correspond one to one in terms of timestamps, and the historical temperature data is the temperature of a non-target area; An error value is calculated based on the output of the long short-term memory network model and the true value, the error value is converted into a model parameter through a back propagation algorithm, and the weight and bias of the long short-term memory network model are updated according to the model parameters to obtain a power consumption prediction model; the output of the power consumption prediction model is the power consumption data corresponding to the temperature of the non-target area at the target timestamp; The second power consumption is input into the power consumption prediction model to obtain a predicted energy consumption coefficient.

[0009] Optionally, the energy consumption threshold value calculated according to the predicted energy consumption coefficient, the activity coefficient and the environmental parameters includes: Energy consumption threshold calculation formula Calculate the energy consumption threshold; Among them, F is the energy consumption threshold, YC is the predicted energy consumption coefficient, AC is the activity coefficient, environmental parameters include: absolute temperature difference, absolute humidity difference and air quality, T is the temperature in the target sub-area, is the preset standard temperature, H is the humidity in the target sub-area, To preset standard humidity, is the air quality score in the target area, AQ is the carbon dioxide concentration in the target area, α and β are constant proportional coefficients and both are not zero.

[0010] Optionally, determining an energy management solution for each target sub-area in the target area includes: If the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold, the operating power and operating mode of the power equipment are changed; If the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold, the working power of the power equipment is changed and the working mode of the power equipment is maintained; If the activity coefficient in the target sub-area is not greater than the activity threshold, and the environmental parameter is not greater than the environmental parameter threshold, the working power and working mode of the power equipment are maintained; If the activity coefficient in the target sub-area is not greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold, the working mode of the power equipment is changed and the working power of the power equipment is maintained.

[0011] In a second aspect of the present invention, a building energy management device based on an embedded industrial computer is provided, comprising: a first power consumption acquisition module, a video stream acquisition module, a second power consumption acquisition module, a threshold determination module and a management scheme determination module: The first power consumption acquisition module is used to acquire the power consumption of each target sub-area in the target area to obtain a first power consumption data set, and use the first power consumption as the actual energy consumption coefficient; the first power consumption data set includes the first power consumption data of each target sub-area; The video stream acquisition module is used to periodically acquire the video stream of each target sub-area, identify the flow of people in the video stream and calculate the activity coefficient; The second power consumption acquisition module is used to acquire the second power consumption, and input the second power consumption into the power consumption prediction model to obtain the predicted energy consumption coefficient; the second power consumption is the power consumption of the previous cycle of the first power consumption data; The threshold determination module is used to obtain the environmental parameters of each target sub-area, and calculate the energy consumption threshold according to the predicted energy consumption coefficient, the activity coefficient and the environmental parameters; the environmental parameters are the absolute temperature difference, the absolute humidity difference and the air quality in the target sub-area; The management scheme determination module is used to determine an energy management scheme for each target sub-area in the target area if the actual energy consumption coefficient is greater than the energy consumption threshold.

[0012] Optionally, the video stream acquisition module includes: a frame extraction frequency determination module, a foreground feature extraction module, a human frame diagram construction module, a gait recognition module and an activity coefficient calculation module: The frame extraction frequency determination module is used to determine the working state of the embedded industrial computer, and determine the frame extraction frequency of the video stream according to the working state; The foreground feature extraction module is used to extract frames from the video stream according to the frame extraction frequency to obtain an image set, perform background recognition on the images in the image set, and remove background features in the image to obtain foreground features; The human frame diagram construction module is used to perform two-dimensional contour recognition on the foreground features and determine key nodes, and construct a human frame diagram according to the key nodes; the key nodes are determined by the activity areas of human joints; The gait recognition module is used to extract the gait cycle in the image set according to the human frame diagram, and perform gait recognition according to the gait cycle to obtain gait features; the gait cycle is the time from when the heel of one side touches the ground to when the heel of the same side touches the ground again during the movement of the human body; The activity coefficient calculation module is used to extract pixel features from foreground features, determine the flow of people in the target sub-area according to the gait features and the pixel features, and calculate the activity coefficient.

[0013] Optionally, the second power consumption acquisition module includes: a training data input module, a model parameter updating module and a predicted energy consumption coefficient acquisition module: The training data input module is used to obtain historical power consumption data and historical temperature data in the target area, and input the historical power consumption data and the historical temperature data into the long short-term memory network model for training; the historical power consumption data and the historical temperature data correspond one to one in terms of timestamps, and the historical temperature data is the temperature of the non-target area; The model parameter updating module is used to calculate an error value according to the output of the long short-term memory network model and the true value, convert the error value to obtain model parameters through a back propagation algorithm, and update the weights and biases of the long short-term memory network model according to the model parameters to obtain a power consumption prediction model; the output of the power consumption prediction model is the power consumption data corresponding to the non-target area temperature at the target timestamp; The predicted energy consumption coefficient acquisition module is used to input the second power consumption into the power consumption prediction model to obtain the predicted energy consumption coefficient.

[0014] Optionally, the threshold determination module is further configured to calculate the energy consumption threshold through the energy consumption threshold calculation formula Calculate the energy consumption threshold; Among them, F is the energy consumption threshold, YC is the predicted energy consumption coefficient, AC is the activity coefficient, environmental parameters include: absolute temperature difference, absolute humidity difference and air quality, T is the temperature in the target sub-area, is the preset standard temperature, H is the humidity in the target sub-area, To preset standard humidity, is the air quality score in the target area, AQ is the carbon dioxide concentration in the target area, α and β are constant proportional coefficients and both are not zero.

[0015] Optionally, the management solution determination module includes: a first management solution module, a second management solution module, a third management solution module and a fourth management solution module: The first management solution module is used to change the working power and working mode of the power equipment if the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold; The second management solution module is used to change the working power of the power equipment and maintain the working mode of the power equipment if the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold; The third management solution module is used to maintain the working power and working mode of the power equipment if the activity coefficient in the target sub-area is not greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold; The fourth management scheme module is used to change the working mode of the power equipment and maintain the working power of the power equipment if the activity coefficient in the target sub-area is not greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold.

[0016] Beneficial effects of the present invention: The present invention proposes a building energy management method based on an embedded industrial computer, which obtains a first power consumption as an actual energy consumption coefficient by obtaining the power consumption of each target sub-area in the target area; periodically obtains the video stream of each target sub-area, identifies the flow of people in the video stream and calculates the activity coefficient; obtains a second power consumption, and inputs the second power consumption into a power consumption prediction model to obtain a predicted energy consumption coefficient; obtains environmental parameters of each target sub-area, and calculates an energy consumption threshold according to the predicted energy consumption coefficient, the activity coefficient and the environmental parameters; if the actual energy consumption coefficient is greater than the energy consumption threshold, an energy management plan is determined for each target sub-area in the target area; comprehensively considering the predicted energy consumption coefficient, the activity coefficient and the environmental parameters, the energy consumption threshold can be dynamically set, and when the actual energy consumption exceeds the threshold, a targeted energy management plan can be formulated in time, so that the plan formulation is more efficient and more accurate, and the comfort of personnel is guaranteed while reducing energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below in conjunction with the accompanying drawings.

[0018] Figure 1 A flowchart of a building energy management method based on an embedded industrial computer is provided for an embodiment of the present invention; Figure 2 A structural schematic diagram of a building energy management device based on an embedded industrial computer is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the present invention, the description of "first", "second", etc. is only used for descriptive purposes, and cannot be understood as indicating or implying its relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" can explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0020] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of the present invention.

[0021] The embodiment of the present invention provides a building energy management method based on an embedded industrial computer. Figure 1 , Figure 1 A flow chart of a building energy management method based on an embedded industrial computer provided by an embodiment of the present invention. The method comprises the following steps: S101, obtaining the power consumption of each target sub-area in the target area to obtain a first power consumption data set, and using the first power consumption as an actual energy consumption coefficient; S102, periodically obtaining the video stream of each target sub-area, identifying the flow of people in the video stream and calculating the activity coefficient; S103, obtaining a second power consumption, and inputting the second power consumption into a power consumption prediction model to obtain a predicted energy consumption coefficient; S104, obtaining environmental parameters of each target sub-area, and calculating an energy consumption threshold according to the predicted energy consumption coefficient, activity coefficient and environmental parameters; S105, if the actual energy consumption coefficient is greater than the energy consumption threshold, determining an energy management plan for each target sub-area in the target area; The first power consumption data set includes the first power consumption data of each target sub-area; the second power consumption is the power consumption of the previous cycle of the first power consumption data; the environmental parameters are the absolute temperature difference, absolute humidity difference and air quality in the target sub-area; A building energy management method based on an embedded industrial computer provided in an embodiment of the present invention can dynamically set an energy consumption threshold by comprehensively considering the predicted energy consumption coefficient, activity coefficient and environmental parameters, and formulate targeted energy management plans in time when the actual energy consumption exceeds the threshold, so that the plan formulation is more efficient and more accurate, reducing energy consumption while ensuring the comfort of personnel.

[0022] In one implementation, the target area is a building, such as an office building; the target sub-area is any floor in the target area, such as any floor in the office building; the power consumption of the power equipment in the target area is obtained through a power sensor; the video stream is obtained through a monitoring device installed in the target sub-area, and the monitoring device monitors the flow of people in the public area; the actual energy consumption coefficient and the predicted energy consumption coefficient are both power consumption, for example, the power consumption is 500KW·h, in kilowatt-hours.

[0023] In one implementation, the environmental parameters of each target sub-area, including temperature, humidity and air quality, are measured by actual IoT devices. The infrared temperature sensor is responsible for measuring temperature in °C, for example: MLX90614; the humidity sensor is responsible for measuring indoor humidity in %, for example: BS4-BR-2000HTB; the air quality sensor is responsible for measuring indoor carbon dioxide content in %, for example: DAQ-GP-CO2 4G; the absolute temperature difference is the difference between the actual measured temperature and the preset standard temperature, and the difference is taken as an absolute value, wherein the preset standard temperature is for example: 26 °C; the absolute humidity difference is the difference between the actual measured humidity and the preset standard humidity, and the difference is taken as an absolute value, wherein the preset standard humidity is for example: 50%; the energy consumption threshold is calculated based on the predicted energy consumption coefficient, the activity coefficient and the environmental parameters, and the dynamically changing energy consumption threshold can be adjusted in real time according to the environmental parameters, the predicted energy consumption coefficient and the activity coefficient of each target sub-area, so as to more accurately reflect the actual energy consumption, optimize energy utilization efficiency and reduce unnecessary energy waste.

[0024] In one implementation, for each target sub-area (each floor) in the target area (office building), a specific energy management plan is determined based on the comparison result between the actual energy consumption coefficient and the dynamic energy consumption threshold, which can significantly improve the pertinence and efficiency of energy use, reduce energy waste, and improve the operation and management level of the office building. The energy management plan is a management plan formulated for indoor power equipment, such as lighting systems, air conditioners, etc. The energy management plan is for example: controlling the working power and working mode of the power equipment; the working power is for example: the wind intensity of the air conditioner in the 18°C ​​cooling mode, the wind intensity includes: weak, medium and strong, the wind intensity is for example: weak corresponds to 500-800 rpm, medium corresponds to 800-1200 rpm, strong corresponds to 1200-1800 rpm, change the temperature of the air conditioner, and increase the temperature of the 18°C ​​cooling mode to 20°C; the working mode is for example: switching of the air conditioner's cooling, dehumidification, heating and fresh air modes.

[0025] In one embodiment, step S102 includes: Determine the working status of the embedded industrial computer, and determine the frame extraction frequency of the video stream according to the working status; The video stream is extracted at a frame extraction frequency to obtain an image set, background recognition is performed on the images in the image set, and the background features in the image are removed to obtain the foreground features; Perform 2D contour recognition on foreground features and determine key nodes, and build a human frame diagram based on the key nodes; the key nodes are determined by the activity areas of human joints; The gait cycle in the image set is extracted according to the human frame diagram, and gait recognition is performed according to the gait cycle to obtain gait features; the gait cycle is the time from when the heel of one side touches the ground to when the heel of the same side touches the ground again during the movement of the human body; Pixel features in the foreground features are extracted, and the flow of people in the target sub-area is determined based on the gait features and pixel features, and the activity coefficient is calculated.

[0026] In one implementation, an embedded industrial computer, for example, has: CPU ARM926EJ, main frequency 300MHZ; high-performance 64MB DDR2 memory, onboard 128MB SLC Flash; integrated high-reliability industrial-grade Cat1 4G wireless module; onboard 1 fully industrial-isolated RS485 / RS232 interface; 100M / 10M adaptive network card, supporting AUTO MDI / MDIX; integrated dedicated encryption chip; the frame extraction frequency of the video stream is determined according to the working status of the embedded industrial computer. The main benefit of this step is that it can optimize resource utilization and ensure the real-time and accuracy of video processing. By dynamically adjusting the frame extraction frequency, it is possible to reduce unnecessary computing burden and storage space occupation while ensuring the integrity of video information, thereby improving the overall performance and efficiency of the system.

[0027] In one implementation, the gait cycle refers to the time from when one side's heel touches the ground to when the heel touches the ground again during walking or running; the gait cycle is an important parameter that describes the gait characteristics of the human body when walking. It reflects the continuity and periodicity of human walking. Each person's gait is different, which directly reflects personal habits. The gait cycle is also different. The walking posture and the time to complete a gait cycle can be used as an important basis for character recognition.

[0028] In one implementation, background recognition is performed on images in the image set, and background features in the image are removed to obtain foreground features; that is, a basic background image of the target sub-area is obtained, and the basic background image contains no people and only static objects. The background image is used as a background feature, and background features are removed from the image, that is, the background is deleted to obtain the outline of the person. The outline of the person is two-dimensional in the picture at this time, and the outline of the person is also the foreground feature.

[0029] In one implementation, two-dimensional contour recognition is performed on foreground features and key nodes are determined, and a human frame diagram is constructed based on the key nodes; the two-dimensional contour is recognized, and the head and limbs of the human body are recognized first, which can reduce the time for locating the human body, because the head and limbs of the human body have obvious characteristics; the key nodes are determined by the activity areas of the human joints, that is, the joints of the human body are regarded as key nodes, for example: the knee joint at the connection of the thigh and the calf is a key node, and the elbow joint at the connection of the upper arm and the forearm is a key node.

[0030] In one implementation, gait recognition is performed by using a posture estimation algorithm on a continuous human frame diagram within a gait cycle. The posture estimation algorithm, for example, YOLO-pose, extracts key points from the human frame diagram. The key points include the positions of joints, and the connection of the key points defines the human skeleton. Among them, the color of human clothing in the foreground feature is identified and used as a pixel feature, that is, the pixel feature and the gait feature are jointly identified to enhance the accuracy of recognition. The pixel feature and the gait feature are input into a database. If there are repeated people in the video stream within a unit time, they can be deleted to ensure that there is no repeated count of people in the target area. The activity coefficient, that is, the flow of people within a unit time.

[0031] In one embodiment, step S103 includes: Obtain historical electricity consumption data and historical temperature data in the target area, and input the historical electricity consumption data and historical temperature data into the long short-term memory network model for training; the historical electricity consumption data and historical temperature data correspond one to one in timestamps, and the historical temperature data is the temperature of the non-target area; The error value is calculated based on the output of the long short-term memory network model and the true value, the model parameters are obtained by converting the error value through the back propagation algorithm, and the weights and biases of the long short-term memory network model are updated according to the model parameters to obtain the power consumption prediction model; the output of the power consumption prediction model is the power consumption data corresponding to the temperature of the non-target area at the target timestamp; The second power consumption is input into the power consumption prediction model to obtain the predicted energy consumption coefficient.

[0032] In one implementation, temperature is an important factor affecting power consumption. In the hot summer, the use of air conditioners and refrigeration equipment will increase power consumption significantly; in the cold winter, the use of heating equipment or air conditioners will also increase power consumption. Incorporating temperature data into the prediction model can significantly improve the accuracy of power consumption prediction.

[0033] In one implementation, the long short-term memory network model is LSTM, which can capture long-term dependencies in time series data. It can use the long-term trends and periodic changes in historical electricity consumption data and temperature data to predict future electricity consumption; the relationship between electricity consumption and temperature may be nonlinear. LSTM has strong nonlinear modeling capabilities and can handle such complex nonlinear relationships, thereby more accurately predicting electricity consumption; the LSTM model can be trained to adapt to different data sets and scenarios. This means that it can adjust the prediction strategy according to different regions, seasons, and climate conditions, thereby providing more accurate prediction results. LSTM is highly robust to noise and outliers in time series data. For example: if there are some inaccurate or abnormal values ​​in the input data, LSTM can still give relatively accurate prediction results.

[0034] In one implementation, the output of the LSTM network model and the true value are used to calculate the error value, which quantifies the degree of inconsistency between the predicted result and the actual result. The core of the backpropagation algorithm is to calculate the gradient of the loss function with respect to the model parameters. The backpropagation algorithm propagates the gradient of the loss function backward layer by layer through the chain rule, starting from the output layer and propagating to the input layer. In the iterative update process, the four steps of forward propagation (calculating the predicted output), calculating the loss, backpropagation (calculating the gradient) and parameter update are repeated many times until the loss value converges to a lower level or the predetermined number of iterations is reached.

[0035] In one embodiment, step S104 includes: Energy consumption threshold calculation formula Calculate the energy consumption threshold; Among them, F is the energy consumption threshold, YC is the predicted energy consumption coefficient, AC is the activity coefficient, environmental parameters include: absolute temperature difference, absolute humidity difference and air quality, T is the temperature in the target sub-area, is the preset standard temperature, H is the humidity in the target sub-area, To preset standard humidity, is the air quality score in the target area, AQ is the carbon dioxide concentration in the target area, α and β are constant proportional coefficients and both are not zero.

[0036] In one implementation, is the absolute temperature difference, is the absolute humidity difference. When any one of the absolute temperature difference, absolute humidity difference and air quality increases, the air conditioning system will run, and the energy consumption will increase. When the activity coefficient increases, it means that there is a large flow of people in the target sub-area. At this time, the air conditioning system needs to increase its power to make the indoor environment comfortable. For example, in summer, the indoor temperature is very high, so the air conditioning system starts cooling. The room is closed and there are many people, which leads to a decrease in air quality, that is, an increase in carbon dioxide concentration, and the fresh air system needs to be turned on to improve ventilation.

[0037] In one embodiment, step S105 includes: If the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold, the operating power and operating mode of the power equipment are changed; If the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold, the working power of the power equipment is changed and the working mode of the power equipment is maintained; If the activity coefficient in the target sub-area is not greater than the activity threshold, and the environmental parameter is not greater than the environmental parameter threshold, the working power and working mode of the power equipment are maintained; If the activity coefficient in the target sub-area is not greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold, the working mode of the power equipment is changed and the working power of the power equipment is maintained.

[0038] In one implementation, the working mode is specifically determined by environmental parameters, the absolute temperature difference corresponds to the cooling or heating mode, the absolute humidity difference corresponds to the dehumidification mode, and the air quality corresponds to the fresh air mode; the working power is determined by the activity coefficient, when the activity coefficient is less than the first number threshold, the wind intensity is determined to be weak, when the first number threshold is less than the activity coefficient and less than the second number threshold, the wind intensity is determined to be medium, and when the second number threshold is less than the activity coefficient, the wind intensity is determined to be strong; environmental parameters When the activity coefficient is greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold, it means that there are many people indoors and the indoor environment is very poor. In this case, the working power and working mode of the power equipment are changed, for example, the power of the air conditioner is adjusted. The power adjustment is specifically to change the intensity of the air conditioner. The intensity includes: weak, medium and strong. The working mode is controlled, for example: cooling, heating, fresh air or dehumidification. When the activity coefficient is greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold, that is, there are many people indoors but the indoor environment is very suitable for people to work, the working mode is kept unchanged and the intensity of the air is increased.

[0039] In one implementation, when the activity coefficient is not greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold, that is, there are not many people indoors and the environment is suitable, the working power and working mode of the power equipment are maintained and no changes are required; when the activity coefficient is not greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold, indicating that there are not many people and the environment is not suitable, the working mode of the power equipment is changed and the working power of the power equipment is maintained, that is, switching between cooling, heating, fresh air or dehumidification, and the wind intensity does not change.

[0040] Based on the same inventive concept, the embodiment of the present invention also provides a building energy management device based on an embedded industrial computer. Figure 2 , Figure 2 A schematic diagram of the structure of a building energy management device based on an embedded industrial computer provided by an embodiment of the present invention includes: a first power consumption acquisition module, a video stream acquisition module, a second power consumption acquisition module, a threshold determination module and a management scheme determination module: A first power consumption acquisition module is used to acquire the power consumption of each target sub-area in the target area to obtain a first power consumption data set, and use the first power consumption as the actual energy consumption coefficient; the first power consumption data set includes the first power consumption data of each target sub-area; The video stream acquisition module is used to periodically acquire the video stream of each target sub-area, identify the flow of people in the video stream and calculate the activity coefficient; A second power consumption acquisition module is used to acquire a second power consumption, and input the second power consumption into the power consumption prediction model to obtain a predicted energy consumption coefficient; the second power consumption is the power consumption of the previous cycle of the first power consumption data; A threshold determination module is used to obtain the environmental parameters of each target sub-area and calculate the energy consumption threshold according to the predicted energy consumption coefficient, activity coefficient and environmental parameters; the environmental parameters are the absolute temperature difference, absolute humidity difference and air quality in the target sub-area; The management scheme determination module is used to determine an energy management scheme for each target sub-area in the target area if the actual energy consumption coefficient is greater than the energy consumption threshold.

[0041] A building energy management device based on an embedded industrial computer provided in an embodiment of the present invention can dynamically set an energy consumption threshold by comprehensively considering the predicted energy consumption coefficient, activity coefficient and environmental parameters, and can promptly formulate a targeted energy management plan when the actual energy consumption exceeds the threshold, so that the plan formulation is more efficient and more accurate, reducing energy consumption while ensuring the comfort of personnel.

[0042] In one embodiment, the video stream acquisition module includes: a frame extraction frequency determination module, a foreground feature extraction module, a human frame diagram construction module, a gait recognition module and an activity coefficient calculation module: A frame extraction frequency determination module is used to determine the working state of the embedded industrial computer and determine the frame extraction frequency of the video stream according to the working state; The foreground feature extraction module is used to extract frames from the video stream according to the frame extraction frequency to obtain an image set, perform background recognition on the images in the image set, and remove the background features in the image to obtain the foreground features; A human frame diagram construction module is used to perform two-dimensional contour recognition on foreground features and determine key nodes, and construct a human frame diagram based on the key nodes; the key nodes are determined by the activity areas of human joints; A gait recognition module is used to extract the gait cycle in the image set according to the human frame diagram, and perform gait recognition according to the gait cycle to obtain gait features; the gait cycle is the time from when the heel of one side touches the ground to when the heel of the same side touches the ground again during the movement of the human body; The activity coefficient calculation module is used to extract pixel features from foreground features, determine the flow of people in the target sub-area based on gait features and pixel features, and calculate the activity coefficient.

[0043] In one embodiment, the second power consumption acquisition module includes: a training data input module, a model parameter updating module and a predicted energy consumption coefficient acquisition module: The training data input module is used to obtain the historical electricity consumption data and historical temperature data in the target area, and input the historical electricity consumption data and historical temperature data into the long short-term memory network model for training; the historical electricity consumption data and the historical temperature data correspond one to one in terms of timestamps, and the historical temperature data is the temperature of the non-target area; The model parameter updating module is used to calculate the error value according to the output of the long short-term memory network model and the true value, convert the error value to obtain the model parameter through the back propagation algorithm, and update the weight and bias of the long short-term memory network model according to the model parameters to obtain the power consumption prediction model; the output of the power consumption prediction model is the power consumption data corresponding to the temperature of the non-target area at the target timestamp; The predicted energy consumption coefficient acquisition module is used to input the second power consumption into the power consumption prediction model to obtain the predicted energy consumption coefficient.

[0044] In one embodiment, the threshold determination module is further configured to calculate the energy consumption threshold through the energy consumption threshold calculation formula Calculate the energy consumption threshold; Among them, F is the energy consumption threshold, YC is the predicted energy consumption coefficient, AC is the activity coefficient, environmental parameters include: absolute temperature difference, absolute humidity difference and air quality, T is the temperature in the target sub-area, is the preset standard temperature, H is the humidity in the target sub-area, To preset standard humidity, is the air quality score in the target area, AQ is the carbon dioxide concentration in the target area, α and β are constant proportional coefficients and both are not zero.

[0045] In one embodiment, the management solution determination module includes: a first management solution module, a second management solution module, a third management solution module and a fourth management solution module: A first management scheme module is used to change the working power and working mode of the electric equipment if the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold; A second management scheme module is used to change the working power of the power equipment and maintain the working mode of the power equipment if the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold; A third management solution module is used to maintain the working power and working mode of the power equipment if the activity coefficient in the target sub-area is not greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold; The fourth management scheme module is used to change the working mode of the power equipment and maintain the working power of the power equipment if the activity coefficient in the target sub-area is not greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold.

[0046] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A building energy management method based on embedded industrial computers, characterized in that: The method comprises: Acquire the power consumption of each target sub-area in the target area to obtain a first power consumption data set, and use the first power consumption as the actual energy consumption coefficient; the first power consumption data set includes the first power consumption data of each target sub-area; Periodically obtain the video stream of each target sub-area, identify the flow of people in the video stream and calculate the activity coefficient; Obtaining a second power consumption, and inputting the second power consumption into a power consumption prediction model to obtain a predicted energy consumption coefficient; the second power consumption is the power consumption of a period before the first power consumption data; Acquire environmental parameters of each target sub-area, and calculate an energy consumption threshold value according to the predicted energy consumption coefficient, activity coefficient and environmental parameters; the environmental parameters are the absolute temperature difference, absolute humidity difference and air quality in the target sub-area; If the actual energy consumption coefficient is greater than the energy consumption threshold, an energy management plan is determined for each target sub-area in the target area.

2. The building energy management method based on embedded industrial computer according to claim 1 is characterized in that: Identifying the flow of people in the video stream and calculating the activity coefficient include: Determine the working state of the embedded industrial computer, and determine the frame extraction frequency of the video stream according to the working state; Extracting frames from the video stream at the frame extraction frequency to obtain an image set, performing background recognition on images in the image set, and removing background features in the images to obtain foreground features; Performing two-dimensional contour recognition on the foreground features and determining key nodes, and constructing a human frame diagram according to the key nodes; the key nodes are determined by the activity areas of human joints; Extracting a gait cycle from the image set according to the human frame diagram, and performing gait recognition according to the gait cycle to obtain gait features; the gait cycle is the time from when the heel of one side touches the ground to when the heel of the same side touches the ground again during the movement of the human body; Pixel features in the foreground features are extracted, and the flow of people in the target sub-area is determined according to the gait features and the pixel features, and the activity coefficient is calculated.

3. The building energy management method based on embedded industrial computer according to claim 1 is characterized in that: Inputting the second power consumption into the power consumption prediction model to obtain the predicted energy consumption coefficient includes: Obtain historical power consumption data and historical temperature data in the target area, and input the historical power consumption data and the historical temperature data into a long short-term memory network model for training; the historical power consumption data and the historical temperature data correspond one to one in terms of timestamps, and the historical temperature data is the temperature of a non-target area; An error value is calculated based on the output of the long short-term memory network model and the true value, the error value is converted into a model parameter through a back propagation algorithm, and the weight and bias of the long short-term memory network model are updated according to the model parameters to obtain a power consumption prediction model; the output of the power consumption prediction model is the power consumption data corresponding to the temperature of the non-target area at the target timestamp; The second power consumption is input into the power consumption prediction model to obtain a predicted energy consumption coefficient.

4. The building energy management method based on embedded industrial computer according to claim 1 is characterized in that: The energy consumption threshold calculated according to the predicted energy consumption coefficient, activity coefficient and environmental parameters includes: Energy consumption threshold calculation formula Calculate the energy consumption threshold; Among them, F is the energy consumption threshold, YC is the predicted energy consumption coefficient, AC is the activity coefficient, environmental parameters include: absolute temperature difference, absolute humidity difference and air quality, T is the temperature in the target sub-area, is the preset standard temperature, H is the humidity in the target sub-area, To preset standard humidity, is the air quality score in the target area, AQ is the carbon dioxide concentration in the target area, α and β are constant proportional coefficients and both are not zero.

5. The building energy management method based on embedded industrial computer according to claim 1, characterized in that: Determining an energy management plan for each target sub-area within the target area includes: If the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold, the operating power and operating mode of the power equipment are changed; If the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold, the working power of the power equipment is changed and the working mode of the power equipment is maintained; If the activity coefficient in the target sub-area is not greater than the activity threshold, and the environmental parameter is not greater than the environmental parameter threshold, the working power and working mode of the power equipment are maintained; If the activity coefficient in the target sub-area is not greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold, the working mode of the power equipment is changed and the working power of the power equipment is maintained.

6. A building energy management device based on an embedded industrial computer, characterized in that: The device comprises: a first power consumption acquisition module, a video stream acquisition module, a second power consumption acquisition module, a threshold determination module and a management scheme determination module: The first power consumption acquisition module is used to acquire the power consumption of each target sub-area in the target area to obtain a first power consumption data set, and use the first power consumption as the actual energy consumption coefficient; the first power consumption data set includes the first power consumption data of each target sub-area; The video stream acquisition module is used to periodically acquire the video stream of each target sub-area, identify the flow of people in the video stream and calculate the activity coefficient; The second power consumption acquisition module is used to acquire a second power consumption, and input the second power consumption into the power consumption prediction model to obtain a predicted energy consumption coefficient; the second power consumption is the power consumption of the previous cycle of the first power consumption data; The threshold determination module is used to obtain the environmental parameters of each target sub-area, and calculate the energy consumption threshold according to the predicted energy consumption coefficient, the activity coefficient and the environmental parameters; the environmental parameters are the absolute temperature difference, the absolute humidity difference and the air quality in the target sub-area; The management scheme determination module is used to determine an energy management scheme for each target sub-area in the target area if the actual energy consumption coefficient is greater than the energy consumption threshold.

7. A building energy management device based on an embedded industrial computer according to claim 6, characterized in that: The video stream acquisition module includes: a frame extraction frequency determination module, a foreground feature extraction module, a human frame diagram construction module, a gait recognition module and an activity coefficient calculation module: The frame extraction frequency determination module is used to determine the working state of the embedded industrial computer, and determine the frame extraction frequency of the video stream according to the working state; The foreground feature extraction module is used to extract frames from the video stream according to the frame extraction frequency to obtain an image set, perform background recognition on the images in the image set, and remove background features in the image to obtain foreground features; The human frame diagram construction module is used to perform two-dimensional contour recognition on the foreground features and determine key nodes, and construct a human frame diagram according to the key nodes; the key nodes are determined by the activity areas of human joints; The gait recognition module is used to extract the gait cycle in the image set according to the human frame diagram, and perform gait recognition according to the gait cycle to obtain gait features; the gait cycle is the time from when the heel of one side touches the ground to when the heel of the same side touches the ground again during the movement of the human body; The activity coefficient calculation module is used to extract pixel features from foreground features, determine the flow of people in the target sub-area according to the gait features and the pixel features, and calculate the activity coefficient.

8. The building energy management device based on embedded industrial computer according to claim 6, characterized in that: The second power consumption acquisition module includes: a training data input module, a model parameter update module and a predicted energy consumption coefficient acquisition module: The training data input module is used to obtain historical power consumption data and historical temperature data in the target area, and input the historical power consumption data and the historical temperature data into the long short-term memory network model for training; the historical power consumption data and the historical temperature data correspond one to one in terms of timestamps, and the historical temperature data is the temperature of the non-target area; The model parameter updating module is used to calculate an error value according to the output of the long short-term memory network model and the true value, convert the error value to obtain model parameters through a back propagation algorithm, and update the weights and biases of the long short-term memory network model according to the model parameters to obtain a power consumption prediction model; the output of the power consumption prediction model is the power consumption data corresponding to the non-target area temperature at the target timestamp; The predicted energy consumption coefficient acquisition module is used to input the second power consumption into the power consumption prediction model to obtain the predicted energy consumption coefficient.

9. The building energy management device based on embedded industrial computer according to claim 6, characterized in that: The threshold determination module is also used to calculate the energy consumption threshold through the energy consumption threshold calculation formula Calculate the energy consumption threshold; Among them, F is the energy consumption threshold, YC is the predicted energy consumption coefficient, AC is the activity coefficient, environmental parameters include: absolute temperature difference, absolute humidity difference and air quality, T is the temperature in the target sub-area, is the preset standard temperature, H is the humidity in the target sub-area, To preset standard humidity, is the air quality score in the target area, AQ is the carbon dioxide concentration in the target area, α and β are constant proportional coefficients and both are not zero.

10. A building energy management device based on an embedded industrial computer according to claim 6, characterized in that: The management scheme determination module includes: a first management scheme module, a second management scheme module, a third management scheme module and a fourth management scheme module: The first management solution module is used to change the working power and working mode of the power equipment if the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold; The second management solution module is used to change the working power of the power equipment and maintain the working mode of the power equipment if the activity coefficient in the target sub-area is greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold; The third management solution module is used to maintain the working power and working mode of the power equipment if the activity coefficient in the target sub-area is not greater than the activity threshold and the environmental parameter is not greater than the environmental parameter threshold; The fourth management scheme module is used to change the working mode of the power equipment and maintain the working power of the power equipment if the activity coefficient in the target sub-area is not greater than the activity threshold and the environmental parameter is greater than the environmental parameter threshold.

Images