Thermal environment regulation system and method based on natural cold source and multi-source information fusion

Abstract

The present application relates to the field of building energy saving and environment control technology, and discloses a thermal environment regulation system and method based on natural cold source and multi-source information fusion; the system comprises an environment perception module, a distributed execution module, a regional collaborative control module and a system optimization module; the method comprises the following steps: multi-source information real-time acquisition and fusion; thermal load evaluation and natural cold source availability judgment; collaborative control decision based on multi-objective optimization; hierarchical precise execution and dynamic collaboration; global strategy optimization and self-learning. The present application realizes thermal environment regulation through intelligent collaborative control, realizes the regulation and control change from passive response to active prediction, effectively solves the problems of response lag and energy waste of traditional systems, and has the remarkable characteristics of high energy efficiency, precise control, rapid response, collaborative control, stability and reliability, self-adaptation and strong generalizability.

Description

Technical Field

[0001] This invention relates to the field of building energy conservation and environmental control technology, and more specifically, to a thermal environment control system and method based on the fusion of natural cold sources and multi-source information. Background Technology

[0002] With the increasing prevalence of high-rise buildings and glass curtain walls in modern architecture, the problem of unbalanced indoor thermal environments has become increasingly prominent. During winter and the transitional seasons of spring and autumn, the sunny sides of buildings (such as the south and west sides) absorb a large amount of heat due to intense solar radiation, leading to severe overheating in some areas, while the shaded areas remain at normal or even below-normal temperatures. Furthermore, alterations to the original floor layout by residents disrupt the original airflow pattern, exacerbating the thermal imbalance within the building. This uneven heating severely impacts indoor comfort and leads to inefficient operation of the central air conditioning system.

[0003] A fundamental limitation of traditional central air conditioning systems is that, during a unified heating or cooling season, the system cannot simultaneously provide differentiated heating or cooling output to areas with different orientations or functions. Furthermore, existing HVAC systems are typically designed based on steady-state, uniform load models, resulting in global and lagging control that struggles to handle instantaneous, dynamic, and spatially unevenly distributed solar radiation heat loads. While improvements exist, such as curtain wall ventilators and localized cooling equipment, these often suffer from problems like inefficient control, interference with the central system, disruption of indoor air pressure balance, high energy consumption, or implementation difficulties.

[0004] Therefore, there is an urgent need in this field for a ventilation control system and method that can be deeply integrated with existing building HVAC systems, intelligently coordinate, precisely regulate the local thermal environment, and is easy to implement and highly energy-efficient. Summary of the Invention

[0005] This invention aims to overcome the shortcomings of existing technologies and provide a thermal environment control system and method based on the fusion of natural cold sources and multi-source information. This system deploys distributed intelligent ventilation terminals and sensor networks and integrates them deeply with the building automation system (BAS) to achieve precise and adaptive ventilation and cooling of local overheated areas of the building, while maintaining overall indoor thermal comfort and stable air pressure, efficiently utilizing outdoor natural cold sources, and reducing the overall energy consumption of the system.

[0006] To achieve the above objectives, the present invention employs the following technical solution.

[0007] In a first aspect, the present invention provides a thermal environment control system based on the fusion of natural cold sources and multi-source information, the system comprising an environmental sensing module, a distributed execution module, a regional collaborative control module, and a system optimization module.

[0008] The environmental sensing module includes a sensor network deployed inside and outside the building to collect real-time environmental data and obtain operating status information from the building's existing heating, ventilation, and air conditioning (HVAC) system; the environmental data includes solar radiation intensity, outdoor meteorological parameters, indoor multi-point temperature and indoor-outdoor air pressure difference;

[0009] The sensor network includes: solar radiation sensors, outdoor temperature and humidity sensors, and wind speed and direction sensors deployed on the building roof; temperature sensors deployed in the controlled areas inside the building; and differential pressure sensors used to measure the air pressure difference between the indoor area and the corridor on the same floor.

[0010] The distributed execution module includes multiple intelligent ventilation terminals deployed at the ventilation openings of the building's external envelope (such as glass curtain walls), used to adjust the air supply volume according to the received air volume control commands;

[0011] The intelligent ventilation terminal includes a fan, a damper, an airflow detection unit, and a local controller (using an embedded microprocessor). The local controller, based on the target airflow command received from the regional collaborative control module, precisely tracks and controls the airflow by adjusting the fan speed and damper opening, and reports its operating status to the regional collaborative control module. The intelligent ventilation terminal is equipped with a wind pressure sensor and a speed decoder for measuring the incoming airflow.

[0012] The regional collaborative control module includes regional coordination controllers installed on building floors. Each regional coordination controller is communicatively connected to all intelligent ventilation terminals and corresponding sensors in the environmental sensing module on its floor, with outdoor sensors shared by all floors. The regional coordination controller communicates with the central controller of the existing HVAC system through the building automation system. Based on the fused information collected by the environmental sensing module, it performs rolling optimization through a model predictive control algorithm to generate and send optimal airflow control instructions (sending valve opening adjustment requests) to the intelligent ventilation terminals, as well as collaborative adjustment instructions with the existing HVAC system (supply air temperature setpoint adjustment suggestions).

[0013] The system optimization module is communicatively connected to the regional collaborative control module and is used to perform global strategy learning based on building schedules, weather forecasts and historical operation data, and to optimize the control parameters in the regional collaborative control module (perform adaptive tuning).

[0014] Secondly, the present invention provides a control method for the above-mentioned thermal environment control system based on the fusion of natural cold sources and multi-source information, comprising the following steps:

[0015] S1: Real-time acquisition and fusion of multi-source information:

[0016] The solar radiation intensity I and outdoor meteorological parameters T are collected synchronously through the environmental sensing module.out Indoor local temperature distribution T zone_i The system collects the indoor and outdoor air pressure difference ΔP and the status information of the existing HVAC system, and performs spatiotemporal registration and feature fusion.

[0017] S2: Heat load assessment and assessment of natural cooling source availability:

[0018] Based on fused information, the regional collaborative control module identifies local overheated areas and their heat load magnitude; at the same time, it determines whether the outdoor air temperature is lower than the indoor overheated area temperature to determine the availability of natural cooling sources.

[0019] S3: Cooperative control decision based on multi-objective optimization:

[0020] When a natural cooling source is available, the regional collaborative control module aims to minimize the total energy consumption of the system, maintain indoor thermal comfort and air pressure stability. Based on the model predictive control algorithm, it uses fused information (especially the feedforward information of solar radiation and the feedback information of indoor temperature) to solve for the optimal target air volume of each intelligent ventilation terminal and generate collaborative action commands with the original HVAC system.

[0021] The determination that a natural cold source is available requires that the following conditions be met simultaneously: the outdoor air temperature is lower than the set threshold for the indoor overheated area temperature; the outdoor air humidity is lower than the set upper limit and there is no risk of condensation; and the outdoor wind speed is lower than the set upper limit.

[0022] The model predictive control algorithm employs a state-space predictive model that integrates solar radiation feedforward characteristics, indoor state feedback characteristics, and system coupling characteristics for rolling optimization.

[0023] S4: Layered Precise Execution and Dynamic Collaboration:

[0024] The distributed execution module sends the optimal target air volume command to each intelligent ventilation terminal for execution (by adjusting the fan speed and the opening of the air valve for precise execution), and dynamically adjusts the indoor air pressure balance. At the same time, it sends the collaborative action command to the original HVAC system to achieve collaborative operation with the building's original HVAC system.

[0025] The dynamic adjustment of indoor air pressure balance is achieved by using an independent air pressure closed-loop control loop and a proportional-integral control algorithm to dynamically adjust the air volume ratio between the air intake and exhaust terminals in order to maintain the set indoor air pressure.

[0026] S5: Global Strategy Optimization and Self-Learning

[0027] The system optimization module performs medium- and long-term strategy optimization based on historical operating data, weather forecasts, and construction schedules, and updates the control parameters through self-learning.

[0028] Furthermore, step S1, real-time acquisition and fusion of multi-source information, specifically includes:

[0029] S11: Synchronous acquisition of multi-source heterogeneous data. Through a distributed sensor network and the building automation system interface, the following five types of heterogeneous data are acquired in real time:

[0030] Radiated disturbance data: ,in Let t be the solar radiation intensity (W / m²) at time t.

[0031] Outdoor meteorological data: These represent outdoor temperature, relative humidity, wind speed, and wind direction, respectively.

[0032] Indoor thermal environment data: ,in Let be the indoor temperature at the i-th monitoring point, and n be the total number of monitoring points. The difference between indoor and outdoor air pressure;

[0033] Status data of the building's original HVAC system : Obtained from the building automation system (via standard protocols such as BACnet), including system operating modes. (Refrigeration, heating, and ventilation), critical damper opening air supply temperature Return air temperature and air volume ;

[0034] S12: Spatiotemporal registration and data preprocessing, performing spatiotemporal alignment and cleaning on the acquired asynchronous and heterogeneous raw data; the first step is time synchronization for all data streams. Inject a unified high-precision timestamp k is the sampling sequence number, forming a synchronous dataset. The second step is spatial mapping, which maps the data from each temperature monitoring point. In relation to its spatial location within the building, a vector representation of the temperature field distribution is constructed. The third step is validity verification, which uses sliding window mean filtering and the Laida criterion to remove outliers and ensure data reliability.

[0035] S13: The feature quantities mentioned in feature-level fusion and key state quantity extraction are:

[0036] Feedforward characteristics based on radiation and meteorology:

[0037] ,

[0038] in, This represents the solar radiation value. The rate of change of radiation, Outdoor temperature, used to predict heat load trends;

[0039] Feedback characteristics based on indoor conditions:

[0040] ,

[0041] in, The degree of thermal imbalance is quantified by the maximum regional temperature difference; The standard deviation of the temperature field; For regional pressure difference, This represents the average predicted vote value based on current temperature and humidity.

[0042] System coupling characteristics based on the existing HVAC system status:

[0043]

[0044] in, The average indoor temperature. It reflects the driving force of the temperature difference between the central air conditioning supply air and the indoor environment.

[0045] Furthermore, the specific process of step S2, heat load assessment and natural cold source availability determination (this step involves thermal state identification and cold source feasibility analysis), is as follows:

[0046] S21: Dynamic heat load identification and overheated region location, based on radiation feedforward characteristics Temperature feedback characteristics Construct a heat load estimation model:

[0047] ,

[0048] in, Heat is gained from solar radiation. The overall heat gain coefficient of the curtain wall, For the area of ​​the radiation-affected curtain wall, For conduction and heat transfer, Let be the heat transfer coefficient of the j-th building envelope. Let j be the area of ​​the j-th enclosure structure. Let j be the temperature of the building envelope. The target area temperature; by comparing the temperatures of different areas With heat dissipation capacity, the location of overheated core areas is denoted as . ;

[0049] S22: Multi-criteria evaluation of the quality and availability of natural cooling sources, establishing access criteria for the use of natural cooling sources, which must simultaneously meet the following:

[0050] Temperature feasibility: ,in Outdoor air temperature This represents the lowest temperature at each monitoring point in the overheated area. The minimum effective temperature difference threshold;

[0051] Humidity safety: and ,in The relative humidity of outdoor air. The maximum permissible outdoor humidity level, Outdoor air temperature This refers to the indoor air dew point temperature. To ensure a safety margin against condensation, and to prevent the introduction of excessively humid air or condensation on the building envelope (such as glass curtain walls);

[0052] Meteorological suitability: ,in Outdoor wind speed, To allow for the maximum wind speed limit, and to prevent strong winds from causing airflow to become uncontrollable or equipment to overload;

[0053] If all conditions are met, the natural cooling source is deemed usable, and its maximum theoretical cooling power is calculated. ,in air density, The specific volume of air at constant pressure. For maximum ventilation, The average temperature of the superheated zone. Outdoor air temperature;

[0054] S23: Controlling triggering and mode selection, defining trigger functions. ,in For indicator functions, This represents the maximum temperature difference indoors. Set a threshold for the temperature difference; when When the system is activated, the intelligent collaborative ventilation mode is triggered; otherwise, regular monitoring or standby mode is maintained.

[0055] Furthermore, S3, based on multi-objective optimization and collaborative control decision-making, includes the following steps:

[0056] S31: Construction of a multi-objective optimization problem, in each control cycle The goal is to minimize the overall cost function and solve for the optimal control vector. ,in The target air volume for each ventilation terminal. To adjust the opening of its blinds. The suggested adjustment amount for the opening of the fresh air system damper in the existing HVAC system; cost function Defined as:

[0057] ;

[0058] in, For the collection of overheated regions, Let be the actual indoor temperature at the i-th monitoring point. For dynamic temperature setpoint, This represents the actual air pressure difference between indoors and outdoors. Set the target air pressure difference between indoors and outdoors. The total energy consumption of all intelligent ventilation terminal fans. This refers to the additional energy consumption of the original HVAC system caused by ventilation adjustment. To predict average voting bias, , , , These represent the weighting coefficients for temperature deviation, air pressure deviation, energy consumption, and thermal comfort deviation, respectively.

[0059] Constraints include equipment constraints: , ;

[0060] in The target air volume for the j-th intelligent ventilation terminal is... , These are the minimum and maximum airflow of a single intelligent ventilation terminal, respectively. This indicates the target opening degree of the air valve of the j-th intelligent ventilation terminal;

[0061] Comfort and safety constraints: ;

[0062] in For air supply speed, For maximum comfortable wind speed, For supply air temperature, This refers to the indoor air dew point temperature. To prevent condensation and ensure a safety margin.

[0063] System coupling constraints:

[0064] ;

[0065] in air density, This represents the total air intake volume for all intelligent ventilation terminals. This refers to the air supply volume of the original HVAC system. This refers to the return air volume of the original HVAC system. This is the instantaneous rate of change of indoor air quality; this formula mandates that the ventilation system and the existing HVAC system coordinate in terms of airflow to maintain quality balance.

[0066] S32: Solving the model predictive control problem based on fused features, using a model predictive control framework for rolling optimization:

[0067] Prediction model: using fusion features extracted from S13 , , The building thermodynamics model is updated online; the model is simplified to state-space form.

[0068] ;

[0069] Among them, state variables , Let the indoor temperature field vector be... The control quantity is the difference between indoor and outdoor air pressure. , The target airflow vector for the intelligent ventilator. As a feedforward input for measurable disturbances; , , These are the state transition matrix, control input matrix, and disturbance input matrix, respectively.

[0070] Scrolling optimization: in each Time, based on the current state and future Perturbation prediction sequence of the step (Based on weather forecasts and radiation trends), solving for the future under constraints. Step control sequence The total cost during the forecast period minimize;

[0071] Feedback correction: Only the first set of control variables in the optimized sequence are adjusted. As the actual instruction output; in the next cycle, using the actual measurement. Correct the model prediction error to achieve closed-loop optimization.

[0072] S33: Cooperative instruction package generation and distribution, which encapsulates the optimization solution results into standardized cooperative instruction packages:

[0073] For intelligent ventilation terminals: issue instructions to , Let the j-th intelligent ventilator be the optimal target airflow. The optimal valve opening for the j-th intelligent ventilator;

[0074] For existing HVAC systems: Generate collaborative requests The message is sent to the building automation system, suggesting adjustments to the damper opening or supply air temperature setting to achieve overall energy savings. The optimal opening degree for the return air valve in the existing HVAC system. The setpoint for the supply air temperature of the existing HVAC system.

[0075] Furthermore, the layered precise execution and dynamic collaboration in S4 includes the following steps:

[0076] S41: Execution layer adaptive flow tracking control. After receiving the command, the intelligent ventilation terminal's local controller adopts a feedforward-feedback composite control strategy. The fan speed control formula and the ventilator damper opening formula are as follows:

[0077] ,

[0078] ,

[0079] in, The feedforward control gain coefficient is... Set the target air supply volume. This is the proportional control gain coefficient. This is the integral gain coefficient. For air volume deviation, This refers to the actual air volume. To determine the static pressure of the air duct The ventilator damper opening compensation function obtained through dynamic table lookup ensures accurate tracking under varying operating conditions. ;

[0080] S42: Rapid closed-loop pressure balance regulation, establishing an independent pressure regulation loop; As a set value, with For feedback, proportional control is used to dynamically adjust the ratio of air intake and exhaust terminal air volume. This ensures that the pressure deviation converges quickly, among which... This is the proportional gain coefficient of the pneumatic circuit. Set the pressure difference value. This is a measurement of the air pressure difference;

[0081] S43: Inter-system data exchange and status synchronization; during execution, the intelligent ventilation system continuously transmits the actual air volume. ,power consumption The status variables are fed back to the building automation system in real time through the system interface, enabling the system to dynamically update the total building load calculation and optimize the operation strategy of the cooling and heating units accordingly, thereby achieving global integrated optimization.

[0082] Furthermore, the S5 global policy optimization and self-learning includes the following steps:

[0083] S51: Run database construction, and run long-term data. Stored in a time-series database to form a training dataset. ;

[0084] S52: Parameter self-tuning based on machine learning, periodically utilizing... Training a deep neural network or reinforcement learning agent with fused features as input. , , The output is an optimized parameter set that minimizes the long-term average cost; the updated parameters will be sent offline to the regional collaborative control module.

[0085] This invention is particularly suitable for solving the problem of localized indoor overheating caused by solar radiation in buildings with large glass curtain walls during winter and transitional seasons, and achieves precise, efficient and energy-saving thermal environment regulation through intelligent collaborative control.

[0086] Compared with the prior art, the present invention has the following significant features:

[0087] (1) High efficiency and energy saving: Directly and actively utilize the abundant outdoor natural cold sources during winter and transitional seasons to precisely cool local overheated areas, which can significantly reduce or even avoid starting high-energy-consuming central refrigeration units, resulting in significant energy-saving benefits.

[0088] (2) Precise control and rapid response: By integrating multi-source information such as solar radiation (feedforward), indoor multi-point temperature (feedback), and air pressure (constraint), the system achieves predictive regulation and rapid response to dynamic heat load, solving the problem of lag in traditional systems.

[0089] (3) System collaboration, stability and reliability: A deep collaboration mechanism with the original HVAC system is proposed. Through air pressure balance closed loop and communication interface, the new system is ensured to work harmoniously with the original system, avoid conflicts, and ensure the overall stability and comfort of the indoor environment.

[0090] (4) Strong adaptability and scalability: The system has self-learning and global optimization capabilities, and can adaptively adjust according to building characteristics and usage patterns. It is suitable for various glass curtain wall buildings with similar thermal imbalance problems. Attached Figure Description

[0091] Figure 1 This is a diagram showing the overall architecture of the thermal environment control system based on the fusion of natural cold sources and multi-source information according to the present invention.

[0092] Figure 2 This is a diagram showing the structure and distribution of the environmental sensing module in this invention.

[0093] Figure 3 This is a flowchart of the intelligent collaborative control method in this invention. Detailed Implementation

[0094] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0095] The thermal environment control system based on the fusion of natural cold sources and multi-source information of the present invention includes an environmental sensing module, a distributed execution module, a regional collaborative control module, and a system optimization module.

[0096] The environmental sensing module includes a sensor network deployed indoors and outdoors to collect real-time environmental data and obtain operational status information from the building's existing HVAC system; the environmental data includes solar radiation intensity, outdoor meteorological parameters, indoor multi-point temperature and indoor-outdoor air pressure difference.

[0097] The sensor network includes at least indoor zone temperature sensors, indoor differential pressure sensors, and outdoor solar radiation sensors, outdoor temperature and humidity sensors, and anemometers. Temperature sensors are deployed in indoor areas where the effects of heat radiation need to be eliminated to monitor the ambient temperature of indoor zones in real time. Differential pressure sensors are deployed both indoors and in the same-floor corridors to calculate the pressure difference between the indoor areas and the common areas of the corridors. Solar radiation sensors are deployed on the building roof to acquire real-time solar radiation intensity. Outdoor temperature and humidity sensors are deployed on the building roof to acquire outdoor ambient temperature and humidity. Anemometers are deployed on the building roof to acquire outdoor wind speed and direction.

[0098] The distributed execution module includes multiple intelligent ventilation terminals deployed at ventilation openings in the building's external envelope (such as glass curtain walls) to adjust the air supply volume based on received airflow control commands. Each intelligent ventilation terminal includes a fan, a damper, an airflow detection unit, and a local controller (using an embedded microprocessor). The local controller, upon receiving the optimal target airflow command from the regional collaborative control module, precisely tracks and controls the airflow by adjusting the fan speed and damper opening, and reports the operating status. The intelligent ventilation terminal is equipped with a wind pressure sensor and a speed decoder to measure the incoming airflow.

[0099] The regional collaborative control module includes regional coordination controllers installed on building floors. Each regional coordination controller is communicatively connected to all intelligent ventilation terminals and corresponding sensors in the environmental sensing module on its floor, with outdoor sensors shared by all floors. The regional coordination controllers communicate with the central controller of the existing HVAC system via the building automation network. Based on the fused information collected by the environmental sensing module, they perform rolling optimization using a model predictive control algorithm to generate and send optimal airflow control commands (sending valve opening adjustment requests) to the intelligent ventilation terminals, as well as collaborative adjustment commands with the existing HVAC system (supply air temperature setpoint adjustment suggestions). The model predictive control algorithm uses a state-space prediction model that integrates solar radiation feedforward characteristics, indoor state feedback characteristics, and system coupling characteristics for rolling optimization.

[0100] The system optimization module includes a central management platform, which is communicatively connected to the regional collaborative control module. The central management platform is used for global load forecasting and operational strategy optimization (global strategy learning) based on building schedules, weather forecasts, and historical operational data. It also optimizes control parameters in the regional collaborative control module (performing adaptive tuning) and issues advanced strategy parameters such as dynamic temperature setpoints and the timing of terminal operation mode switching.

[0101] The above-mentioned control method for a thermal environment control system based on the fusion of natural cold sources and multi-source information includes the following steps:

[0102] S1: Real-time acquisition and fusion of multi-source information:

[0103] The solar radiation intensity I and outdoor meteorological parameters T are collected synchronously through the environmental sensing module. out Indoor local temperature distribution T zone_i The system collects the indoor and outdoor air pressure difference ΔP and the status information of the existing HVAC system, and performs spatiotemporal registration and feature fusion.

[0104] S2: Heat load assessment and assessment of natural cooling source availability:

[0105] Based on fused information, the regional collaborative control module identifies local overheated areas and their heat load magnitude; at the same time, it determines whether the outdoor air temperature is lower than the indoor overheated area temperature to determine the availability of natural cooling sources.

[0106] S3: Cooperative control decision based on multi-objective optimization:

[0107] When a natural cooling source is available (the following conditions must be met simultaneously: outdoor air temperature is lower than the indoor overheated area temperature setting threshold; outdoor air humidity is lower than the set upper limit and there is no risk of condensation; outdoor wind speed is lower than the set upper limit), the regional collaborative control module aims to minimize the total system energy consumption, maintain indoor thermal comfort and air pressure stability. Based on the model predictive control algorithm, it uses fused information (especially feedforward information of solar radiation and feedback information of indoor temperature) to solve for the optimal target air volume of each intelligent ventilation terminal and generate collaborative action commands with the original HVAC system.

[0108] S4: Layered Precise Execution and Dynamic Collaboration:

[0109] The distributed execution module sends the target airflow command to each intelligent ventilation terminal, precisely executing it by adjusting fan speed and valve opening, and dynamically adjusting indoor air pressure balance. Through an independent air pressure closed-loop control loop, based on air pressure difference feedback, it dynamically adjusts the airflow ratio between the intake and exhaust terminals using a proportional-integral control algorithm, or links and controls relevant valves in the existing HVAC system to maintain the set indoor air pressure. Simultaneously, it sends the coordinated action command to the existing HVAC system, causing it to adjust its airflow output accordingly, thus achieving coordinated operation between the thermal environment control system and the building's original HVAC system.

[0110] S5: Global Strategy Optimization and Self-Learning

[0111] The system optimization module (central management platform) performs medium- and long-term strategy optimization based on historical operating data, weather forecasts, and construction schedules, and updates the control parameters through self-learning (using machine learning to continuously optimize the control parameters of the area controller), thereby improving the system's adaptive capabilities.

[0112] Example 1

[0113] This embodiment introduces a thermal environment control system based on the fusion of natural cold sources and multi-source information, such as... Figure 1 and 2 As shown, the system is deployed in an office building with a glass curtain wall to address the problem of excessive heat in the western area during the afternoon.

[0114] 1. Distributed execution module.

[0115] Intelligent ventilation terminals are installed at pre-designated access ports or ventilator locations on the glass curtain wall of each floor on the west side. The intelligent ventilation terminals use low-noise EC fans with continuously adjustable airflow from 50-400 m³ / h. The air valves are servo motor-driven throttle valves. The local controller uses an embedded microprocessor, pre-stores airflow-speed-opening characteristic curves, and supports RS-485 or wireless communication.

[0116] 2. Environmental perception module, its composition is described in [reference needed]. Figure 2 .

[0117] Solar radiation sensor: A total radiation meter is installed on the roof to measure the total solar radiation intensity in real time, and the data is transmitted through the building network.

[0118] Outdoor temperature and humidity sensor and wind speed and direction sensor: Utilize the existing weather station on the roof to obtain outdoor temperature, humidity, wind speed and direction data.

[0119] Temperature sensors: A high-precision digital temperature sensor is deployed under the ceiling near the window in each bay on the west side to form a temperature monitoring array.

[0120] Differential pressure sensor: A micro differential pressure transmitter is installed in the central area of ​​a typical floor to monitor the air pressure difference between the room and the corridor;

[0121] Building Automation System Interface: Through the BACnet / IP protocol gateway, the air supply temperature, damper opening degree, and FCU operating mode of the VAV box on the west floor are read from the building's BAS system.

[0122] 3. Regional Cooperative Control Module

[0123] Zone controller: One unit is configured on each floor. It receives all data from the sensing modules via a wired network. Its core algorithm integrates feedforward control based on radiation intensity, PID feedback control based on indoor temperature deviation, and constraint control based on air pressure difference.

[0124] Central Management Platform: Deployed on a cloud server or local server, providing a web management interface. The platform integrates a weather forecast API, automatically downloads daily weather forecasts, and runs model predictive control algorithms to generate recommended ventilation strategies for each floor for different time periods the following day (e.g., from 1:00 PM to 4:00 PM, the baseline ventilation volume on the west side is increased to 60%).

[0125] Example 2

[0126] This embodiment describes the control method of the thermal environment control system based on the fusion of natural cold sources and multi-source information in Embodiment 1, such as... Figure 3 As shown, it includes the following steps.

[0127] S1: Real-time acquisition and fusion of multi-source information;

[0128] S11: The system collects data synchronously at a 1-minute interval; measurements are taken from the rooftop weather station. =580W / m², T out =5°C, V wind =2.1m / s (northwest wind); Indoor T measured by indoor sensor array zThe room temperature was set at 22°C, between 26.5°C and 28.2°C. A differential pressure sensor measured ΔP = +7 Pa. The building's BAS system data communication acquired the VAV supply air temperature T. s =22℃, damper opening α vav =65%;

[0129] S12: Perform spatiotemporal registration, marking all data with timestamps t. k =13:00:00, the temperature point is associated with the location on the building floor plan, forming a temperature field vector T. z (t k The data is valid after filtering.

[0130] S13: The regional controller calculates the fusion features, where the feedforward feature is F. ff : [I sol =580,dI sol / dt≈15,T out =5];Feedback characteristic is F fb :[ΔT max =1.7, σ T =0.5, ΔP=7, PMV est ≈+1.5 (slightly hot)]; the system characteristic is F sys : [M sys =Heating,α vav =65%,Q s ,(T s -avg(T z ))≈-5].

[0131] S2: Heat load assessment and assessment of natural cooling source availability;

[0132] S21: Heat load identification, based on I sol =580W / m² and curtain wall parameters, the instantaneous radiative heat gain Q in the west zone was calculated. gain >35kW, significantly higher than the North Zone, clearly indicating that the overheated area is located in the western window area Z. hot ;

[0133] S22: Cold source assessment, check cold air access criteria: ①T out (5) <min(T z_hot (26.5)-3; ② Low outdoor humidity, no risk of condensation; ③ v wind =2.1<5m / s; the natural cold source is deemed to be of high quality and usable, with a theoretical cooling power P. cool,max considerable;

[0134] S23: Trigger determination, due to ΔT maxWhen the temperature is above 2°C and a cold source is available, triggering function Ψ(t_k)=1, the system immediately enters the intelligent collaborative ventilation mode.

[0135] S3: Cooperative control decision based on multi-objective optimization;

[0136] S31: The area controller constructs the optimization problem at the current moment, with the optimization variable being the target air volume Q of the four smart ventilators in the overheated area. vent [1,2,…,n]; The core constraints of the cost function are as described above;

[0137] S32: MPC rolling optimization, based on weather forecast, predicting the next 30 minutes I sol It will remain high and then decline slowly. This prediction sequence {F} ff (t k+1|k Input state-space prediction model:

[0138] [T z [(t+1);ΔP(t+1)]=A*[T] z (t);ΔP(t)]+B*Q vent (t)+B d *F ff (t);

[0139] Under constraints such as air volume, opening degree, and wind speed comfort, solve for the optimal control sequence for the next 15 minutes to minimize the total cost over the prediction period. The optimal command for the current moment is calculated as: Q* vent =[120,115,...] m³ / h (4 values ​​in total), and calculate that to balance this increased air volume, it is recommended to adjust the opening α of the return air damper in the west zone of this floor. vav Increased by 8%;

[0140] S33: Increase airflow command Q* vent The request was sent to the four corresponding ventilation terminals; simultaneously, a coordination request {Δα} was sent. vav =+8%} is sent to the building's BAS via BACnet / IP.

[0141] S4: Precise execution and dynamic collaboration;

[0142] S41: Terminal execution, for example, ventilation terminal 1 receives Q. vent *[1]=120m³ / h. Its local controller is based on K ff *120 Quickly set the initial fan speed. Simultaneously, based on the actual airflow Q fed back by the air pressure sensor... actual via PI controller K p *e(t)+K i *∫e(t) fine-tunes the rotation speed, eventually stabilizing at 120 m³ / h. The ventilator's inlet valve is based on α=fmap (120,ΔP duct (Refer to the table and automatically set the opening to 55° to optimize the flow field;)

[0143] S42: Intra-area air pressure regulation. After execution begins, the indoor ΔP fluctuates briefly. The air pressure closed loop operates independently. When ΔP rises to 9.5Pa, it fine-tunes the ratio of the intake and return air volume at the terminal, stabilizing ΔP back to 8±0.5Pa within 30 seconds.

[0144] S43: Building system coordination, BAS receives Δα vav After a request of +8%, the system automatically adjusts and feeds back the actual total air volume and total power consumption to the BAS for building-wide energy metering.

[0145] S5: Global strategy optimization and self-learning. In this example, the central management platform, based on weather forecasts and execution history, adjusted the MPC weights for the afternoon based on the weather forecasts and execution history. T It has been slightly increased by 5% to enhance the proactive nature of cooling. All operating data (F ff ,F fb The values ​​(u*, J) are stored in the database. Weekly machine learning model analysis on the platform reveals that, under similar weather conditions, starting ventilation slightly earlier (10 minutes earlier) results in a lower overall comfort cost (J). Therefore, the system automatically updates the strategy library for future predictive control.

[0146] Through the above closed-loop operation, the low-temperature outdoor air is successfully transformed into a free cooling resource. Through the deep integration of multi-source information and intelligent decision-making, precise, efficient and green control of the building thermal environment is achieved, which significantly improves indoor comfort and reduces building energy consumption.

[0147] Although the embodiments of this application have been described above, this application is not limited to the specific embodiments and application fields described above. The specific embodiments described above are merely illustrative and instructive, and not restrictive. Those skilled in the art, based on the guidance of this specification and without departing from the scope of protection of the claims of this application, can make many other forms, all of which are within the scope of protection claimed in this application.

Claims

1. A thermal environment regulation system based on natural cold source and multi-source information fusion, characterized in that, It includes an environmental perception module, a distributed execution module, a regional collaborative control module, and a system optimization module; The environmental sensing module includes a sensor network deployed inside and outside the building to collect real-time environmental data and obtain operating status information from the building's existing HVAC system; the environmental data includes solar radiation intensity, outdoor meteorological parameters, indoor temperature and indoor-outdoor air pressure difference; The distributed execution module includes an intelligent ventilation terminal deployed at the ventilation openings of the building's external envelope, used to adjust the air supply volume according to the received air volume control commands and dynamically adjust the indoor air pressure balance; The regional collaborative control module includes regional coordination controllers installed on building floors. Each regional coordination controller is communicatively connected to all intelligent ventilation terminals and corresponding sensors in the environmental sensing module on its floor, with outdoor sensors shared by all floors. The regional coordination controller communicates with the central controller of the original HVAC system through the building automation system. Based on the fused information collected by the environmental sensing module, it identifies local overheated areas and their heat load magnitude. At the same time, it determines whether the outdoor air temperature is lower than the indoor overheated area temperature to determine the availability of natural cooling sources. When a natural cooling source is available, the regional collaborative control module aims to minimize the total energy consumption of the system, maintain indoor thermal comfort and air pressure stability. Based on the model predictive control algorithm, it uses fused information to solve for the optimal target air volume of each intelligent ventilation terminal and generates collaborative action commands with the original HVAC system. The model predictive control algorithm employs a state-space prediction model that integrates feedforward features of radiation and meteorology, feedback features of indoor conditions, and system coupling features of the original HVAC system conditions for rolling optimization. The system optimization module is communicatively connected to the regional collaborative control module and is used to perform global strategy learning based on building schedules, weather forecasts, and historical operation data to optimize the control parameters in the regional collaborative control module.

2. The thermal environment regulation system based on natural cold source and multi-source information fusion according to claim 1, characterized in that, The sensor network includes: solar radiation sensors, outdoor temperature and humidity sensors, and wind speed and direction sensors deployed on the roof of the building; temperature sensors deployed in the controlled areas inside the building; and differential pressure sensors for measuring the air pressure difference between the indoor area and the corridor on the same floor.

3. The thermal environment regulation system based on natural cold source and multi-source information fusion according to claim 1, characterized in that, The intelligent ventilation terminal includes a fan, a damper, an air volume detection unit, and a local controller. The local controller is used to accurately track and control the air volume by adjusting the fan speed and the damper opening according to the target air volume command sent by the regional collaborative control module, and reports the operating status to the regional collaborative control module. The intelligent ventilation terminal is equipped with a wind pressure sensor and a speed decoder for measuring the intake air volume.

4. The method of claim 1-3, wherein the method is characterized in that, Includes the following steps: S1: Real-time acquisition and fusion of multi-source information: The solar radiation intensity I and outdoor meteorological parameters T are collected synchronously through the environmental sensing module. out Indoor local temperature distribution T zone_i The system collects the indoor and outdoor air pressure difference ΔP and the status information of the existing HVAC system, and performs spatiotemporal registration and feature fusion. S2: Heat load assessment and assessment of natural cooling source availability: Based on fused information, the regional collaborative control module identifies local overheated areas and their heat load magnitude; at the same time, it determines whether the outdoor air temperature is lower than the indoor overheated area temperature to determine the availability of natural cooling sources. S3: Cooperative control decision based on multi-objective optimization: When a natural cooling source is available, the regional collaborative control module aims to minimize the total energy consumption of the system, maintain indoor thermal comfort and air pressure stability. Based on the model predictive control algorithm, it uses fused information to solve for the optimal target air volume of each intelligent ventilation terminal and generates collaborative action commands with the original HVAC system. The determination that a natural cold source is available requires that the following conditions be met simultaneously: the outdoor air temperature is lower than the set threshold for the indoor overheated area temperature; the outdoor air humidity is lower than the set upper limit and there is no risk of condensation; and the outdoor wind speed is lower than the set upper limit. The model predictive control algorithm employs a state-space prediction model that integrates feedforward features of radiation and meteorology, feedback features of indoor conditions, and system coupling features of the original HVAC system conditions for rolling optimization. S4: Layered Precise Execution and Dynamic Collaboration: The distributed execution module sends the optimal target air volume command to each intelligent ventilation terminal for execution and dynamically adjusts the indoor air pressure balance; at the same time, it sends the collaborative action command to the original HVAC system to achieve collaborative operation with the building's original HVAC system. The dynamic adjustment of indoor air pressure balance is achieved by using an independent air pressure closed-loop control loop and a proportional-integral control algorithm to dynamically adjust the air volume ratio between the air intake and exhaust terminals in order to maintain the set indoor air pressure. S5: Global Strategy Optimization and Self-Learning The system optimization module performs medium- and long-term strategy optimization based on historical operating data, weather forecasts, and construction schedules, and updates the control parameters through self-learning.

5. The method according to claim 4, wherein the method is characterized by, Step S1, real-time acquisition and fusion of multi-source information, specifically includes: S11: Synchronous acquisition of multi-source heterogeneous data. Through a distributed sensor network and the building automation system interface, the following five types of heterogeneous data are acquired in real time: Radiation perturbation data: wherein is the solar radiation intensity at time t; Outdoor weather data: , representing respectively the outdoor temperature, the relative humidity, the wind speed and the wind direction; Indoor thermal environment data: ,in Let be the indoor temperature at the i-th monitoring point, and n be the total number of monitoring points. The difference between indoor and outdoor air pressure; Building original HVAC system state data : Obtained from building automation system, including system operation mode , key air valve opening , supply air temperature , return air temperature and air volume ; S12: Spatiotemporal registration and data preprocessing, performing spatiotemporal alignment and cleaning on the acquired asynchronous and heterogeneous raw data; the first step is time synchronization for all data streams. Inject a unified high-precision timestamp k is the sampling sequence number, forming a synchronous dataset. The second step is spatial mapping, which maps the data from each temperature monitoring point. In relation to its spatial location within the building, a vector representation of the temperature field distribution is constructed. The third step is validity verification, which uses sliding window mean filtering and the Laida criterion to remove outliers and ensure data reliability. S13: The feature quantities in feature-level fusion and key state quantity extraction are: Feedforward characteristics based on radiation and meteorology: ； wherein is the solar radiation intensity, is the rate of change of radiation, for predicting heat load trends; Feedback characteristics based on indoor conditions: ； in, The degree of thermal imbalance is quantified by the maximum regional temperature difference; The standard deviation of the temperature field; For regional pressure difference, The average vote value is the forecast estimate based on the current temperature and humidity. System coupling characteristics based on the existing HVAC system status: ； wherein, T is the indoor average temperature, reflects the temperature difference driving potential of the central air conditioning supply air and the indoor environment.

6. The method of claim 4, wherein the method further comprises: The specific process of step S2, heat load assessment and natural cooling source availability determination, is as follows: S21: Dynamic thermal load identification and overheating area positioning, based on radiation and meteorological feedforward characteristic quantities with feedback characteristic quantities of the indoor state , construction of a thermal load estimation model: ， in, Heat is gained from solar radiation. The overall heat gain coefficient of the curtain wall, For the area of ​​the radiation-affected curtain wall, The intensity of solar radiation. Let be the heat transfer coefficient of the j-th building envelope. Let j be the area of ​​the j-th enclosure structure. Let j be the temperature of the building envelope. The target area temperature; by comparing the temperatures of different areas With heat dissipation capacity, the location of overheated core areas is denoted as . ; S22: Multi-criteria evaluation of the quality and availability of natural cooling sources, establishing access criteria for the use of natural cooling sources, which must be met simultaneously: Temperature feasibility: ,in Outdoor air temperature This represents the lowest temperature at each monitoring point in the overheated area. The minimum effective temperature difference threshold; Humidity safety: and , in The relative humidity of outdoor air. The maximum permissible outdoor humidity level, Outdoor air temperature This refers to the indoor air dew point temperature. To ensure a safety margin against condensation, and to prevent the introduction of excessively humid air or condensation on the building envelope; Weather suitability: wherein is the outdoor wind speed, is the allowed maximum wind speed upper limit to avoid strong winds leading to loss of control of the airflow organization or equipment overload; If both are satisfied, then the natural cold source is determined to be available and its maximum theoretical cooling power is calculated where is the air density, is the constant pressure specific volume of air, , is the average temperature of the superheated zone, is the outdoor air temperature; S23: Controlling triggering and mode selection, defining trigger functions. ,in For indicator functions, This represents the maximum temperature difference indoors. Set a threshold for the temperature difference; when When the system is activated, the intelligent collaborative ventilation mode is triggered; otherwise, regular monitoring or standby mode is maintained.

7. The method of claim 4, wherein the method further comprises: The S3-based collaborative control decision based on multi-objective optimization includes the following steps: S31: Construction of a multi-objective optimization problem, in each control cycle The goal is to minimize the overall cost function and solve for the optimal control vector. ,in The target air volume for each ventilation terminal. To adjust the opening of its blinds. The suggested adjustment amount for the opening of the fresh air system damper in the existing HVAC system; cost function Defined as: in, For the collection of overheated regions, Let be the actual indoor temperature at the i-th monitoring point. For dynamic temperature setpoint, This represents the actual air pressure difference between indoors and outdoors. Set the target air pressure difference between indoors and outdoors. The total energy consumption of all intelligent ventilation terminal fans. This refers to the additional energy consumption of the original HVAC system caused by ventilation adjustment. To predict average voting bias, These represent the weighting coefficients for temperature deviation, air pressure deviation, energy consumption, and thermal comfort deviation, respectively. Constraints include equipment constraints: ， ； wherein is the target air volume of the jth intelligent ventilation terminal, are the minimum and maximum air volume of the single intelligent ventilation terminal, respectively, represents the target opening of the air valve of the jth intelligent ventilation terminal; Comfort and safety constraints: ， ； wherein is the supply air velocity, is the maximum comfort air velocity, is the supply air temperature, is the indoor air dew point temperature, is the anti-condensation safety margin; System coupling constraints: ； in air density, This represents the total air intake volume for all intelligent ventilation terminals. This refers to the air supply volume of the original HVAC system. This refers to the return air volume of the original HVAC system. The instantaneous rate of change of indoor air quality; S32: Solving the model predictive control problem based on fused features, using a model predictive control framework for rolling optimization: Prediction model: using the fused feature quantity extracted in S13 Online updating of the building thermal dynamic model; model reduction to state space form: ； Among them, state variables , Let the indoor temperature field vector be... The control quantity is the difference between indoor and outdoor air pressure. , The target airflow vector for the intelligent ventilator. As a feedforward input for measurable disturbances; These are the state transition matrix, control input matrix, and disturbance input matrix, respectively. Scrolling optimization: in each Time, based on the current state and future Perturbation prediction sequence of the step Solving the future under constraints Step control sequence The total cost during the forecast period minimize; Feedback correction: only the first set of control quantities of the optimized sequence as actual instruction output; in the next cycle, with the actual measured correcting for model prediction errors, implementing closed-loop optimization; S33: Cooperative instruction package generation and distribution, which encapsulates the optimization solution results into standardized cooperative instruction packages: For intelligent ventilation terminal: issue instructions to , is the optimal target air volume of the jth intelligent ventilation device, is the optimal air valve opening of the jth intelligent ventilation device; For original HVAC system: generate a coordination request to the building automation system, suggesting to adjust the damper opening or the supply air temperature set point to achieve overall energy saving, wherein the optimal return air damper opening for the original HVAC system, the optimal supply air temperature set point for the original HVAC system.

8. The method of claim 4, wherein the method further comprises: The layered precise execution and dynamic collaboration in S4 include the following steps: S41: Execution layer adaptive flow tracking control. After receiving the command, the intelligent ventilation terminal's local controller adopts a feedforward-feedback composite control strategy. The fan speed control formula and the ventilator damper opening formula are as follows: ， ， The feedforward control gain coefficient is... Set the target air supply volume. This is the proportional control gain coefficient. This is the integral gain coefficient. For air volume deviation, This refers to the actual air volume. To determine the static pressure of the air duct The ventilator damper opening compensation function obtained through dynamic table lookup ensures accurate tracking under varying operating conditions. ; S42: Rapid closed-loop pressure balance regulation, establishing an independent pressure regulation loop; As a set value, with For feedback, proportional control is used to dynamically adjust the ratio of air intake and exhaust terminal air volume. This ensures that the pressure deviation converges quickly, among which... This is the proportional gain coefficient of the pneumatic circuit. Set the pressure difference value. This is a measurement of the air pressure difference; S43: Inter-system data exchange and status synchronization; during execution, the intelligent ventilation system continuously transmits the actual air volume. ,power consumption The status variables are fed back to the building automation system in real time through the system interface, enabling the system to dynamically update the total building load calculation and optimize the operation strategy of the cooling and heating units accordingly, thereby achieving global integrated optimization.

9. The method of claim 4, wherein the method further comprises: The S5 global policy optimization and self-learning includes the following steps: S51 : Run database build, long run data stored to time series database, forming training data set ; S52: Parameter self-tuning based on machine learning, periodically utilizing... Training a deep neural network or reinforcement learning agent with fused features as input. The output is an optimized parameter set that minimizes the long-term average cost; the updated parameters will be sent offline to the regional collaborative control module.

Images