CN114383295B - Ventilation control method based on indoor pollution concentration estimation model - Google Patents
Ventilation control method based on indoor pollution concentration estimation model Download PDFInfo
- Publication number
- CN114383295B CN114383295B CN202111535815.6A CN202111535815A CN114383295B CN 114383295 B CN114383295 B CN 114383295B CN 202111535815 A CN202111535815 A CN 202111535815A CN 114383295 B CN114383295 B CN 114383295B
- Authority
- CN
- China
- Prior art keywords
- indoor
- gate
- output
- network
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F11/00—Control or safety arrangements
- F24F11/0001—Control or safety arrangements for ventilation
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F11/00—Control or safety arrangements
- F24F11/62—Control or safety arrangements characterised by the type of control or by internal processing, e.g. using fuzzy logic, adaptive control or estimation of values
- F24F11/63—Electronic processing
- F24F11/64—Electronic processing using pre-stored data
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F11/00—Control or safety arrangements
- F24F11/70—Control systems characterised by their outputs; Constructional details thereof
- F24F11/72—Control systems characterised by their outputs; Constructional details thereof for controlling the supply of treated air, e.g. its pressure
- F24F11/74—Control systems characterised by their outputs; Constructional details thereof for controlling the supply of treated air, e.g. its pressure for controlling air flow rate or air velocity
- F24F11/77—Control systems characterised by their outputs; Constructional details thereof for controlling the supply of treated air, e.g. its pressure for controlling air flow rate or air velocity by controlling the speed of ventilators
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F11/00—Control or safety arrangements
- F24F11/70—Control systems characterised by their outputs; Constructional details thereof
- F24F11/72—Control systems characterised by their outputs; Constructional details thereof for controlling the supply of treated air, e.g. its pressure
- F24F11/79—Control systems characterised by their outputs; Constructional details thereof for controlling the supply of treated air, e.g. its pressure for controlling the direction of the supplied air
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F2110/00—Control inputs relating to air properties
- F24F2110/30—Velocity
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F2110/00—Control inputs relating to air properties
- F24F2110/50—Air quality properties
- F24F2110/64—Airborne particle content
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F2110/00—Control inputs relating to air properties
- F24F2110/50—Air quality properties
- F24F2110/65—Concentration of specific substances or contaminants
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F2110/00—Control inputs relating to air properties
- F24F2110/50—Air quality properties
- F24F2110/65—Concentration of specific substances or contaminants
- F24F2110/70—Carbon dioxide
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F24—HEATING; RANGES; VENTILATING
- F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
- F24F2130/00—Control inputs relating to environmental factors not covered by group F24F2110/00
- F24F2130/10—Weather information or forecasts
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
- Y02B30/00—Energy efficient heating, ventilation or air conditioning [HVAC]
- Y02B30/70—Efficient control or regulation technologies, e.g. for control of refrigerant flow, motor or heating
Abstract
The invention relates to a ventilation control method based on an indoor pollution concentration estimation model. The method comprises the following steps: step 1, building an indoor air detection system; step 2, PM2.5 and CO collected in the step 1 are collected 2 Carrying out normalization processing on smoke signal data; step 3, establishing an indoor pollution concentration estimation model; step 4, establishing an indoor ventilation control algorithm model; step 5, embedding the designed indoor pollution concentration estimation model and the indoor ventilation control algorithm model into an air detection data processing module to complete the control of indoor ventilation; and 6, the indoor air detection data processing module detects outdoor wind speed and raindrop value in real time, and when the defined threshold value is exceeded, the controller closes the ventilation device. The invention establishes indoor PM2.5 and CO 2 Based on the dynamic estimation model of concentration and smoke signals, PM2.5 and CO are dynamically estimated according to the rewarding value of ventilation control quantity 2 The control of the window opening and the fan is completed by the concentration and smoke signals, so that the energy consumption can be saved and the service life of equipment can be prolonged compared with the traditional mechanical ventilation system.
Description
Technical Field
The invention relates to the field of indoor pollution, in particular to a ventilation control method based on an indoor pollution concentration estimation model.
Background
With the development of modern technology and the improvement of life quality requirements of people, the functions of modern buildings are more and more perfect, and more human activities can be transferred from outdoor to indoor. Research shows that modern people spend most of their time in building rooms, so that the quality of the air in the rooms has a very important influence on the health and comfort of the human body. In recent years, pollution problems of indoor air around the world are frequent, and indoor air quality problems are also becoming more and more important.
Indoor air pollutants are mainly classified into solid and gas. The solid pollutants mainly comprise suspended particles, dust, pollen of partial plants, cells of microorganisms, partial smoke and the like; the gas pollutants mainly comprise Volatile Organic Compounds (VOCs), ozone (O3), nitrogen oxides (NOx), sulfur dioxide (SO 2), radon gas and the like. Prolonged exposure of the human body to these indoor air pollutants can suffer from serious respiratory diseases, cardiovascular diseases, skin diseases, cancers, etc., and even die. The invention provides a ventilation control method based on an indoor pollution concentration estimation model, which is used for establishing an indoor PM2.5, CO2 and smoke concentration dynamic estimation model and controlling ventilation equipment according to the dynamic estimation model.
Disclosure of Invention
In order to solve the problems, the invention collects PM2.5 and CO in the room 2 Based on smoke concentration data and outdoor raindrops and wind speed data, a ventilation control method based on an indoor pollution concentration estimation model is provided, and indoor PM2.5 and CO are established 2 Based on the dynamic estimation model of concentration and smoke signals, PM2.5 and CO are dynamically estimated according to the rewarding value of ventilation control quantity 2 The concentration and smoke signals are used for controlling the opening of a window and a fan, and the specific steps of the invention are as follows, and the invention is characterized in that:
step 1, building an indoor air detection system: the indoor air detection system is composed of PM2.5 detection sensor and CO 2 The device comprises a detection sensor, a smoke detection sensor, a wind speed detection sensor, a raindrop sensor, a signal conditioning circuit module, a power supply circuit module and an NI acquisition card;
step 2, PM2.5 and CO collected in the step 1 are collected 2 Normalizing the smoke signal data to obtain PM2.5 and CO 2 Normalizing the smoke signal data to a range of 0 to 1 to reduce dimensional effects between the data;
step 3, establishing an indoor pollution concentration estimation model: PM2.5 and CO after normalization in the step 2 2 The smoke data and the output value of the controller to the actuating mechanism are used as input, and the indoor PM2.5 and CO at the next moment are used 2 The smoke data is used as output to establish an indoor pollution concentration estimation model;
step 4, building an indoor ventilation control algorithm model: indoor PM2.5 and CO at the next moment calculated by wind speed, ventilation action and indoor pollution concentration estimation model 2 Smoke data and current indoor PM2.5, CO 2 The smoke data is used as an input training depth deterministic strategy gradient network, and the depth deterministic strategy gradient network is utilized to output ventilation values;
step 5, embedding the designed indoor pollution concentration estimation model and the indoor ventilation control algorithm model into an air detection data processing module to complete the control of indoor ventilation;
and 6, the indoor air detection data processing module detects outdoor wind speed and raindrop value in real time, and when the defined threshold value is exceeded, the controller closes the ventilation device.
Further, the process of building the indoor air detection system in step 1 may be represented as follows:
PM2.5 detection sensor and CO are respectively detected by a power supply circuit 2 Detection sensor, smoke detection sensor, wind speed detection sensor, and raindropThe sensor is powered and installed in proper position to detect PM2.5 and CO 2 And the smoke, wind speed and raindrop signals are converted by the conditioning circuit and then are uniformly connected to the NI acquisition card, and the NI acquisition card transmits the signals to the indoor air detection data processing module of the upper computer through the PCIE interface.
Further, the process of establishing the indoor pollution concentration estimation model in step 3 may be expressed as follows:
step 3.1: the indoor pollution concentration estimation model consists of an input door, a forgetting door, a cell state and an output door; firstly, an input door selectively memorizes input, and a forget door selectively forgets input transmitted by a cell state; secondly, the outputs of the input gate and the forgetting gate are overlapped to the next cell state; finally, the output gate scales the output of the cell state and outputs the predicted value of the next moment;
step 3.2: input door for constructing indoor pollution concentration estimation model and filtering PM2.5 and CO 2 The smoke data and the invalid information of the output value of the controller to the actuating mechanism, and inputting valid information to the cell state
Entry is indicated as follows:
i t =g(W xi x t +W hi h t-1 +W ci c t-1 +b i +W ni n t (1)
wherein x is t Is normalized PM2.5 and CO 2 Smoke data and data of output value of controller to actuator, g () is activation function, W xi Is an input gate and x t Weight coefficient between h t-1 Is a hidden state at t-1, W hi Is the weight coefficient between the input gate and the hidden state, c t-1 Is the cell state at t-1, W ci Is the weight coefficient between the input gate sample and the cell state, b i Is the bias term of the input gate, i t To input gate sample, n t Is the noise interference at t, W ni The weight coefficient between the input gate and the noise interference is used for improving the robustness of the model by adding the noise interference;
step 3.3: and constructing a forgetting gate of an indoor pollution concentration estimation model to delete part of invalid information in the cell state, wherein the forgetting gate is expressed as follows:
f t =g(W xf x t +W hf h t-1 +W cf c t-1 +b f ) (2)
wherein f t For forgetting gate output, W xf Is forgetting door and x t Weight coefficient of each other, W hf Is the weight coefficient between the forget gate and the hidden state, W cf For forgetting door sample and cell state c t-1 Weight coefficient of b f Is forgetting
A bias term for the gate;
step 3.3: constructing a model cell state:
c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c ) (3)
in which W is xc Is the cell state and x t Weight coefficient of the two, W hc Is cell state and hidden state h t-1 Weight coefficient of b c A bias term for the state of the cell;
step 3.4: building a model output gate to output indoor PM2.5 and CO at the next moment 2 Smoke data:
s t =g(W xs x t +W hs h t-1 +W cs c t +b s ) (4)
wherein s is t PM2.5 and CO in the room at the next moment of output of the output gate 2 Normalized data of smoke, W xs Is an output gate and x t Weight coefficient of each other, W hs Is to output the door sample and h t-1 Weight coefficient of the two, W cs Is the output gate sample, c t Weight coefficient between b s Bias item for output gate
Step 3.5: hidden state h of t moment through output gate t Updating:
h t =s t tanh(c t ) (5)
and 3.6, repeating the steps 3.2-3.5, and updating the weight and the bias parameters in the model by utilizing an SGD algorithm until the loss function reaches a convergence threshold value, wherein the convergence threshold value is set to be 1e-4.
Further, the process of building the indoor ventilation control algorithm model in step 4 may be represented as follows:
step 4.1, the indoor ventilation control algorithm model comprises a strategy network and a value network, wherein the value network aims at selecting an optimal action value, the value network aims at minimizing a loss function to update network parameters, the strategy network aims at selecting a current action according to a current state, and generating a next moment action according to an environment state;
step 4.2, respectively initializing a strategy network and a value network, wherein the strategy network and the value network respectively comprise a real network and a target network, and randomly initializing parameters of the strategy network and the value network: θ Q 、θ μ And the following steps:
θ Q′ =θ Q (6)
θ μ′ =θμ (7)
in θ Q Representing real value network parameters, theta Q′ Representing the target value network parameter, θ μ Representing realistic policy network parameters, θ μ′ Representing target policy network parameters
Step 4.3, selecting actuator action a under the current state t :
a t =μ(s t |θ μ )+N t (8)
Wherein N is t Is random noise, and the function mu () is the optimal behavior strategy;
step 4.4, by performing action a t Obtain the prize r t+1 And a state s estimated by an indoor pollution concentration estimation model t+1 Save the current state s t Action a t Prize r t+1 Sum state s t+1 And placing the model into an experience pool; ,
step 4.5, when the empirical pool data reaches the training conditions, randomly selecting N samples {(s) from the empirical pool t ,a t ,r t+1 ,s t+1 )};
Step 4.6, strategy network and value network parameters θ are measured using the samples in the experience pool Q 、θ μ Updating the target action value y calculated through the value network:
y=r t+1 +γQ(s t+1 ,a t+1 /θ Q′ ) (9)
gamma is the discount factor, a t+1 Is an action at time t+1;
step 4.7, updating the target network once at intervals:
θ Q′ =τθ Q +(1-τ)θ Q ′ (10)
θ μ′ =τθ μ +(1-τ)θ μ′ (11)
wherein τ is a learning rate;
step 4.8, returning to the step 4.3 to carry out loop iteration, and knowing the maximum round number;
step 4.9, outputting ventilation control quantity a according to the indoor ventilation control algorithm model after training is completed t 。
The ventilation control method based on the indoor pollution concentration estimation model has the beneficial effects that:
according to the invention, on the basis of establishing a dynamic estimation model of indoor PM2.5, CO2 concentration and smoke signals, the control of the window opening and the fan is finished according to the rewarding value of the ventilation control quantity and the dynamic estimation of the PM2.5, CO2 concentration and smoke signals, and compared with the traditional mechanical ventilation system, the energy consumption can be saved and the service life of equipment can be prolonged;
according to the invention, the outdoor wind speed and the raindrop value are detected in real time, and when the air speed and the raindrop value exceed the limiting threshold value, the controller turns off the ventilation device, so that the safe operation of the ventilation system is ensured;
in order to increase the self-adaption and anti-interference capacity of the ventilation system, the invention designs a novel ventilation control method based on an indoor concentration estimation model, and the system executor is controlled by utilizing the online weight learning capacity, so that the uncertainty and the interference of the ventilation system can be automatically controlled;
the invention provides an important technical means for an indoor pollution ventilation control method.
Drawings
FIG. 1 is a flow chart of the present invention;
fig. 2 is a model diagram of an indoor ventilation control algorithm of the present invention.
Detailed Description
The invention provides a ventilation control method based on an indoor pollution concentration estimation model, aims to improve the running stability and service life of ventilation control equipment while improving the indoor air quality, and provides a principle and a prediction framework based on an indoor pollution concentration estimation algorithm and a ventilation control algorithm. FIG. 1 is a flow chart of the present invention, and the present invention is further described below with reference to the drawings and the detailed description:
step 1, building an indoor air detection system: the indoor air detection system is composed of PM2.5 detection sensor and CO 2 The device comprises a detection sensor, a smoke detection sensor, a wind speed detection sensor, a raindrop sensor, a signal conditioning circuit module, a power supply circuit module and an NI acquisition card;
PM2.5 detection sensor and CO are respectively detected by a power supply circuit 2 The detection sensor, the smoke detection sensor, the wind speed detection sensor and the raindrop sensor are powered, and meanwhile, the sensors are installed at proper positions to detect PM2.5 and CO 2 And the smoke, wind speed and raindrop signals are converted by the conditioning circuit and then are uniformly connected to the NI acquisition card, and the NI acquisition card transmits the signals to the indoor air detection data processing module of the upper computer through the PCIE interface.
Step 2, PM2.5 and CO collected in the step 1 are collected 2 Normalizing the smoke signal data to obtain PM2.5 and CO 2 Normalizing the smoke signal data to a range of 0 to 1 to reduce dimensional effects between the data;
step 3, establishing an indoor pollution concentration estimation model: PM2.5 and CO after normalization in the step 2 2 Pair of smoke data and controllerThe constructed output value is used as input to make indoor PM2.5 and CO at the next moment 2 The smoke data is used as output to establish an indoor pollution concentration estimation model;
step 3.1: the indoor pollution concentration estimation model consists of an input door, a forgetting door, a cell state and an output door; firstly, an input door selectively memorizes input, and a forget door selectively forgets input transmitted by a cell state; secondly, the outputs of the input gate and the forgetting gate are overlapped to the next cell state; finally, the output gate scales the output of the cell state and outputs the predicted value of the next moment;
step 3.2: input door for constructing indoor pollution concentration estimation model and filtering PM2.5 and CO 2 And the smoke data and the invalid information of the output value of the controller to the execution mechanism, and inputting valid information to the cell state, wherein the input gate is expressed as follows:
i t =gW xi x t +W hi h t-1 +W ci c t-1 +b i +W ni n t (I)
wherein x is t Is normalized PM2.5 and CO 2 Smoke data and data of output value of controller to actuator, g () is activation function, W xi Is an input gate and x t Weight coefficient between h t-1 Is a hidden state at t-1, W hi Is the weight coefficient between the input gate and the hidden state, c t-1 Is the cell state at t-1, W ci Is the weight coefficient between the input gate sample and the cell state, b i Is the bias term of the input gate, i t To input gate sample, n t Is the noise interference at t, W ni The weight coefficient between the input gate and the noise interference is used for improving the robustness of the model by adding the noise interference;
step 3.3: and constructing a forgetting gate of an indoor pollution concentration estimation model to delete part of invalid information in the cell state, wherein the forgetting gate is expressed as follows:
f t =g(W xf x t +W hf h t-1 +W cf c t-1 +b f ) (2)
wherein f t For forgetting gate output, W xf Is forgetting door and x t Weight coefficient of each other, W hf Is the weight coefficient between the forget gate and the hidden state, W cf For forgetting door sample and cell state c t-1 Weight coefficient of b f Is an offset item of the forgetting door;
step 3.3: constructing a model cell state:
c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c ) (3)
in which W is xc Is the cell state and x t Weight coefficient of the two, W hc Is cell state and hidden state h t-1 Weight coefficient of b c A bias term for the state of the cell;
step 3.4: building a model output gate to output indoor PM2.5 and CO at the next moment 2 Smoke data:
s t =g(W xs x t +W hs h t-1 +W cs c t +b s ) (4)
wherein s is t PM2.5 and CO in the room at the next moment of output of the output gate 2 Normalized data of smoke, W xs Is an output gate and x t Weight coefficient of each other, W hs Is to output the door sample and h t-1 Weight coefficient of the two, W cs Is the output gate sample, c t Weight coefficient between b s Bias item for output gate
Step 3.5: hidden state h of t moment through output gate t Updating:
h t =s t tanh(c t ) (5)
and 3.6, repeating the steps 3.2-3.5, and updating the weight and the bias parameters in the model by utilizing an SGD algorithm until the loss function reaches a convergence threshold value, wherein the convergence threshold value is set to be 1e-4.
Step 4, building an indoor ventilation control algorithm model: wind speed and ventilationIndoor PM2.5 and CO at next moment calculated by action and indoor pollution concentration estimation model 2 Smoke data and current indoor PM2.5, CO 2 The smoke data is used as an input training depth deterministic strategy gradient network, and the depth deterministic strategy gradient network is utilized to output ventilation values;
the indoor ventilation control algorithm model comprises a strategy network and a value network, wherein the value network aims at selecting the optimal action value, the value network aims at minimizing a loss function to update network parameters, the strategy network aims at selecting the current action according to the current state and generating the action at the next moment according to the environmental state, and the indoor ventilation control algorithm model diagram is shown in figure 2;
step 4.1, respectively initializing a strategy network and a value network, wherein the strategy network and the value network respectively comprise a real network and a target network, and randomly initializing parameters of the strategy network and the value network: θ Q 、θ μ And the following steps:
θ Q′ =θ Q (6)
θ μ′ =θμ (7)
in θ Q Representing real value network parameters, theta Q′ Representing the target value network parameter, θ μ Representing realistic policy network parameters, θ μ′ Representing target policy network parameters
Step 4.2, selecting actuator action a under the current state t :
a t =μ(s t |θ μ )+N t (8)
Wherein N is t Is random noise, and the function mu () is the optimal behavior strategy;
step 4.3 by performing action a t Obtain the prize r t+1 And a state s estimated by an indoor pollution concentration estimation model t+1 Save the current state s t Action a t Prize r t+1 Sum state s t+1 And placing the model into an experience pool; ,
step 4.4, when the experience pool data reaches the training conditionAt this time, N samples {(s) are randomly selected from the experience pool t ,a t ,r t+1 ,s t+1 )};
Step 4.5, strategy network and value network parameters θ are measured using the samples in the experience pool Q 、θ μ Updating the target action value y calculated through the value network:
=r t+1 +γQ(s t+1 ,a t+1 /θ Q′ ) (9)
gamma is the discount factor, a t+1 Is an action at time t+1;
step 4.6, updating the target network once at intervals:
θ Q′ =τθ Q +(1-τ)θ Q′ (10)
θ μ′ =τθ μ +(1-τ)θ μ′ (11)
wherein τ is a learning rate;
step 4.7, returning to the step 4.2 to carry out loop iteration, and knowing the maximum round number;
step 4.8, outputting ventilation control quantity a according to the indoor ventilation control algorithm model after training t 。
Step 5, embedding the designed indoor pollution concentration estimation model and the indoor ventilation control algorithm model into an air detection data processing module to complete the control of indoor ventilation;
and 6, the indoor air detection data processing module detects outdoor wind speed and raindrop value in real time, and when the defined threshold value is exceeded, the controller closes the ventilation device.
The above description is only of the preferred embodiment of the present invention, and is not intended to limit the present invention in any other way, but is intended to cover any modifications or equivalent variations according to the technical spirit of the present invention, which fall within the scope of the present invention as defined by the appended claims.
Claims (1)
1. The ventilation control method based on the indoor pollution concentration estimation model comprises the following specific steps of:
step 1, lappingBuilding an indoor air detection system: the indoor air detection system is composed of PM2.5 detection sensor and CO 2 The device comprises a detection sensor, a smoke detection sensor, a wind speed detection sensor, a raindrop sensor, a signal conditioning circuit module, a power supply circuit module and an NI acquisition card;
the process of building the indoor air detection system in step 1 can be expressed as follows:
PM2.5 detection sensor and CO are respectively detected by a power supply circuit 2 The detection sensor, the smoke detection sensor, the wind speed detection sensor and the raindrop sensor are powered, and meanwhile, the sensors are installed at proper positions to detect PM2.5 and CO 2 The smoke, wind speed and raindrop signals are converted by the conditioning circuit and then are uniformly connected to the NI acquisition card, and the NI acquisition card transmits the signals to the indoor air detection data processing module of the upper computer through the PCIE interface;
step 2, PM2.5 and CO collected in the step 1 are collected 2 Normalizing the smoke signal data to obtain PM2.5 and CO 2 Normalizing the smoke signal data to a range of 0 to 1 to reduce dimensional effects between the data;
step 3, establishing an indoor pollution concentration estimation model: PM2.5 and CO after normalization in the step 2 2 The smoke data and the output value of the controller to the actuating mechanism are used as input, and the indoor PM2.5 and CO at the next moment are used 2 The smoke data is used as output to establish an indoor pollution concentration estimation model;
the process of establishing the indoor pollution concentration estimation model in the step 3 is represented as follows:
step 3.1: the indoor pollution concentration estimation model consists of an input door, a forgetting door, a cell state and an output door; firstly, an input door selectively memorizes input, and a forget door selectively forgets input transmitted by a cell state; secondly, the outputs of the input gate and the forgetting gate are overlapped to the next cell state; finally, the output gate scales the output of the cell state and outputs the predicted value of the next moment;
step 3.2: input door for constructing indoor pollution concentration estimation model and filtering PM2.5 and CO 2 Pair of smoke data and controllerConstruct invalid information of output value, and input valid information to cell state, input gate is expressed as follows:
i t =g(W xi x t +W hi h t-1 +W ci c t-1 +b i )+W ni n t (1)
wherein x is t Is normalized PM2.5 and CO 2 Smoke data and data of output value of controller to actuator, g () is activation function, W xi Is an input gate and x t Weight coefficient between h t-1 Is a hidden state at t-1, W hi Is the weight coefficient between the input gate and the hidden state, c t-1 Is the cell state at t-1, W ci Is the weight coefficient between the input gate sample and the cell state, b i Is the bias term of the input gate, i t To input gate sample, n t Is the noise interference at t, W ni The weight coefficient between the input gate and the noise interference is used for improving the robustness of the model by adding the noise interference;
step 3.3: and constructing a forgetting gate of an indoor pollution concentration estimation model to delete part of invalid information in the cell state, wherein the forgetting gate is expressed as follows:
f t =g(W xf x t +W hf h t-1 +W cf c t-1 +b f ) (2)
wherein f t For forgetting gate output, W xf Is forgetting door and x t Weight coefficient of each other, W hf Is the weight coefficient between the forget gate and the hidden state, W cf For forgetting door sample and cell state c t-1 Weight coefficient of b f Is an offset item of the forgetting door;
step 3.3: constructing a model cell state:
c t =f t c t-1 +i t tanh(W xc x t +W hc h t-1 +b c ) (3)
in which W is xc Is the cell state and x t Weight coefficient of the two, W hc Is cell state and hidden state h t-1 Weight coefficient of b c A bias term for the state of the cell;
step 3.4: building a model output gate to output indoor PM2.5 and CO at the next moment 2 Smoke data:
s t =g(W xs x t +W hs h t-1 +W cs c t +b s ) (4)
wherein s is t PM2.5 and CO in the room at the next moment of output of the output gate 2 Normalized data of smoke, W xs Is an output gate and x t Weight coefficient of each other, W hs Is to output the door sample and h t-1 Weight coefficient of the two, W cs Is the output gate sample, c t Weight coefficient between b s Bias item for output gate
Step 3.5: hidden state h of t moment through output gate t Updating:
h t =s t tanh(c t ) (5)
step 3.6, repeating the steps 3.2-3.5, and updating the weight and the bias parameters in the model by utilizing an SGD algorithm until the loss function reaches a convergence threshold value, wherein the convergence threshold value is set to be 1e-4;
step 4, building an indoor ventilation control algorithm model: indoor PM2.5 and CO at the next moment calculated by wind speed, ventilation action and indoor pollution concentration estimation model 2 Smoke data and current indoor PM2.5, CO 2 The smoke data is used as an input training depth deterministic strategy gradient network, and the depth deterministic strategy gradient network is utilized to output ventilation values;
the process of building the indoor ventilation control algorithm model in step 4 can be expressed as follows:
step 4.1, respectively initializing a strategy network and a value network, wherein the strategy network and the value network respectively comprise a real network and a target network, and randomly initializing parameters of the strategy network and the value network: θ Q 、θ μ And the following steps:
θ Q′ =θ Q (6)
θ μ′ =θ μ (7)
in θ Q Representing real value network parameters, theta Q′ Representing the target value network parameter, θ μ Representing realistic policy network parameters, θ μ′ Representing target policy network parameters
Step 4.2, selecting actuator action a under the current state t :
a t =μ(s t |θ μ )+N t (8)
Wherein N is t Is random noise, and the function mu () is the optimal behavior strategy;
step 4.3 by performing action a t Obtain the prize r t+1 And a state s estimated by an indoor pollution concentration estimation model t+1 Save the current state s t Action a t Prize r t+1 Sum state s t+1 And placing the model into an experience pool;
step 4.4, when the empirical pool data reaches the training conditions, randomly selecting N samples {(s) from the empirical pool t ,a t ,r t+1 ,s t+1 )};
Step 4.5, strategy network and value network parameters θ are measured using the samples in the experience pool Q 、θ μ Updating the target action value y calculated through the value network:
y=r t+1 +γQ(s t+1 ,a t+1 /θ Q′ ) (9)
gamma is the discount factor, a t+1 Is an action at time t+1;
step 4.6, updating the target network once at intervals:
θ Q′ =τθ Q +(1-τ)θ Q′ (10)
θ μ′ =τθ μ +(1-τ)θ μ′ (11)
wherein τ is a learning rate;
step 4.7, returning to the step 4.2 to carry out loop iteration, and knowing the maximum round number;
step 4.8, modeling according to the indoor ventilation control algorithm after trainingOutput ventilation control amount a t ;
Step 5, embedding the designed indoor pollution concentration estimation model and the indoor ventilation control algorithm model into an air detection data processing module to complete the control of indoor ventilation;
and 6, the indoor air detection data processing module detects outdoor wind speed and raindrop value in real time, and when the defined threshold value is exceeded, the controller closes the ventilation device.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111535815.6A CN114383295B (en) | 2021-12-15 | 2021-12-15 | Ventilation control method based on indoor pollution concentration estimation model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111535815.6A CN114383295B (en) | 2021-12-15 | 2021-12-15 | Ventilation control method based on indoor pollution concentration estimation model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114383295A CN114383295A (en) | 2022-04-22 |
CN114383295B true CN114383295B (en) | 2023-05-16 |
Family
ID=81198537
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111535815.6A Active CN114383295B (en) | 2021-12-15 | 2021-12-15 | Ventilation control method based on indoor pollution concentration estimation model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114383295B (en) |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014167837A1 (en) * | 2013-04-12 | 2014-10-16 | パナソニック株式会社 | Air-conditioning system and controller |
CN106707999B (en) * | 2017-02-09 | 2023-04-18 | 苏州科技大学 | Building energy-saving system based on adaptive controller, control method and simulation |
CN110348592A (en) * | 2019-05-21 | 2019-10-18 | 华电电力科学研究院有限公司 | A kind of load model prediction technique and forecasting system based on artificial neural network |
WO2021010505A1 (en) * | 2019-07-12 | 2021-01-21 | 엘지전자 주식회사 | Intelligent air purifier, and indoor air quality control method and control device using intelligent air purifier |
KR102264855B1 (en) * | 2019-11-26 | 2021-06-15 | 전남대학교산학협력단 | Method and system for predicting ventilation mode based on artificial neural network |
CN111103220B (en) * | 2019-12-31 | 2021-04-27 | 西安交通大学 | Method and system for predicting and regulating concentration of atmospheric pollutants |
US11892187B2 (en) * | 2020-03-26 | 2024-02-06 | Generac Power Systems, Inc. | Thermostat device with improved energy optimization |
CN112628955B (en) * | 2020-12-23 | 2022-03-01 | 杭州电子科技大学 | Indoor ventilation control method based on LSTM neural network and krill swarm optimization algorithm |
-
2021
- 2021-12-15 CN CN202111535815.6A patent/CN114383295B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN114383295A (en) | 2022-04-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11648504B2 (en) | Air conditioning device and control method thereof | |
CN108426812B (en) | PM2.5 concentration value prediction method based on memory neural network | |
CN110395286B (en) | Method and system for monitoring air quality and regulating ventilation in train | |
Chenari et al. | Towards sustainable, energy-efficient and healthy ventilation strategies in buildings: A review | |
Li et al. | A coupled computational fluid dynamics and back-propagation neural network-based particle swarm optimizer algorithm for predicting and optimizing indoor air quality | |
Dai et al. | An artificial neural network model using outdoor environmental parameters and residential building characteristics for predicting the nighttime natural ventilation effect | |
CN106918113A (en) | The management method and system of a kind of smart home | |
An et al. | A reinforcement learning approach for control of window behavior to reduce indoor PM2. 5 concentrations in naturally ventilated buildings | |
CN111561764A (en) | Control method and device for fresh air system and storage medium | |
CN114383295B (en) | Ventilation control method based on indoor pollution concentration estimation model | |
CN106707999B (en) | Building energy-saving system based on adaptive controller, control method and simulation | |
CN106368551A (en) | Method and device of automatically controlling working state of window | |
Zhuang et al. | Probabilistic occupancy forecasting for risk-aware optimal ventilation through autoencoder Bayesian deep neural networks | |
Majdi et al. | A novel method for Indoor Air Quality Control of Smart Homes using a Machine learning model | |
CN107544375B (en) | Household residence air quality control strategy analysis method | |
CN104102139B (en) | Simulation platform of variable air rate air-conditioner control system | |
CN205533374U (en) | Automatic open control device of bathroom air discharge fan | |
CN112070317A (en) | Hotel air conditioner energy consumption prediction method | |
CN207065822U (en) | A kind of toilet ventilation system | |
CN108036473B (en) | Intelligent temperature and humidity control method and device | |
Rastogi et al. | An IoT-based system to evaluate indoor air pollutants using grey relational analysis | |
TWI699637B (en) | System for controlling environmental comfort using deep reinforcement learning and method thereof | |
CN105571997A (en) | Measurement and control method and device for particle concentration of indoor air | |
CN108549232B (en) | A kind of room air self-adaptation control method based on approximate model planning | |
CN205374235U (en) | Room air particle concentration measuring device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |