Personalized precise energy management method for nursing home based on portrait model

Abstract

The invention discloses a nursing home personalized precise energy management method based on a portrait model, and the method comprises the steps: dividing a nursing home into an old person bedroom and a functional room, enabling the nursing home to be provided with a corresponding energy supply device, and building a nursing home energy system digital twinborn model; collecting multi-source data of each room of the nursing home, and constructing a portrait model of each room; based on the nursing home energy system digital twinborn model and the portrait model of each room, in combination with the monitored indoor and outdoor air environment information and the dynamic influence factors, establishing a time-phased air environment multi-index demand prediction model of each room of the nursing home; based on the air environment multi-index demand prediction model of each room of the nursing home in different time periods and the portrait model of each room, and in combination with the operation characteristics of the corresponding energy supply device matched with each room, establishing an energy supply device regulation and control model of each room in different time periods, and obtaining a regulation and control strategy of the energy supply device; and based on the nursing home energy system digital twinborn model, verification and strategy issuing execution are carried out on the regulation and control strategy of the energy supply device.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy management in nursing homes, and particularly relates to a personalized and precise energy management method for nursing homes based on a portrait model. Background Art

[0002] With the deepening of the global aging degree, the issue of elderly care has attracted more and more attention. The construction and management of nursing homes have become important social issues. Traditional nursing homes generally have problems such as irregular management, simple facilities, and low service quality. With the continuous development and popularization of Internet of Things technology, intelligent nursing home management systems have received extensive attention and applications, and through technical means such as sensors and network communication, the health management and monitoring of the elderly are realized, improving the quality of life of the elderly.

[0003] The quality of life of the elderly in nursing homes is closely related to the indoor living environment. For example, the ability of the elderly to regulate heat and cold declines, and they are sensitive to changes in the external environment and need a comfortable indoor temperature; in addition, too low indoor humidity is likely to cause diseases and breed bacteria, and when the indoor humidity is too high, it is stuffy in summer and cold in winter. In addition, the facilities in the nursing home are mainly distributed in the elderly's bedroom rooms and functional rooms such as restaurants, medical rooms, rehabilitation rooms, and activity rooms. The physical health conditions of different elderly people in different rooms in the nursing home are different, and their daily life and behavioral activities are also different, indirectly resulting in different energy demands for temperature, humidity, ventilation volume, etc. Therefore, how to carry out personalized and precise energy management for nursing homes based on the energy demands of each room in the nursing home, create a more comfortable and healthy living environment for the elderly in the nursing home, and promote the digital transformation of the elderly care service industry is an urgent problem to be solved at present.

[0004] Based on the above technical problems, it is necessary to design a new personalized and precise energy management method for nursing homes based on a portrait model. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the deficiencies of the prior art and provide a personalized and precise energy management method for nursing homes based on a portrait model. By constructing a portrait model for different rooms in the nursing home, accurately predicting the air environment demands of each room at different times, formulating the optimal regulation strategy, providing a personalized air environment adjustment plan, improving the comfort of the elderly's living and daily activities, and promoting the digital transformation of the elderly care service industry.

[0006] To solve the above technical problems, the technical solution of the present invention is as follows:

[0007] The present invention provides a personalized and precise energy management method for nursing homes based on a portrait model, which includes:

[0008] Step S1: Divide the nursing home into elderly people's bedroom rooms and functional rooms, and equip them with corresponding energy supply devices. Establish a digital twin model of the nursing home energy system using mechanism modeling and data identification methods;

[0009] Step S2: Collect multi-source data of each room in the nursing home and construct a portrait model for each room;

[0010] Step S3: Based on the digital twin model of the nursing home energy system and the portrait models of each room, combined with the monitored indoor and outdoor air environment information and dynamic influencing factors, establish a multi-index demand prediction model for the air environment of each room in the nursing home at different times;

[0011] Step S4: Based on the multi-index demand prediction model for the air environment of each room in the nursing home at different times and the portrait models of each room, combined with the operating characteristics of the corresponding energy supply devices equipped in each room, establish a regulation model for the energy supply devices of each room at different times, and obtain the regulation strategy of the energy supply devices;

[0012] Step S5: Verify the regulation strategy of the energy supply devices based on the digital twin model of the nursing home energy system and issue and execute the strategy.

[0013] Furthermore, in the above-mentioned Step S1, dividing the nursing home into elderly people's bedroom rooms and functional rooms and equipping them with corresponding energy supply devices includes:

[0014] According to the actual facility distribution of the nursing home, divide it into elderly people's bedroom rooms and functional rooms; the apartment types of the elderly people's bedroom rooms include single-bedroom, double-bedroom, and triple-bedroom; the functional rooms include a dining room, an activity room, a rehabilitation room, a medical room, and a conference hall;

[0015] Each room is equipped with an energy supply device, including at least a winter heating device, a summer cooling device, a humidity adjustment device, a ventilation and air change adjustment device, and an oxygen supply device, which are used to regulate the air environment indicators of each room.

[0016] Furthermore, in the above-mentioned Step S1, establishing a digital twin model of the nursing home energy system using mechanism modeling and data identification methods specifically includes:

[0017] Based on the actual facility distribution of each room in the nursing home and the equipped energy supply devices, determine the physical entities of the nursing home energy system, including room energy-using physical entities and energy supply physical entities;

[0018] Establish a geometric model, a physical model, a behavior model, and a rule model of the physical entities of the nursing home energy system. After integrating, fusing, and performing consistency verification on the models using virtual reality technology, construct a virtual entity mechanism model that is highly consistent with the actual physical entities of the nursing home energy system;

[0019] Among them, the geometric model is used to describe the spatial positions, sizes, and shape characteristics of each room and the supporting energy supply devices in the nursing home; the physical model is used to add the physical properties, constraints, and characteristics of the physical entities of the nursing home energy system on the basis of the geometric model; the behavior model is used to describe the behaviors and behavior evolutions of the physical entities of the nursing home energy system under different internal and external influencing factors and action mechanisms; the rule model is used to describe the criteria, knowledge, and experiences of the physical entities of the nursing home energy system in actual scenario applications.

[0020] Taking the virtual entity mechanism model as the main body, calculating the residuals of the virtual entity mechanism model using historical data, preprocessing the input and the model residual results of the virtual entity mechanism model and respectively using them as the input and output of the machine learning network, obtaining a digital twin data-driven model after training and learning, and compensating and correcting the residuals of the virtual entity mechanism model.

[0021] Fusing the corrected virtual entity mechanism model and the digital twin data-driven model to establish a digital twin model of the nursing home energy system.

[0022] Furthermore, the specific steps of step S2 include:

[0023] Collecting multi-source data of the elderly's bedroom rooms and functional rooms in the nursing home through the data support layer, including static physical data and dynamic behavior data.

[0024] Performing cleaning operations such as standardization processing, outlier processing, dimensionality reduction processing, filling in null values, and removing duplicate values on the multi-source data of each room in the nursing home collected through the data processing layer to generate preprocessed multi-source data of each room in the nursing home.

[0025] Using deep learning methods in the label extraction layer to extract different attribute features from the preprocessed multi-source data of each room in the nursing home to form multi-dimensional labels of the portraits of each room, including at least static attribute labels, health attribute labels, daily routine behavior labels, and energy consumption behavior labels of the elderly's bedroom rooms, as well as static attribute labels, activity behavior labels, and energy consumption behavior labels of functional rooms.

[0026] Using the multi-dimensional labels of the portraits of each room in the portrait modeling layer to construct portrait models of each room.

[0027] Furthermore, the static physical data of the elderly's bedroom rooms includes: the housing type of the elderly's bedroom rooms, the room area, the room spatial layout, the age, gender, physical function, and medical history of the elderly.

[0028] The dynamic behavior data of the elderly's bedroom rooms includes: the energy consumption habits of the elderly, the daily routines in the elderly's rooms, the activity habits outside the elderly's rooms, and the operation information of the energy supply equipment.

[0029] The static physical data of the functional rooms include: room function type, room area, and room spatial layout;

[0030] The dynamic behavior data of the functional rooms include: the number of uses and duration in different time periods of the rooms, personnel flow behavior, personnel participation, comfort evaluation information of each room, the changing trend of energy consumption in each room, and operation information of the energy supply equipment.

[0031] Furthermore, after constructing the portrait models of each room, it also includes: setting an update period, regularly re - collecting multi - source data of each room in the nursing home to update the portrait models of each room, and when it is detected that the static physical data of the elderly's bedrooms and functional rooms in the nursing home change, or the dynamic behavior data changes exceed the preset range, actively updating the portrait models of each room.

[0032] Furthermore, step S3 specifically includes:

[0033] Obtain the indoor and outdoor air environment information, energy consumption data, and dynamic influencing factors monitored in each room at different time periods based on the digital twin model of the nursing home energy system;

[0034] Take the indoor and outdoor air environment information, energy consumption data, dynamic influencing factors monitored in each room at different time periods, and the portrait label data reflecting the characteristics of each room obtained based on the portrait models of each room as the input data of the multi - index demand prediction model of the air environment, and divide the input data into input data for multiple time intervals;

[0035] With room comfort as the goal and the multi - index of the air environment that conforms to the portrait models of each room as the output data, extract the important data features that affect the multi - index of the air environment of each room, and input them into a machine learning algorithm for training and learning to establish a multi - index demand prediction model of the air environment for each room in the nursing home at different time periods, and output the multi - index demand prediction model of the air environment for each room at different time periods; the multi - index of the air environment includes winter heat demand index, summer cooling demand index, humidity demand index, fresh air volume demand index, and oxygen content demand index.

[0036] Furthermore, the indoor and outdoor air environment information at least includes indoor and outdoor temperature, wind speed, humidity, oxygen concentration, carbon dioxide concentration, and PM2.5 concentration; the energy consumption data includes the operation data of winter heating devices, summer cooling devices, humidity adjustment devices, ventilation and air exchange adjustment devices, and oxygen supply devices; the dynamic influencing factors include the monitored room occupancy and the physical indicators of the elderly such as body temperature and heart rate.

[0037] Furthermore, step S4 specifically includes:

[0038] Based on the multi-index demand prediction model for the air environment of each room in the nursing home in different time periods, the multi-index demand prediction value for the air environment of each room in different time periods is obtained, and combined with the portrait label data obtained based on the portrait model of each room, the operating data of the corresponding energy supply device in each room, the performance change curve of the energy supply device under different operating conditions and the regulation data of the energy supply device, they are used as the sample data set;

[0039] With the goal of meeting the multi-indicator demand prediction values ​​of the air environment in each room of the nursing home in different time periods, the sample data set is input into the machine learning algorithm for training and learning, and the energy supply device control model for each room in different time periods is established. The control strategy of the energy supply device is obtained, including the control parameters of the actuator components of the winter heating device, summer cooling device, humidity control device, ventilation control device and oxygen supply device in each room in different time periods.

[0040] Further, the step S5 specifically includes:

[0041] After simulating and deducing the control strategy of the energy supply device based on the digital twin model of the nursing home energy system, it is verified and evaluated whether the multiple indicators of the air environment in each room in different time periods meet the expected goals. If so, the strategy is issued and executed; otherwise, the control strategy of the energy supply device in each room in different time periods is readjusted.

[0042] The beneficial effects of the present invention are:

[0043] By constructing portrait models for different rooms in a nursing home, the present invention can provide personalized air environment adjustment solutions based on the needs of the elderly in different rooms, thereby improving the comfort of the elderly's living and daily activities; in addition, by utilizing the digital twin model, the portrait models of each room and multi-source data analysis, the air environment requirements of each room at different time periods can be accurately predicted to avoid energy waste caused by excessive power supply, and based on the established power supply device control model, the optimal control strategy can be formulated to improve the equipment operation efficiency and meet the air environment requirements of each room. Ultimately, it realizes the comprehensive perception, analysis, prediction and control of the environment and data in the nursing home, improves the service quality, reduces operating costs, and creates a more comfortable and healthy living environment for the elderly in nursing homes, promoting the digital transformation and technological progress of the elderly care service industry.

[0044] Other features and advantages will be described in the following description, and partly become apparent from the description, or understood by practicing the invention. The purpose and other advantages of the invention are realized and obtained by the structures particularly pointed out in the description and the drawings.

[0045] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0047] Figure 1 Flow chart of a personalized and precise energy management method for a nursing home based on an image model of the present invention;

[0048] Figure 2 Principle block diagram for constructing the image model of each room in the nursing home of the present invention. Specific embodiments

[0049] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions of the present invention with reference to the drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0050] As Figure 1 shown, this embodiment provides a personalized and precise energy management method for a nursing home based on an image model, which includes:

[0051] Step S1: Divide the nursing home into elderly people's bedroom rooms and functional rooms, and be equipped with corresponding energy supply devices. Use mechanism modeling and data identification methods to establish a digital twin model of the nursing home energy system;

[0052] Step S2: Collect multi-source data of each room in the nursing home and construct an image model for each room;

[0053] Step S3: Based on the digital twin model of the nursing home energy system and the image models of each room, combined with the monitored indoor and outdoor air environment information and dynamic influencing factors, establish a multi-index demand prediction model for the air environment of each room in the nursing home at different time periods;

[0054] Step S4: Based on the multi-index demand prediction model for the air environment of each room in the nursing home at different time periods and the image models of each room, combined with the operating characteristics of the corresponding energy supply devices equipped in each room, establish a regulation model for the energy supply devices of each room at different time periods, and obtain the regulation strategy for the energy supply devices;

[0055] Step S5: Verify the regulation strategy for the energy supply devices based on the digital twin model of the nursing home energy system and issue and execute the strategy.

[0056] In this embodiment, in step S1, the nursing home is divided into elderly people's bedroom rooms and functional rooms, and corresponding energy supply devices are provided, including:

[0057] According to the actual facility distribution of the nursing home, it is divided into elderly people's bedroom rooms and functional rooms; the types of the elderly people's bedroom rooms include single-bedroom rooms, double-bedroom rooms, and triple-bedroom rooms; the functional rooms include a dining hall, an activity room, a rehabilitation room, a medical room, and a conference hall;

[0058] Each room is equipped with an energy supply device, which at least includes a winter heating device, a summer cooling device, a humidity adjustment device, a ventilation and air change adjustment device, and an oxygen supply device, for regulating the air environment indexes of each room.

[0059] In actual applications, a nursing home usually has multiple rest buildings, mainly double rooms and multi-person rooms, with a small number of single rooms and suites; it is also equipped with a dining hall, an activity room, a rehabilitation room, a medical room, and a conference hall for the elderly to dine, engage in interest activities, and receive rehabilitation physiotherapy. For example, the rehabilitation room can be equipped with rehabilitation therapists to conduct rehabilitation training for the elderly with rehabilitation requirements, the medical room is responsible for diagnosis and treatment services by professional medical staff, with outpatient departments for various departments and emergency services provided, and the activity room is equipped with complete leisure and entertainment facilities and diverse entertainment programs. The winter heating device and summer cooling device of the nursing home can choose central air conditioning for heating and cooling, or choose a central heating method for heating in winter, and each room can be individually regulated for heating, choose an air conditioner unit for cooling in summer, use the ventilation and air change adjustment device to adjust the fresh air volume, and use the oxygen supply device to adjust the indoor oxygen content.

[0060] In this embodiment, in step S1, a digital twin model of the nursing home energy system is established by using mechanism modeling and data identification methods, specifically including:

[0061] Based on the actual facility distribution of each room in the nursing home and the supporting energy supply devices, determine the physical entities of the nursing home energy system, including the physical entities of room energy consumption and the physical entities of energy supply;

[0062] Establish a geometric model, a physical model, a behavior model, and a rule model of the physical entities of the nursing home energy system, and after integrating, fusing, and performing consistency verification on the models by using virtual reality technology, construct a virtual entity mechanism model that is highly consistent with the physical entities of the actual nursing home energy system;

[0063] Among them, the geometric model is used to describe the spatial positions, sizes, and shape characteristics of each room and the supporting energy supply devices in the nursing home; the physical model is used to add the physical properties, constraints, and characteristics of the physical entities of the nursing home energy system on the basis of the geometric model; the behavior model is used to describe the behaviors and behavior evolutions of the physical entities of the nursing home energy system under different internal and external influencing factors and action mechanisms; the rule model is used to describe the criteria, knowledge, and experiences of the physical entities of the nursing home energy system in actual scenario applications.

[0064] Taking the virtual entity mechanism model as the main body, the residuals of the virtual entity mechanism model are calculated using historical data. After preprocessing the input and the model residual results of the virtual entity mechanism model, they are respectively used as the input and output of the machine learning network. After training and learning, a digital twin data-driven model is obtained, and the residuals of the virtual entity mechanism model are compensated and corrected.

[0065] Fuse the corrected virtual entity mechanism model and the digital twin data-driven model to establish a digital twin model of the nursing home energy system.

[0066] As Figure 2 shown, in this embodiment, the step S2 specifically includes:

[0067] Collect multi-source data of the elderly's bedroom rooms and functional rooms in the nursing home through the data support layer, including static physical data and dynamic behavior data.

[0068] Through the data processing layer, perform cleaning operations on the multi-source data of each room in the nursing home, such as standardization processing, outlier processing, dimensionality reduction processing, filling in null values, and removing duplicate values, to generate preprocessed multi-source data of each room in the nursing home.

[0069] Through the label extraction layer, use deep learning methods to extract different attribute characteristics from the preprocessed multi-source data of each room in the nursing home to form multi-dimensional labels of each room portrait, including at least static attribute labels, health attribute labels, daily routine behavior labels, and energy consumption behavior labels of the elderly's bedroom rooms, as well as static attribute labels, activity behavior labels, and energy consumption behavior labels of functional rooms.

[0070] Through the portrait modeling layer, use the multi-dimensional labels of each room portrait to construct each room portrait model.

[0071] In practical applications, the structure for establishing portrait models of each room in a nursing home includes a data support layer, a data processing layer, a label extraction layer, and a portrait modeling layer. The data support layer is used to collect multi-source data required for constructing portraits of each room, including data of the room itself and data information of the people in the room, etc.; the data processing layer is used to preprocess the collected multi-source data; the label extraction layer is used to extract feature labels by means of various tools and deep learning algorithms; the portrait modeling layer is used to fuse the extracted multi-dimensional feature labels to realize the construction of the portrait model.

[0072] Generally, the multi-source data collected is text information. When extracting labels from the preprocessed multi-source data, the text features of each room in the nursing home are comprehensively captured from two dimensions of characters and words to improve the utilization rate of text data and semantic representation ability. The text feature vectors of the two dimensions are concatenated in sequence to obtain a multi-dimensional fusion vector, which is then input into the feature extraction layer. The BiGRU-DAE-Attention feature module is used for feature extraction to obtain high-quality key information, and finally it is input into the softmax classifier for classification to complete the establishment of portrait labels for each room.

[0073] BiGRU is an improved model based on the gated recurrent unit network. It adds another layer of GRU network to process data in reverse on the basis of the GRU network. At each moment, the state calculation of the GRU network includes an update gate, a reset gate, a candidate value, and a hidden state. The update gate and the reset gate are used to control the flow of information, and the candidate value and the hidden state control the output of the node. There will be a large amount of redundant or irrelevant information in the text data, which will interfere with the learning of key features. The denoising autoencoder DAE can effectively process redundant information and noise, reduce the influence of the original data noise, and improve the learning effect of key features. Then, the Attention module is used to calculate the weights between different positions of the features, reduce the dependence of the model on external information, and improve the performance of the model.

[0074] In this embodiment, the static physical data of the elderly's bedroom includes: the house type of the elderly's bedroom, the room area, the room space layout, the age, gender, physical function, and medical history of the elderly;

[0075] The dynamic behavior data of the elderly's bedroom includes: the energy consumption habits of the elderly, the daily routine in the elderly's room, the activity habits outside the elderly's room, and the operation information of the energy supply equipment;

[0076] The static physical data of the functional room includes: the room function type, the room area, and the room space layout;

[0077] The dynamic behavior data of the functional rooms include: the usage times and durations of the rooms at different time periods, personnel flow behaviors, personnel participation levels, comfort evaluation information for each room, the changing trends of energy consumption in each room, and the operation information of the energy supply equipment.

[0078] In actual applications, before the elderly move into the nursing home, the basic information, hobbies, etc. of the elderly will be understood, and the physical health status of the elderly will be evaluated. The elderly with the same living habits and hobbies will be arranged in the same room as much as possible. The physical functions of the elderly include self-care, semi-self-care, and non-self-care conditions; the medical history of the elderly is used to understand whether the elderly have diseases such as diabetes, heart disease, asthma, etc., which is convenient for attention during the daily meals of the elderly, and certain diseases have corresponding requirements for the indoor air environment.

[0079] The energy usage habits and preferences of the elderly will be correspondingly affected by their physiological characteristics, which can be considered when building the user portraits of the rooms. As the elderly age, their physiological mechanisms will degenerate, resulting in a significant decline in the body's mechanisms. Therefore, compared with the young, the self-regulation and adaptation abilities of the elderly's bodies to changes in the external environment are constantly declining, which is mainly reflected in the following aspects:

[0080] 1) Decline in sensory function

[0081] During the aging process of the elderly, their sensory abilities also gradually decline, which in turn affects the elderly's adaptability to environmental changes;

[0082] 2) Degeneration of immune function

[0083] The degeneration of immune function will cause the resistance of the elderly to decline, the ability to adapt to environmental changes to weaken, be more sensitive to the environment, and have a weaker resistance to diseases. Therefore, the indoor environment must have stable and comfortable thermal conditions, and at the same time, the ventilation is also required to meet higher requirements;

[0084] 3) Degeneration of the motor system

[0085] The activity behaviors of the elderly are related to the changes in age. As they age, their limb flexibility, energy, and vitality also decline, and their exercise volume will also decrease significantly, and they may even be unable to perform strenuous exercises;

[0086] According to the physiological characteristics of the elderly, to meet the requirements of the elderly's physiological comfort, first of all, the indoor environmental temperature needs to be maintained stable and ensure that it can be in a suitable temperature range. Secondly, the indoor air humidity value should be kept appropriate to reduce the related diseases caused by the humid environment for the elderly. For example, too high humidity is likely to cause rheumatism-related diseases and is also prone to breeding bacteria. And a hot and dry indoor environment can also cause diseases of the mouth, nose and respiratory system. In addition, indoor ventilation is also very important. In summer, it is hot, and in winter, it is cold. The frequency of opening windows for ventilation and air exchange is relatively limited. And dirty indoor air is likely to cause bacteria to breed and related diseases. And ventilating through doors and windows is likely to disrupt the stability of the indoor thermal environment. Therefore, a good ventilation system also has an important impact on improving the comfort of the elderly. In addition, to ensure the quality of life of the elderly, especially in special environments (such as high altitude or heavily polluted areas), measures should be taken to ensure good indoor air quality and the oxygen content is within a healthy range.

[0087] In this embodiment, after constructing the portrait models of each room, it further includes: setting an update period, regularly re-collecting multi-source data of each room in the nursing home to update the portrait models of each room, and when it is monitored that the static physical data or the dynamic behavior data of the elderly's bedroom rooms and functional rooms in the nursing home changes beyond the preset range, actively updating the portrait models of each room.

[0088] In this embodiment, step S3 specifically includes:

[0089] Obtain the indoor and outdoor air environment information, energy consumption data and dynamic influencing factors monitored in each room at different time periods based on the digital twin model of the nursing home energy system;

[0090] Take the indoor and outdoor air environment information, energy consumption data, dynamic influencing factors monitored in each room at different time periods and the portrait label data reflecting the characteristics of each room obtained based on the portrait models of each room as the input data of the multi-index demand prediction model of the air environment, and divide the input data into input data of multiple time intervals;

[0091] With the room comfort as the goal and the multi-index of the air environment conforming to the portrait models of each room as the output data, extract the important data features affecting the multi-index of the air environment of each room and input them into the machine learning algorithm for training and learning to establish a multi-index demand prediction model of the air environment for each room in the nursing home at different time periods, and output the multi-index demand prediction model of the air environment for each room at different time periods; the multi-index of the air environment includes the winter heat demand index, the summer cooling demand index, the humidity demand index, the fresh air volume demand index and the oxygen content demand index.

[0092] In actual applications, there are certain requirements for the daily living and living environment of the elderly, including:

[0093] Comfortable temperature range: Temperature has a direct impact on indoor thermal comfort. The ability of the elderly to regulate heat and cold declines, and they are more sensitive to changes in the external environment, requiring stable and comfortable temperature conditions. In addition, the environmental temperature also affects the individual's metabolic rate. When the environmental temperature is between 20 degrees and 30 degrees and the human body is in a static state, the body has a relatively stable metabolic rate.

[0094] Appropriate relative humidity: When the indoor humidity is below 30%, it is easy to cause irritating coughs and respiratory diseases and is also prone to bacterial growth. When the humidity is greater than 80%, the human body will feel stuffy and even suffer from heatstroke in summer, and will feel cold and damp and catch a cold in winter. Too low or too high indoor humidity is not conducive to the health of the elderly.

[0095] Ventilation: In summer, the air is humid, and bacteria are likely to grow in a closed environment. In winter, the weather is cold, and opening windows for ventilation will harm physical health. Due to the decline in the mobility of the elderly, the bedroom becomes the main activity place, and strengthening indoor ventilation is an important means to improve indoor comfort.

[0096] Oxygen content: Under normal circumstances, under normal sea-level conditions, the oxygen concentration in the air is about 21%. This is the environment that most people are adapted to in daily life. Suitable range: Although there are no strict regulations on the specific oxygen content in nursing homes, it is generally considered ideal to keep the oxygen content in the air close to the natural level (i.e., about 21%). If the oxygen content is too low, it may cause symptoms of hypoxia; while too high may increase the risk of oxygen poisoning. In high-altitude areas, due to the lower atmospheric pressure, the partial pressure of oxygen in the air will also decrease, which may affect the comfort and health of the occupants.

[0097] In this embodiment, the indoor and outdoor air environment information at least includes indoor and outdoor temperature, wind speed, humidity, oxygen concentration, carbon dioxide concentration, and PM2.5 concentration; the energy consumption data includes the operation data of the winter heating device, summer cooling device, humidity adjustment device, ventilation and air change adjustment device, and oxygen supply device; the dynamic influencing factors include the monitored room traffic flow and the physical indicators of the elderly such as body temperature and heart rate.

[0098] In this embodiment, step S4 specifically includes:

[0099] Based on the multi-index demand prediction model of the air environment in each room of the nursing home at different times, obtain the predicted values of the multi-index demand of the air environment in each room at different times, and combine the portrait label data obtained based on the portrait model of each room, the operation data of the corresponding energy supply devices equipped in each room, the performance change curves of the energy supply devices under different operating conditions, and the regulation data of the energy supply devices as the sample data set.

[0100] Taking the predicted values of multi-index air environment requirements for each room in the nursing home at different time periods as the goal, the sample data set is input into a machine learning algorithm for training and learning to establish a regulation model for the energy supply devices in each room at different time periods, and a regulation strategy for the energy supply devices is obtained, including the regulation parameters of the execution components of the winter heating device, summer cooling device, humidity regulation device, ventilation and air change regulation device, and oxygen supply device in each room at different time periods.

[0101] In actual applications, the execution components of the winter heating device include boilers or heat pump controllers, electric valves, water pumps, etc. Or when the winter heating device or summer cooling device is an air conditioner, the execution components are air conditioner unit controllers, frequency converters, expansion valves, etc. The execution components of the ventilation and air change regulation device include fans, fresh air units, exhaust valves, and intake valves. The execution components of the oxygen supply device include the controllers of air purifiers and air fresheners.

[0102] The deep reinforcement learning algorithm is used to train the regulation model of the energy supply devices in each room at different time periods. The specific implementation process is as follows:

[0103] 1) Define the problem framework

[0104] Environment: Each room in the nursing home and its supporting energy supply devices;

[0105] Agent: A software or hardware entity responsible for selecting the optimal regulation strategy according to the current state;

[0106] State: Includes predicted values of multi-index air environment requirements, portrait label data, energy supply device operation data, performance change curves, etc.;

[0107] Action: Corresponding to the specific regulation parameter settings of the execution components of the energy supply devices;

[0108] Reward: Set one or more objective functions, such as comfort index, energy consumption cost, air quality score, etc., to evaluate the effect after taking specific actions;

[0109] 2) Model design

[0110] Select the DRL algorithm, which is suitable for dealing with complex decision-making problems in a continuous action space;

[0111] Define the network structure: Construct a deep neural network as a value function approximator or a policy function approximator, with the state vector as the input and the action probability distribution or Q value as the output;

[0112] 3) Environment simulation and interaction

[0113] Build a simulation platform: Use the digital twin model of the nursing home energy system for simulation, allowing the agent to learn from trial and error in a virtual environment;

[0114] Iterative training: Use the digital twin model of the nursing home energy system to repeatedly run the simulated intelligent agent, explore the best solution by trying different control strategies, and adjust the behavior of the intelligent agent based on the results of each attempt, gradually optimizing its decision-making ability;

[0115] 4) Reinforcement learning training process

[0116] Initialization parameters: set initial strategy, value function weight, learning rate and other hyperparameters;

[0117] Sampling experience replay: Store the historical interaction records of the intelligent agent, randomly extract samples for offline training, break the correlation between data, and improve generalization ability;

[0118] Exploration and utilization: Balance the agent's exploration of unknown areas and the application of existing knowledge to avoid falling into local optimal solutions;

[0119] Update rule: Use gradient descent or other optimization algorithms to minimize the loss function and continuously update network parameters so that the agent gradually learns how to maximize the cumulative reward;

[0120] 5) Testing and verification

[0121] The trained time-sharing energy supply device control model for each room is deployed in the actual nursing home environment. Its performance is observed and compared with historical data to evaluate the effectiveness and stability of the model.

[0122] In this embodiment, step S5 specifically includes:

[0123] After simulating and deducing the control strategy of the energy supply device based on the digital twin model of the nursing home energy system, it is verified and evaluated whether the multiple indicators of the air environment in each room in different time periods meet the expected goals. If so, the strategy is issued and executed; otherwise, the control strategy of the energy supply device in each room in different time periods is readjusted.

[0124] In several embodiments provided in this application, it should be understood that the disclosed systems and methods can also be implemented in other ways. The system embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of systems, methods, and computer program products according to multiple embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0125] In addition, each functional module in various embodiments of the present invention may be integrated together to form an independent part, or each module may exist alone, or two or more modules may be integrated to form an independent part. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories, random access memories, magnetic disks, or optical discs that can store program codes.

[0126] Taking the above ideal embodiments of the present invention as an inspiration, through the above description, relevant workers can completely make various changes and modifications without departing from the technical idea of this invention. The technical scope of this invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A personalized and precise energy management method for nursing homes based on a portrait model, characterized in that: It includes: Step S1: Divide the nursing home into bedrooms and functional rooms for the elderly, and equip them with corresponding energy supply devices, and establish a digital twin model of the nursing home energy system using mechanism modeling and data identification methods; Step S2: Collect multi-source data of each room in the nursing home and construct a portrait model of each room; Step S3: Based on the digital twin model of the nursing home energy system and the portrait models of each room, combined with the monitored indoor and outdoor air environment information and dynamic influencing factors, a multi-index demand prediction model for the air environment of each room in the nursing home in different time periods is established; Step S4: Based on the multi-index demand prediction model of the air environment in each room of the nursing home in different time periods and the portrait model of each room, combined with the operating characteristics of the corresponding energy supply devices in each room, establish an energy supply device control model for each room in different time periods to obtain a control strategy for the energy supply device; Step S5: Verify the control strategy of the energy supply device based on the digital twin model of the nursing home energy system and issue and execute the strategy.

2. The personalized and precise energy management method for nursing homes according to claim 1 is characterized in that: In step S1, the nursing home is divided into bedrooms for the elderly and functional rooms, and equipped with corresponding energy supply devices, including: According to the actual distribution of facilities in the nursing home, it is divided into elderly bedrooms and functional rooms; the elderly bedrooms include single bedrooms, double bedrooms, and triple bedrooms; the functional rooms include dining rooms, activity rooms, rehabilitation rooms, medical rooms, and conference rooms; Each room is equipped with energy supply devices, including at least winter heating devices, summer cooling devices, humidity control devices, ventilation control devices and oxygen supply devices, which are used to regulate the air environment indicators of each room.

3. The personalized and precise energy management method for nursing homes according to claim 2 is characterized in that: In step S1, a digital twin model of the nursing home energy system is established by using a mechanism modeling and data identification method, specifically including: According to the actual distribution of facilities in each room of the nursing home and the supporting energy supply devices, determine the physical entity of the nursing home energy system, including the room energy physical entity and the energy supply physical entity; Establish the geometric model, physical model, behavioral model and rule model of the physical entity of the nursing home energy system, and use virtual reality technology to integrate, merge and check the consistency of the model, and then build a virtual entity mechanism model that is highly consistent with the actual physical entity of the nursing home energy system; Among them, the geometric model is used to describe the spatial position, size and shape characteristics of each room in the nursing home and the supporting energy supply devices; the physical model is used to add the physical properties, constraints and characteristics of the physical entities of the nursing home energy system on the basis of the geometric model; the behavioral model is used to describe the behavior and behavioral evolution of the physical entities of the nursing home energy system under different internal and external factors and mechanisms; the rule model is used to describe the principles, knowledge and experience of the physical entities of the nursing home energy system in actual scene applications; Taking the virtual entity mechanism model as the main body, the residual of the virtual entity mechanism model is calculated using historical data. The input of the virtual entity mechanism model and the model residual result are preprocessed and used as the input and output of the machine learning network respectively. After training and learning, the digital twin data-driven model is obtained, and the residual of the virtual entity mechanism model is compensated and corrected. The revised virtual entity mechanism model and the digital twin data-driven model are integrated to establish a digital twin model of the nursing home energy system.

4. The personalized and precise energy management method for nursing homes according to claim 1 is characterized in that: The step S2 specifically includes: The data support layer collects multi-source data of the bedrooms and functional rooms of the elderly in the nursing home, including static physical data and dynamic behavior data; The data processing layer performs standardization, outlier processing, dimension reduction, null value filling and duplicate value removal on the multi-source data collected from each room in the nursing home, generating pre-processed multi-source data from each room in the nursing home. Through the label extraction layer, a deep learning method is used to extract different attribute features from the pre-processed multi-source data of each room in the nursing home, forming a multi-dimensional label for each room portrait, including at least the static attribute label, health attribute label, work and rest behavior label and energy use behavior label of the elderly bedroom, as well as the static attribute label, activity behavior label and energy use behavior label of the functional room; The portrait modeling layer uses the multi-dimensional labels of each room portrait to build a portrait model for each room.

5. The personalized and precise energy management method for nursing homes according to claim 4 is characterized in that: The static physical data of the elderly bedroom include: the bedroom type, room area, room space layout, the elderly age, gender, physical function, and medical history; The dynamic behavior data of the elderly bedroom includes: the elderly's energy usage habits, daily routines in the elderly's room, activity habits outside the elderly's room, and operation information of energy supply equipment; The static physical data of the functional room includes: room function type, room area, and room space layout; The dynamic behavior data of the functional rooms include: the number of times and duration of use of the rooms in different time periods, personnel flow behavior, personnel participation, comfort evaluation information of each room, energy consumption change trend of each room, and operation information of energy supply equipment.

6. The personalized and precise energy management method for nursing homes according to claim 4 is characterized in that: After constructing the portrait models of each room, it also includes: setting an update cycle, regularly re-collecting multi-source data of each room in the nursing home to update the portrait models of each room, and actively updating the portrait models of each room when it is monitored that the static physical data of the bedrooms and functional rooms of the elderly in the nursing home have changed, or the dynamic behavior data changes exceed a preset range.

7. The personalized and precise energy management method for nursing homes according to claim 1 is characterized in that: The step S3 specifically includes: Based on the digital twin model of the nursing home energy system, the indoor and outdoor air environment information, energy consumption data and dynamic influencing factors monitored in each room at different times are obtained; The indoor and outdoor air environment information, energy consumption data, dynamic influencing factors monitored in each room at different time periods, and the portrait label data reflecting the characteristics of each room based on the portrait model of each room are used as the input data of the air environment multi-index demand prediction model, and the input data is divided into input data of multiple time intervals; Taking room comfort as the goal and multiple air environment indicators that conform to the portrait model of each room as the output data, important data features that affect the multiple air environment indicators of each room are extracted and input into the machine learning algorithm for training and learning, and a multi-indicator demand prediction model for the air environment of each room in the nursing home for different time periods is established, and the multi-indicator demand prediction model for the air environment of each room for different time periods is output; the multiple air environment indicators include winter heat demand index, summer cooling demand index, humidity demand index, fresh air volume demand index and oxygen content demand index.

8. The personalized and precise energy management method for nursing homes according to claim 7 is characterized in that: The indoor and outdoor air environment information includes at least indoor and outdoor temperature, wind speed, humidity, oxygen concentration, carbon dioxide concentration, and PM2.5 concentration; the energy consumption data includes the operating data of winter heating devices, summer cooling devices, humidity control devices, ventilation and air exchange control devices, and oxygen supply devices; the dynamic influencing factors include the monitored room flow and the elderly's body temperature, heart rate and physical indicators.

9. The personalized and precise energy management method for nursing homes according to claim 1 is characterized in that: The step S4 specifically includes: Based on the multi-index demand prediction model for the air environment of each room in the nursing home in different time periods, the multi-index demand prediction value for the air environment of each room in different time periods is obtained, and combined with the portrait label data obtained based on the portrait model of each room, the operating data of the corresponding energy supply device in each room, the performance change curve of the energy supply device under different operating conditions and the regulation data of the energy supply device, they are used as the sample data set; With the goal of meeting the multi-indicator demand prediction values ​​of the air environment in each room of the nursing home in different time periods, the sample data set is input into the machine learning algorithm for training and learning, and the energy supply device control model for each room in different time periods is established. The control strategy of the energy supply device is obtained, including the control parameters of the actuator components of the winter heating device, summer cooling device, humidity control device, ventilation control device and oxygen supply device in each room in different time periods.

10. The personalized and precise energy management method for nursing homes according to claim 1 is characterized in that: The step S5 specifically includes: After simulating and deducing the control strategy of the energy supply device based on the digital twin model of the nursing home energy system, it is verified and evaluated whether the multiple indicators of the air environment in each room in different time periods meet the expected goals. If so, the strategy is issued and executed; otherwise, the control strategy of the energy supply device in each room in different time periods is readjusted.

Images