Green building intelligent lighting and energy collaborative optimization system based on multi-source data fusion

Abstract

The application discloses a kind of green building intelligent lighting and energy collaborative optimization system based on multi-source data fusion.System first constructs accurate real-time environment state by dynamic confidence weighted fusion indoor and outdoor illumination, meteorology, personnel occupancy and multi-source information such as electricity price, then adopts model predictive control combined with multi-objective optimization strategy, and cooperatively adjusts sunshade device, intelligent lighting and heating ventilation air conditioner, to realize the minimization of comprehensive energy consumption or operating cost on the premise of guaranteeing indoor illuminance and temperature comfort requirement.In addition, adaptive module continuously learns and adjusts building thermal characteristics and use mode through online model parameter correction and occupancy probability prediction, to ensure long-term stability of energy-saving effect.System drives each execution device through standard building automation interface, is easy to integrate in new or reconstruction project, and can significantly improve energy utilization efficiency and indoor environment quality.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for green buildings, specifically to an intelligent lighting and energy synergistic optimization system for green buildings based on multi-source data fusion. Background Technology

[0002] With the rise of green building concepts, optimizing building energy consumption and improving indoor environmental quality have become important issues. Traditional building lighting and air conditioning systems are often controlled independently, lacking coordination: lighting systems are typically adjusted according to fixed schedules or simple light sensors, while HVAC systems operate independently based on thermostat settings. This fragmented control approach fails to fully utilize natural light and other available information in the environment, leading to the following problems:

[0003] Energy waste: Artificial lighting may still be turned on when there is sufficient sunlight during the day, resulting in wasted electricity; conversely, opening curtains to increase natural light may introduce excessive solar radiation heat, increasing the air conditioning load. The lack of coordinated optimization between lighting and air conditioning leads to inefficient energy use.

[0004] Insufficient comfort: Relying on a single sensor to control lighting or temperature cannot fully perceive the actual indoor usage conditions. For example, maintaining high brightness and air conditioning when the room is empty, or insufficient lighting and unsuitable temperature when someone is present, both affect comfort.

[0005] Lack of intelligent adaptation: Most existing systems are based on preset rules or simple feedback control, making it difficult to adapt to dynamically changing environments and usage patterns. For example, the system cannot respond and optimize in advance when there are sudden weather changes, changes in sunshine patterns, or fluctuations in electricity prices.

[0006] Insufficient data utilization: Modern buildings are equipped with various sensors (illuminance, temperature and humidity, occupancy detection, etc.) and access to external data (weather forecasts, electricity prices, etc.), but traditional control rarely integrates and utilizes this multi-source data, failing to unleash the potential of data-driven optimization.

[0007] In view of the above shortcomings, there is an urgent need for an intelligent control system that can integrate multi-source data such as indoor and outdoor lighting, occupancy, climate conditions, and electricity prices, and use this data to coordinate and optimize lighting and air conditioning energy. Summary of the Invention

[0008] Technical Objective: To address the problems of energy waste and poor comfort caused by the independent control of lighting and air conditioning in existing systems, this invention discloses a green building intelligent lighting and energy synergistic optimization system and method based on multi-source data fusion. By integrating various data related to indoor and outdoor environment and energy, and applying intelligent algorithms to coordinate and control lighting and energy, the system achieves comprehensive optimization of lighting energy consumption and air conditioning energy consumption, significantly improving the building's energy efficiency and environmental quality.

[0009] Technical solution: To achieve the above technical objectives, the present invention adopts the following technical solution:

[0010] A green building intelligent lighting and energy synergistic optimization system based on multi-source data fusion includes:

[0011] The multi-source data acquisition and fusion module is used to acquire multiple data sources from inside and outside the building, and to fuse the data to generate comprehensive environmental status information for control decisions. The multiple data sources include at least indoor illuminance, outdoor illuminance and meteorological data, indoor occupancy information, indoor environmental parameters and electricity price data.

[0012] The intelligent lighting control module is used to adjust the building's lighting devices and artificial lighting equipment according to the comprehensive environmental status information, so as to realize intelligent control of indoor illuminance. The intelligent lighting control module is connected to the adjustable lighting devices and lighting fixtures. By controlling the opening degree of the adjustable lighting devices and the output brightness of the lighting fixtures, the target illuminance is maintained by giving priority to natural light and supplementing artificial lighting when there is insufficient light, while avoiding glare.

[0013] The energy collaborative optimization control module is used to optimize the control of the building's lighting system and HVAC system by integrating the comprehensive environmental status information, so as to collaboratively reduce energy consumption. The energy collaborative optimization control module is based on a predetermined dual optimization objective function, and performs overall calculation on the lighting control output and the air conditioning control output to balance the lighting energy consumption and air conditioning energy consumption under the indoor illuminance requirements and temperature comfort requirements, so as to minimize the total energy consumption or operating cost.

[0014] An environmental prediction and adaptation module is used to predict future environmental conditions based on historical data and external prediction information and dynamically adjust control strategy parameters. The environmental prediction and adaptation module includes models for predicting changes in outdoor light intensity and indoor heat load, as well as a mechanism for adaptively correcting control model parameters based on real-time feedback.

[0015] The control decision and execution module is used to receive lighting control and air conditioning control commands output by the energy collaborative optimization control module and send them to the execution device to adjust the working status of indoor lighting fixtures and air conditioning equipment. The control decision and execution module includes a lighting execution unit connected to adjustable lighting equipment and lighting fixtures, and an air conditioning execution unit connected to the HVAC system, and controls indoor illuminance and temperature in conjunction with the optimal control decision.

[0016] Preferably, the multi-source data acquisition and fusion module includes an indoor illuminance sensor, an outdoor light sensor, an indoor occupancy sensor, a temperature and humidity sensor, and a data interface for acquiring weather forecasts and electricity price information; the multi-source data acquisition and fusion module filters, synchronizes, and normalizes data from different sources, and uses weighted averaging, state estimation, or machine learning algorithms to fuse the data into one or more comprehensive indicators characterizing the current indoor environmental light and heat state.

[0017] Preferably, the intelligent lighting control module includes motorized curtains or adjustable blinds and dimmable lighting fixtures. The intelligent lighting control module determines the opening angle or light transmittance of the motorized curtains or adjustable blinds and the dimming level of the dimmable lighting fixtures in real time based on the indoor and outdoor light intensity and the indoor illuminance setting value through fuzzy inference and closed-loop control algorithms. When it detects that there are people in the room and the natural light is insufficient, it turns on or brightens the lighting fixtures. When it detects that the natural light is sufficient or there are no people, it dims or turns off the lighting fixtures. When the external light is too strong and causes glare or excessive heat gain, it automatically retracts the motorized curtains or adjustable blinds to reduce the light transmittance.

[0018] Preferably, the control decision and execution module is connected to the lighting and air conditioning equipment through a standard building automation communication interface to implement control commands; the lighting execution unit controls the output of the dimmable lighting fixtures through a DALI or 0-10V interface and controls the opening and closing of electric curtains or adjustable blinds through motor drive; the air conditioning execution unit sends temperature setting, valve opening, or fan speed commands to the HVAC system through BACnet or MODBUS interfaces, and receives execution feedback in real time to ensure that the commands are executed in place and to perform closed-loop verification.

[0019] Preferably, the energy synergistic optimization control module pre-stores mathematical models for evaluating the energy consumption of the lighting and air conditioning systems, and defines an objective function for synergistic optimization of lighting and energy. This objective function weights and combines indoor illuminance deviation, indoor temperature deviation, and energy consumption to simultaneously characterize lighting comfort and energy cost. The energy synergistic optimization control module uses model predictive control or a genetic algorithm to solve the objective function in real time, obtaining optimal lighting and air conditioning adjustment commands. This ensures that, while meeting indoor illuminance and temperature requirements, the total energy consumption or energy cost is optimized. The formula for calculating the objective function is:

[0020] J=α×(I set -I in ) 2 +β×(T in -T set ) 2 +γ×(P L +P HVAC )

[0021] Among them I set I is the indoor illuminance setpoint. in T represents the measured indoor illuminance. set T is the air conditioner temperature setpoint. in P represents the actual indoor temperature. L and P HVAC These represent the power or energy consumption of the lighting system and the air conditioning system, respectively, with α, β, and γ being the weighting coefficients for indoor illuminance deviation, indoor temperature deviation, and energy consumption, respectively.

[0022] Preferably, the environmental prediction and adaptation module utilizes outdoor weather forecast data and indoor sensor historical data to employ a machine learning model to predict outdoor light intensity, indoor cooling load demand, or occupancy within a predetermined time interval. The environmental prediction results are provided to the energy collaborative optimization control module for proactively adjusting the control scheme. Simultaneously, the environmental prediction and adaptation module automatically corrects the parameters of the energy consumption model or fuzzy control rules by monitoring the deviation between the actual control effect and the prediction or model, thereby ensuring that the control model is automatically calibrated with seasonal changes and equipment aging factors.

[0023] A method for intelligent daylighting and energy synergy optimization in green buildings based on multi-source data fusion includes the following steps:

[0024] S1. Collect multi-source data related to indoor and outdoor lighting, occupancy, meteorology and energy, preprocess and time-align the collected data, and use data fusion algorithms to generate comprehensive characterization parameters of the current environmental state;

[0025] S2. Based on the comprehensive characterization parameters of the current environmental state and the obtained external prediction information, predict the changes in outdoor light intensity, indoor thermal environment, or room occupancy within a predetermined period in the future.

[0026] S3. Establish a coordinated energy consumption model for lighting and air conditioning and optimize the objective function. Based on the comprehensive characterization parameters and prediction results of the current environmental state, use the optimization algorithm to calculate the optimal combination decision of lighting control and air conditioning control under the constraints of indoor illuminance and thermal comfort, including the opening degree of adjustable daylighting equipment, the dimming level of lighting fixtures and the set control amount of air conditioning temperature.

[0027] S4. Send the optimal combination decision to the corresponding lighting and air conditioning actuators to adjust the operating status of the indoor lighting and air conditioning systems, thereby achieving coordinated control of the indoor light and thermal environment.

[0028] S5. Monitor the actual indoor illuminance and temperature feedback information after step S4 is executed, and compare it with the predicted value or set value. If there is a deviation, adaptively adjust the model parameters or control strategy in subsequent control cycles, including updating data fusion weights, optimizing algorithm parameters or fuzzy rules, so that the system can adapt to new environmental characteristics.

[0029] S6. Repeat steps S1 to S5 to form a cyclical control process to continuously optimize indoor lighting utilization and energy consumption.

[0030] Preferably, the fusion of multi-source data in step S1 includes: acquiring the illuminance at various points indoors and calculating the average illuminance or minimum illuminance value; acquiring the outdoor horizontal illuminance and converting it into available indoor illuminance based on the window area and light transmittance; combining the occupancy sensor status to determine the room usage demand; and comprehensively forming a light environment index for lighting decision-making. Simultaneously, acquiring the current air conditioning operating parameters and indoor and outdoor temperature and humidity, estimating the current building heat load status, fusing multi-sensor readings through Kalman filtering to obtain smooth illuminance and temperature values, and integrating occupancy information and comfort requirements into the environmental index through fuzzy logic to obtain a fuzzy comprehensive evaluation value describing the current light and heat environment.

[0031] Preferably, the optimization algorithm in step S3 adopts the model predictive control method, specifically including: establishing a prediction model with the current state as the initial state in each control cycle, discretizing candidate lighting and air conditioning control sequences in the prediction time domain, calculating the overall optimization target value by simulating the changes in indoor illuminance and temperature and the accumulation of energy consumption under the action of these control sequences, selecting the control sequence that minimizes the target value as the optimal solution, and only executing the control action of the current cycle in the controlled sequence, and repeatedly performing optimization calculations in the next cycle using new state feedback and prediction information.

[0032] Preferably, the adaptive adjustment in step S5 includes: when the actual indoor illuminance deviates from the target value for a long period of time, adjusting the gain parameter of the illuminance control loop or relearning the lighting energy consumption model; when the actual temperature control overshoot or lag is detected, adjusting the PID parameter of the air conditioning control or the heat transfer coefficient of the prediction model; when environmental or load mode changes cause the original fuzzy rules to no longer be applicable, automatically activating the previously trained and stored rule set for the new mode or using a fuzzy adaptive algorithm to generate new control rules, thereby ensuring that the system can maintain efficient and stable control performance under different environmental conditions.

[0033] Beneficial Effects: The green building intelligent lighting and energy synergistic optimization system and method based on multi-source data fusion provided by this invention has the following beneficial effects:

[0034] 1. This system combines maximizing the use of natural lighting with air conditioning energy consumption control. While ensuring comfortable indoor illuminance and temperature, it reduces overall energy consumption. By coordinating and optimizing lighting and HVAC, it can reduce lighting power consumption and reduce air conditioning load caused by sunlight, achieving comprehensive energy saving. Especially during peak electricity price periods, the system can automatically reduce unnecessary lighting and air conditioning power, avoiding high-cost energy use, thereby saving operating costs.

[0035] 2. Compared to traditional single-sensor control, this invention integrates multi-dimensional data such as illumination, occupancy, climate, and electricity prices, making control decisions more comprehensive and scientific. For example, the system can turn off lights and air conditioning to save energy when it senses that the room is empty and well-lit, and adjust to a comfortable state in advance when it senses that someone is present; it can also adjust lighting in advance before a sudden drop in sunlight, based on weather forecasts, to avoid the impact of light fluctuations. The fusion of multi-source data enables the system to have context awareness and autonomous decision-making capabilities, significantly improving the level of intelligence in control.

[0036] 3. By introducing fuzzy logic rules, the system can handle the uncertainties of sensor data and the fuzzy factors of human comfort, ensuring a smooth and flexible control process. Combined with model predictive control to account for future disturbances, the system can still operate optimally under sudden weather changes or load fluctuations, exhibiting good robustness and foresight. The combination of these two approaches achieves a hardware-software integration of intelligent lighting and energy management, avoiding the control lag or oscillation problems that may occur with traditional rigid control, and ensuring the dual stability of comfort and energy-saving effects.

[0037] 4. The system of this invention can continuously learn the building's usage patterns and environmental characteristics through an adaptive module, and dynamically adjust the control strategy. For example, it automatically corrects the balance strategy between lighting and air conditioning as the seasons change, and automatically adjusts the filtering weights as sensor noise increases due to aging. This ensures that the system remains close to its optimal state in long-term operation, has continuous optimization capabilities, and overcomes the performance degradation caused by the traditional control system's "one-time setting, long-term unchanging" approach. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0039] Figure 1 This is a schematic diagram of the overall system architecture of the present invention;

[0040] Figure 2 This is a flowchart of the method of the present invention;

[0041] Figure 3 This is a schematic diagram illustrating the effect of the shading opening degree on light collection and thermal gain according to the present invention.

[0042] Figure 4 This is a schematic diagram illustrating the occupancy probability prediction of the present invention. Detailed Implementation

[0043] The present invention will now be described more clearly and completely by way of a preferred embodiment in conjunction with the accompanying drawings, but this does not limit the invention to the scope of the described embodiment.

[0044] The system of this invention comprises three layers: a perception layer, a decision-making layer, and an execution layer. The perception layer consists of multi-source sensors and data interfaces, used to acquire data related to the building environment and energy. The decision-making layer, consisting of multi-source data acquisition and fusion, environmental prediction and adaptation, and energy collaborative optimization control algorithms, is the brain of the system, responsible for calculating the optimal control strategy. The execution layer includes equipment such as lighting and air conditioning, which receives instructions through the control decision and execution module to regulate the physical environment. All layers work together to achieve intelligent lighting and energy optimization control.

[0045] The system is deployed within a green building, with various sensors in the sensing layer installed in typical rooms and on the building's exterior: indoor illuminance sensors are positioned at work surface height to detect light intensity, while outdoor light sensors are installed on the building facade to detect changes in natural light throughout the day; indoor temperature and humidity sensors, combined with sensors integrated into the HVAC system, monitor environmental thermal parameters; infrared human body sensors or ultrasonic detectors are installed on the ceiling to detect occupancy (occupied / unoccupied); smart meters and power consumption monitoring devices provide real-time power consumption data for lighting circuits and air conditioning equipment; additionally, it accesses weather forecast services via an internet interface to obtain future weather and sunshine prediction data, and connects to the power company to obtain time-of-use electricity pricing information. This data from various sources is transmitted via wired or wireless networks to the building energy management system (BEMS)'s multi-source data acquisition and fusion module.

[0046] The core controller at the decision-making level can use a high-performance programmable logic controller (PLC) or an industrial computer to run the software algorithm of this invention. The controller is internally divided into several functional modules: a multi-source data acquisition and fusion module, an environmental prediction and adaptation module, and an energy-coordinated optimization control module. The multi-source data acquisition and fusion module is responsible for preprocessing and comprehensively evaluating the multi-source data uploaded from the sensing layer; the environmental prediction and adaptation module uses fused historical data and external information to make short-term predictions and fine-tunes control parameters based on the differences between the recent system output and environmental feedback; the energy-coordinated optimization control module solves for the optimal control output accordingly. The modules exchange information through shared memory or a message bus: for example, the environmental state after data fusion is sent to the environmental prediction and adaptation module, the prediction results are provided to the energy-coordinated optimization control module, and the updated parameters from the environmental prediction and adaptation module are applied to the optimization algorithm.

[0047] The execution layer includes the lighting system and the HVAC system. The lighting subsystem consists of dimmable LED lights, controllable motorized curtains, or adjustable blinds, etc. These devices receive dimming and opening / closing commands from the controller to physically regulate the indoor lighting environment. The HVAC system includes variable air volume (VAV) air handling units, thermostatic valves, and other actuators, which regulate indoor temperature and air quality using commands from the controller, such as temperature settings and valve opening parameters. Furthermore, the execution layer also includes drive interfaces and communication networks, such as DALI (Digital Addressable Lighting Interface) for lighting equipment control, and BACnet or MODBUS protocols for air conditioning equipment control, ensuring that the controller can accurately issue commands and monitor the execution status.

[0048] like Figure 1 As shown, the green building intelligent lighting and energy synergistic optimization system based on multi-source data fusion provided by the present invention specifically includes:

[0049] The multi-source data acquisition and fusion module is used to collect data from multiple data sources inside and outside the building, and to fuse the data to generate comprehensive environmental status information for control decisions. The multiple data sources include at least indoor illuminance, outdoor illuminance and meteorological data, indoor occupancy information, indoor environmental parameters and electricity price data.

[0050] This module collects and timestamps the following data from various sources:

[0051] Indoor illuminance data: Provided by illuminance sensors distributed in different areas of the room, reflecting the current illuminance levels at key points within the room. To fully characterize the light environment, multi-dimensional data such as work surface illuminance and wall brightness can be collected.

[0052] Outdoor lighting and climate data: Acquired through light sensors installed around the building or from weather stations, including outdoor horizontal illuminance, solar radiation intensity, solar angle, and meteorological data such as outdoor temperature and humidity. This data helps determine the potential for natural lighting and its impact on air conditioning load.

[0053] Indoor occupancy information: Acquired by pyroelectric infrared (PIR) sensors, ultrasonic detectors, and camera image recognition, this information is used to determine whether there are people in the room and their activity levels. The occupancy data can be updated at a set frequency to meet real-time requirements, identifying periods of low occupancy and periods of high usage.

[0054] Indoor environmental parameters: The HVAC system’s built-in or independent environmental sensors provide indicators such as indoor temperature, relative humidity, and carbon dioxide concentration, reflecting thermal comfort and air quality.

[0055] Energy and electricity price data: This includes power / heat consumption data for lighting circuits and air conditioning equipment, collected by energy consumption monitoring instruments; as well as energy price signals from the power grid, such as time-of-use pricing and demand control. This type of data provides real-time quantitative data on energy consumption and costs for optimization algorithms.

[0056] For the above multi-source data, necessary filtering and cleaning are first performed. This includes eliminating sensor sampling noise (using moving average or Kalman filtering), filling in occasional missing data, and interpolating or aggregating data at different time granularities to unify the time step. In addition, heterogeneous data are mapped to a unified scale: for example, occupancy status is encoded as 0 / 1 or probability values, and qualitative weather forecast information in the form of "high / medium / low" is converted into quantitative indices.

[0057] This module uses a data fusion algorithm to integrate preprocessed multi-source information to form a comprehensive environmental state vector at the current moment. The following fusion strategies can be adopted:

[0058] Rule-based weighted fusion: For the same physical quantity from data from multiple sensors, a more accurate estimate is obtained by weighting the data according to the sensor's trust level (e.g., fusion of multiple lux meter readings). Different types of information are weighted according to their importance to control decisions, and a comprehensive index is calculated. For example, defining the ambient brightness comprehensive index L... env Indoor illuminance L in outdoor illuminance L out Weighted average yields:

[0059] L env =ω1·L in +ω2·f(L out A window )

[0060] Where f(L) out A window ) represents the equivalent outdoor illuminance that can be used indoors, calculated based on window area and light transmittance, where ω1 and ω2 are weights (which can be calibrated experimentally or adjusted adaptively).

[0061] Model inference fusion: A state-space model for data fusion is established, using state variables to represent implicit environmental states that are difficult to measure directly (such as overall light comfort, thermal comfort index, etc.), and the state estimate is updated through sensor observations. For example, a latent variable X is introduced to represent "the degree to which the current environment meets comfort requirements," and an observation equation is established to map illuminance, temperature, occupancy, etc., to fuzzy measures of X. Bayesian inference or Kalman filtering is used to update the estimate of X in real time, thereby fusing multi-source information into a single evaluation quantity.

[0062] Machine learning fusion: This approach employs pre-trained artificial neural networks or deep learning models, using multi-source sensor data as input and outputting fused environmental features. For example, deep neural networks can be used to obtain high-level features such as "current natural lighting availability score" and "current energy-saving optimization space," which can then guide optimization decisions. Training data can come from historical operational data, and the training process uses control actions that minimize energy consumption while achieving the desired comfort level as supervisory signals.

[0063] The output of the multi-source data acquisition and fusion module is a set of environmental state parameters used for decision-making. For example: comprehensive illuminance L. env Current occupancy status indicator ρ, indoor comfort deviation ΔC (difference relative to ideal illuminance and temperature), and current electricity price P. grid These fused parameters are updated at fixed intervals (e.g., every 5 minutes, or shorter intervals as needed) and passed to subsequent environmental prediction and optimization control modules.

[0064] The intelligent lighting control module is used to adjust the building's lighting devices and artificial lighting equipment according to the comprehensive environmental status information, so as to realize intelligent control of indoor illuminance. The intelligent lighting control module is connected to the adjustable lighting devices and lighting fixtures. By controlling the opening degree of the adjustable lighting devices and the output brightness of the lighting fixtures, the target illuminance is maintained by giving priority to natural light and supplementing artificial lighting when there is insufficient light, while avoiding glare.

[0065] The intelligent daylighting control module is responsible for dynamically adjusting the utilization of natural light and the output of artificial lighting indoors based on control decisions, in order to maintain the desired illuminance level and avoid unsuitable lighting environments. Its specific implementation includes the following aspects:

[0066] This invention preferably employs electrically adjustable light-collecting devices, such as electrically adjustable blinds, electric curtains, or smart dimming glass. The smart light-collecting control module receives instructions from the optimization decision module and controls the state of these light-collecting devices.

[0067] Motorized adjustable blinds / curtains: Adjust the opening angle of the blinds or the opening ratio of the curtains according to the required amount of natural light and anti-glare requirements. For example, open the curtains or increase the angle of the blinds to let more sunlight in when you need to increase indoor natural light; close them appropriately to reduce direct sunlight when glare occurs or the room temperature rises excessively.

[0068] Smart dimming glass: For windows using electrochromic (EC) glass, the light transmittance of the glass is controlled by adjusting the applied voltage. The smart lighting control module adjusts the glass transparency according to the control algorithm output, continuously changing between transparent and opaque, achieving precise control over the intensity of transmitted light.

[0069] The control of the lighting equipment adopts a closed-loop correction mechanism: the module reads real-time feedback from the indoor illuminance sensor, and if the illuminance does not reach the target after adjustment, the opening degree of the equipment is iteratively fine-tuned until the requirements are met or the equipment travel limit is reached. In addition, to avoid mechanical wear caused by frequent actions during the control process, dead zones and delays are set: adjustment is only performed when the illuminance deviation exceeds a certain threshold and persists for a period of time, in order to prevent frequent switching caused by short-term fluctuations such as cloud cover.

[0070] Lighting Fixture Dimming: Indoor lighting fixtures (such as LED downlights, fluorescent lamps, etc.) should support 0-100% stepless dimming or graded control. The intelligent lighting control module sends a dimming signal based on the required supplemental light, adjusting the output luminous flux of the fixtures via DALI or a 0-10V interface. The control strategy follows the principle of "prioritizing the use of natural light, supplemented by artificial lighting."

[0071] When there is sufficient outdoor light during the day to meet indoor illumination requirements, the lights are kept at low brightness or turned off, and environmental changes are monitored only.

[0072] When natural light is insufficient to achieve the set illuminance (L) set At that time, the required artificial light compensation ΔL = L is calculated. set -L env Increase the light output proportionally until the actual indoor illuminance (L) is reached. in Approaching L set If multiple sets of lights are controlled in zones, they can be adjusted separately according to the differences in each zone to achieve uniform illumination.

[0073] At night or in the absence of natural light, different strategies are adopted depending on whether the area is occupied or unoccupied: when occupied, maintain L set Illuminance: When no one is present, reduce the illuminance to a safe level or turn off the lights to save energy, leaving only the necessary guide lights on.

[0074] The intelligent lighting control module employs fuzzy reasoning control and expert rules during autonomous control to enhance its ability to handle complex situations. The specific control rules are as follows:

[0075] Fuzzy inference control: For the adjustment of illuminance and glare, fuzzy variables are introduced, such as "current indoor brightness is too dim / moderate / too bright", "outdoor light is too weak / strong", "glare risk is low / medium / high", etc., and a fuzzy rule base is established, for example:

[0076] Rule 1: If the indoor lighting is dim and the outside light is strong, then increase natural lighting (open the curtains / blinds);

[0077] Rule 2: If the indoor lighting is too dim and the outside light is too weak, then increase artificial lighting.

[0078] Rule 3: If the risk of glare is high, reduce natural light (close the curtains / dim the windows).

[0079] The intelligent lighting control module fuzzifies sensor data in real time and inputs it into a rule base for reasoning, resulting in a preliminary adjustment plan and improving robustness in response to complex changes in the lighting environment. The fuzzy control output, combined with precise PID dimming, balances rapid response and steady-state accuracy.

[0080] The intelligent lighting control module interacts with the environmental prediction and adaptation module. When it anticipates that outdoor sunlight will increase, it gradually reduces the brightness of the lights in advance to avoid overshoot; when it anticipates that sunlight will decrease or even that the sky will darken, it increases the brightness of the lights in advance to ensure a smooth transition. This feedforward control reduces the adverse effects of drastic fluctuations in illuminance on users and improves the stability of the lighting environment.

[0081] The system allows users to preset or learn their preferences for lighting environments, such as higher illuminance for certain tasks or a preference for a slightly darker environment. The intelligent lighting control module will adaptively adjust the lighting for different spaces or time periods while meeting basic standards. For example, in a conference presentation mode, it will automatically dim the ambient light, retaining only the necessary lighting, thus balancing energy saving and usage needs.

[0082] By combining the above multiple strategies, the intelligent daylighting control module can dynamically and intelligently manage indoor daylighting under various ambient light conditions, ensuring that indoor lighting is always in a comfortable and efficient state, and providing flexible daylighting adjustment means for energy synergy optimization.

[0083] The energy collaborative optimization control module is used to optimize the control of the building's lighting system and HVAC system by integrating the comprehensive environmental status information, so as to collaboratively reduce energy consumption. The energy collaborative optimization control module is based on a predetermined dual optimization objective function, and performs overall calculation on the lighting control output and the air conditioning control output to balance the lighting energy consumption and air conditioning energy consumption under the indoor illuminance requirements and temperature comfort requirements, so as to minimize the total energy consumption or operating cost.

[0084] The energy synergy optimization module is the core decision-making unit of this invention. It comprehensively considers the coupling relationship between the two major energy-consuming systems, lighting and air conditioning, to achieve optimal global energy efficiency. Its main functions and methods include:

[0085] The energy-coordinated optimization control module internally establishes an energy model for lighting-air conditioning coordination to evaluate energy consumption under different control settings. The model consists of the following sub-models:

[0086] Lighting energy consumption model: Lighting power is directly related to the dimming level and the number of lamps turned on. Let P be an example. L Indicates the power consumption of the lighting system, u L This indicates the lighting dimming control input (0-1 corresponds to 0%-100% output), then...

[0087] P L =P L,max ·u L ·N on

[0088] Where P L,max For the full output power of a single lamp, N on The number of lights currently on (which can automatically switch on and off depending on the area's occupancy), and the lighting energy consumption E. L It is P L Integrate over time or accumulate over discrete time steps. Note that lighting also generates sensible heat, increasing the indoor cooling load, which will be reflected in the air conditioning model later.

[0089] Air Conditioning Energy Consumption Model: Air conditioning energy consumption depends on the cooling / heating capacity required to maintain the indoor set temperature and the equipment efficiency. First, establish a heat balance:

[0090] Q HVAC =Q int +Q solar +Q out -Q loss

[0091] Q int It is the indoor heat load (heat dissipation from people, equipment, and lighting), Q solar The solar radiation heat enters through the window, Q out It is the heat load brought in by heat transfer from the outside (depending on the temperature difference between the outside and the inside), Q loss This refers to the heat loss (or gain) from ventilation and structural heat dissipation. The amount of cooling / heating that an HVAC system needs to provide to maintain a constant room temperature is Q. HVAC The balance value. Then consider the performance of the air conditioning equipment, using the coefficient of performance (COP) to convert cooling capacity into power consumption: if cooling COP = η cool Heating COP = η heat The power consumption of the air conditioner is

[0092]

[0093] This model characterizes the impact of daylight control on air conditioning load: when the opening and closing of the louvers changes Q... solar Or the light switch affects Q int At that time, P will change accordingly. HVAC Therefore, the energy consumption of lighting and air conditioning is not independent and needs to be optimized together.

[0094] Integrated Energy Consumption / Cost Model: Taking into account both lighting and air conditioning, the total energy consumption E is obtained. total =E L +E HVAC When time-of-use pricing is in place, total operating costs can also be calculated. Where π(t) is the electricity price at time t. The collaborative optimization module aims to reduce E... total Or C total The goal is to keep indoor environmental parameters within a comfortable range.

[0095] To incorporate the trade-off between comfort and energy consumption into mathematical optimization, this module designs a multi-objective optimization strategy, weighting illuminance deviation, temperature deviation, and energy consumption into a single objective function. For example, the objective function J can be defined as follows:

[0096] J=α×(I set -I in ) 2 +β×(T in -T set ) 2 +γ×(P L +P HVAC )

[0097] Among them I set I is the indoor illuminance setpoint. in T represents the measured indoor illuminance. set T is the air conditioner temperature setpoint. in P represents the actual indoor temperature. L and P HVAC These represent the power or energy consumption of the lighting and air conditioning systems, respectively. α, β, and γ are weighting coefficients for indoor illuminance deviation, indoor temperature deviation, and energy consumption, respectively, used to balance the importance of lighting comfort, thermal comfort, and energy consumption. Adjusting these weights can reflect different optimization focuses: for example, minimizing energy consumption while ensuring basic comfort, or prioritizing comfort during critical periods. The task of the collaborative optimization module is to select control actions (such as dimming commands) in real time. L Curtain opening degree u shade (e.g., adjusting the air conditioning temperature setting ΔT) to minimize the objective function J.

[0098] Another objective expression is the constrained optimization form, which uses illuminance and temperature comfort requirements as constraints and directly aims at minimizing energy cost:

[0099]

[0100] constraint:

[0101]

[0102] In this optimization model, the decision variables include the lighting dimming ratio u. L Shading device opening degree u shade Temperature control setting T setThe constraints ensure that the indoor illuminance is not lower than required and the temperature is within a comfortable range. The result of the optimization solution is the optimal combination of control variables in the next time period, which minimizes the total energy consumption. For multi-objective and multi-constraint problems, Pareto solutions can be obtained using weighted sum methods, ∈-constraint methods, etc., but this invention achieves the solution through the above-described comprehensive objective function form, which has higher computational efficiency.

[0103] The energy synergy optimization module needs to solve the above optimization problem under real-time requirements. The present invention preferably implements it in the following ways:

[0104] Model Predictive Control (MPC) Algorithm: Based on the established energy consumption model and indoor environmental state equations, it employs rolling time-domain optimization techniques. Specifically, at each control moment, using the current state as initial conditions, it predicts the environmental state evolution and energy accumulation within a future time range of H (the prediction time domain, where the next hour is divided into 12 five-minute steps). It then optimizes the sequence of lighting dimming, shading, and air conditioning settings over the entire time domain, ensuring the cumulative objective function... Minimum efficiency is achieved. The solution can be obtained using quadratic programming (QP) or mixed-integer programming (if discrete variables are involved). After obtaining the optimal sequence, only the control action at the current time step is executed, and then the rolling window moves forward to the next time step to repeat the calculation. In this way, MPC utilizes future information provided by the prediction module (such as illumination and load forecasts) to coordinate lighting and air conditioning strategies in advance, such as pre-cooling, pre-heating, or pre-dimming, to avoid peak energy consumption and maintain comfort and stability. Furthermore, MPC easily handles multivariate coupling and constraint problems, making it suitable for the needs of this invention.

[0105] Fuzzy optimization control: For scenarios with higher real-time requirements or where accurate models are difficult to obtain, this module can also employ fuzzy logic-based optimization methods. This involves directly providing control correction values ​​through fuzzy inference, without explicitly solving complex equations. For example, fuzzy rules can be designed such as: "If the total power is too high and the illuminance margin is sufficient, reduce the lighting output and increase the set temperature," or "If the illuminance is insufficient and the air conditioning load is low, increase the opening of the curtains." The fuzzy controller determines the control correction based on the current P... total The output relative to inputs such as the deviation from the historical average and the current illuminance margin (the amount by which illuminance exceeds the minimum requirement) is u. L ,u shade ,T set Adjustment suggestion Δu L ,Δu shade ,ΔT. This rule-based optimization is fast and robust, and can be used as a supplement to complex optimization or to take over control under abnormal conditions.

[0106] Combining precise optimization with intelligent algorithms is another approach. For example, genetic algorithms or reinforcement learning can be used offline to find the optimal coordination strategy for lighting and air conditioning in different typical environmental scenarios, and stored in a strategy library. During runtime, the optimal strategy can be approximated by looking up a table or by using a trained neural network. With lower complexity, the current control variable can also be directly optimized step-by-step using a genetic algorithm. However, considering real-time performance, using machine learning to approximate the optimization results online is more feasible. The preferred solution of this invention is MPC (Multi-Purpose Control), because it can combine prediction to achieve active control and can handle the inherent problem of multi-source data input and coupled output.

[0107] Guided by the energy synergy optimization module, this invention implements the following synergy control strategy:

[0108] When there are many people indoors and the lighting requirements are high, the system will open the shading to increase natural light and increase the fresh air volume (to prevent CO2 from rising due to the large number of people). However, if the monitored temperature rises, the air conditioning cooling capacity will be increased in advance. When strong sunlight causes the cooling load to increase, the curtains will be closed appropriately and artificial lighting will be used to supplement the lighting, reducing the air conditioning load. The status of lighting and shading is directly fed back to the air conditioning control, so that the two complement each other and work together to meet comfort needs while minimizing energy consumption.

[0109] Illumination and temperature standards are dynamically adjusted based on occupancy status and usage needs. For example, when a room is unoccupied, the illuminance is reduced to zero and the temperature setting is increased to save energy; when someone is detected entering, the lighting and comfortable temperature are quickly restored, and preparations are even made in advance (by predicting when someone will enter, such as the flow of people in adjacent rooms) to ensure a comfortable environment upon arrival. In meeting or presentation modes, lighting is reduced to enhance screen visibility, while the air conditioning temperature is appropriately increased to prevent people sitting for long periods from getting cold. These systems work together to optimize specific scenarios.

[0110] During periods of high electricity prices, the system tends to sacrifice some comfort margins to save costs. For example, during peak daytime hours, the air conditioning temperature setting is appropriately increased by 0.5°C and the illuminance setting is reduced by 10%, in exchange for a 20% reduction in power consumption. Conversely, during off-peak hours, the building structure is pre-cooled at night to reduce daytime cooling demand the following day. This scheduling optimizes operating costs while ensuring basic comfort. If the building has renewable energy sources such as solar power, the system can also be considered, utilizing air conditioning for cooling or increasing illuminance when solar power is abundant, and strictly conserving energy when solar power is insufficient, thus achieving source-load synergy.

[0111] When sensors malfunction or data becomes abnormal, the collaborative optimization module activates fault-tolerant strategies, such as ignoring obviously erroneous data or using the most recently trusted value. When encountering extreme weather (heavy rain, extreme darkness, or scorching heat with intense sunlight) that exceeds the normal model range, the system automatically switches to a safety mode: for example, forcibly closing the sunroof during heavy rain to prevent water ingress, and running the air conditioner at full capacity during extreme heat to prevent damage to equipment. These safeguards ensure the system operates safely under all conditions.

[0112] Through the above method, the energy collaborative optimization module continuously calculates and outputs the optimal control scheme in a complex and ever-changing environment, guiding the lighting and air conditioning execution modules to make coordinated adjustments. Ultimately, this achieves the goal of ensuring both lighting and thermal comfort while minimizing total energy consumption and costs. Actual measurements show that, compared to traditional independent control, the collaborative control of this invention can effectively reduce overall building energy consumption by more than 20%, smooth fluctuations in environmental parameters, and significantly improve the energy management level of green buildings.

[0113] An environmental prediction and adaptation module is used to predict future environmental conditions based on historical data and external prediction information and dynamically adjust control strategy parameters. The environmental prediction and adaptation module includes models for predicting changes in outdoor light intensity and indoor heat load, as well as a mechanism for adaptively correcting control model parameters based on real-time feedback.

[0114] The environmental prediction and adaptation module, as an auxiliary decision-making unit, runs through the entire control process, providing the system with forward-looking environmental information and self-adjustment capabilities.

[0115] Utilizing advanced data-driven models to predict future environmental conditions, primarily including two aspects: illumination and load forecasting.

[0116] Natural light prediction: Based on weather forecasts, solar position models, and current cloud cover, this forecast predicts outdoor light intensity and available sunshine duration for the near future (e.g., hourly within one hour, or even hourly within one day). For example, a radiative transfer model can be used to calculate solar radiation under given meteorological conditions, and then the transmitted light entering the room can be calculated based on window orientation and time. Machine learning methods can also be employed, such as training a regression model or LSTM neural network to predict future curves using historical light data from similar days. Prediction results are provided in time series format, such as the expected outdoor illuminance L every 5 minutes. out,forecast (t), which provides feedforward information for intelligent daylighting control, enabling lighting adjustment to be proactive, such as turning lights on or off in advance.

[0117] Heat Load and Temperature Forecasting: Considering the significant thermal inertia of buildings, predicting indoor temperature trends in advance is crucial for optimizing air conditioning. The model, based on the heat balance equation, uses current air conditioning operating status, occupancy plans, equipment heat dissipation estimates, and outdoor temperature curves from weather forecasts to predict future room temperature changes and air conditioning load demand. For example, a state-space model can be established:

[0118]

[0119] Where T in (t) represents the indoor air temperature at time t, T out(t) represents the outdoor air temperature at time t, Δt represents the discrete time step, C represents the indoor equivalent heat capacity (including heat storage in air, furniture, and interior surfaces), K represents the overall heat transfer coefficient, ρ represents the air density, and c p V represents the specific heat capacity of air, and V represents the volume of air inside the room.

[0120] Similar models can be used for internal predictions in MPC, or approximate models can be trained using statistical or machine learning methods for real-time predictions. If the temperature is predicted to rise to the comfort limit, the co-optimization module can increase cooling in advance; conversely, if the room temperature is predicted to drop, the air conditioning output can be reduced in advance to save energy.

[0121] Optionally, occupancy can also be predicted, such as using daily patterns to predict the probability distribution of occupancy at different times of the day in office buildings. In some scenarios, changes in CO2 concentration and humidity can also be predicted, but mainly light and temperature, as they have the greatest impact on control decisions.

[0122] like Figure 4 The diagram shows the probability of occupancy rate prediction. The horizontal axis represents each hour of the day, the solid line represents the real-time occupancy observation O(t), and the broken line represents the double exponential smoothing prediction P. occ (t) demonstrates the system's ability to adjust lighting and air conditioning strategies in advance through occupancy prediction, reducing blind responses and energy waste.

[0123] In reality, model parameters and environmental characteristics change over time, such as seasonal changes and equipment aging. The adaptive module maintains the accuracy of the model and control through online learning and parameter adjustment.

[0124] For key parameters in energy consumption and prediction models (such as window light transmittance, room heat capacity C, and heat transfer coefficient K), corrections are made using the discrepancies between actual measured data and model predictions. For example, comparing the measured increase in room temperature with the model's calculated value, if there is a long-term discrepancy, the K or C values ​​are adjusted. Another example is estimating the current cloud cover factor in the illumination model, updating it based on the ratio of measured outdoor illuminance to the theoretical value for clear skies. This ensures the model remains consistent with reality year-round, improving prediction and optimization accuracy.

[0125] The system can record a large amount of operational data, including environmental conditions, control outputs, and user feedback (if any, such as records of human intervention when adjustments are unsatisfactory). Using reinforcement learning or policy gradient methods, with rewards for minimizing energy consumption and satisfying comfort levels, a policy network or the parameters of fuzzy rules are trained, thereby gradually improving the control policy library. Over time, the system's understanding of specific building characteristics becomes increasingly precise, and its control behavior becomes more optimized. For example, it learns that in an office facing a certain direction, the curtains should be opened slightly later at sunrise to avoid direct, glaring sunlight, and then gradually opened. This fine-tuning strategy can be summarized and applied automatically by the adaptive module through analysis of historical effects.

[0126] The adaptive module also acts as a monitoring module, using data fusion results to detect anomalies (such as sensor malfunctions or data jumps). When an anomaly is detected, it triggers backup strategies or uses redundant data sources to ensure the robustness of control decisions. Identified invalid data segments are not included in the learning process to prevent errors from affecting the model.

[0127] The environmental prediction and adaptive modules do not operate independently but are closely integrated with optimal control. Predictive results are input into the optimization algorithm, guiding its decisions to be more forward-looking; the optimized output and actual results are fed back to adaptively adjust the model, forming a closed loop. A typical loop is as follows:

[0128] At time t0, the data fusion module provides the current state, the prediction module provides the environmental prediction from t0 to t0+H, and the adaptive module updates the model parameters.

[0129] Based on the above information, the optimization module calculates the sequence of control actions starting at time t0, and the execution module applies the first control action;

[0130] When time enters t1 = t0 + Δt, the perception layer acquires the new state, the system records the effect of the action at t0 and compares it with the prediction. If the deviation is significant, the adaptive module adjusts the model or policy parameters at t1.

[0131] Then proceed to the next cycle, and continue in this manner.

[0132] Through a long-term prediction-feedback-adjustment mechanism, the system of this invention can adapt to slow changes in building operation (such as seasonal changes in sunlight angle) and maintain stable performance under short-term disturbances (such as sudden changes in pedestrian flow or equipment failure). The environmental prediction and adaptive module endows the system with the ability to "plan ahead" and "self-evolve," making it more intelligent and reliable than traditional fixed-value control, truly meeting the requirements of green buildings for intelligence and high adaptability.

[0133] The control decision and execution module is used to receive lighting control and air conditioning control commands output by the energy collaborative optimization control module and send them to the execution device to adjust the working status of indoor lighting fixtures and air conditioning equipment. The control decision and execution module includes a lighting execution unit connected to adjustable lighting equipment and lighting fixtures, and an air conditioning execution unit connected to the HVAC system, and controls indoor illuminance and temperature in conjunction with the optimal control decision.

[0134] like Figure 2 As shown, the present invention also provides a method for intelligent daylighting and energy synergy optimization in green buildings based on multi-source data fusion, comprising the following steps:

[0135] S1. Collect multi-source data related to indoor and outdoor lighting, occupancy, meteorology and energy, preprocess and time-align the collected data, and use data fusion algorithms to generate comprehensive characterization parameters of the current environmental state;

[0136] The fusion of multi-source data in step S1 includes: acquiring the illuminance at various indoor points and calculating the average or minimum illuminance value; acquiring the outdoor horizontal illuminance and converting it into available indoor illuminance based on window area and light transmittance; combining the occupancy sensor status to determine room usage needs; and comprehensively forming a light environment index for lighting decisions. Simultaneously, acquiring the current air conditioning operating parameters and indoor and outdoor temperature and humidity; estimating the current building heat load status; fusing multi-sensor readings through Kalman filtering to obtain smoothed illuminance and temperature values; and incorporating occupancy information and comfort requirements into the environmental index through fuzzy logic to obtain a fuzzy comprehensive evaluation value describing the current light and heat environment.

[0137] S2. Based on the comprehensive characterization parameters of the current environmental state and the obtained external prediction information, predict the changes in outdoor light intensity, indoor thermal environment, or room occupancy within a predetermined period in the future.

[0138] S3. Establish a coordinated energy consumption model for lighting and air conditioning and optimize the objective function. Based on the comprehensive characterization parameters and prediction results of the current environmental state, use the optimization algorithm to calculate the optimal combination decision of lighting control and air conditioning control under the constraints of indoor illuminance and thermal comfort, including the opening degree of adjustable daylighting equipment, the dimming level of lighting fixtures and the set control amount of air conditioning temperature.

[0139] The optimization algorithm in step S3 adopts the model predictive control method, which specifically includes: establishing a prediction model with the current state as the initial state in each control cycle; discretizing candidate lighting and air conditioning control sequences in the prediction time domain; calculating the overall optimization target value by simulating the changes in indoor illuminance and temperature and the accumulation of energy consumption under the action of these control sequences; selecting the control sequence that minimizes the target value as the optimal solution; and executing only the control action of the current cycle in the controlled sequence. In the next cycle, the optimization calculation is repeated in a rolling manner using new state feedback and prediction information.

[0140] S4. Send the optimal combination decision to the corresponding lighting and air conditioning actuators to adjust the operating status of the indoor lighting and air conditioning systems, thereby achieving coordinated control of the indoor light and thermal environment.

[0141] S5. Monitor the actual indoor illuminance and temperature feedback information after step S4 is executed, and compare it with the predicted value or set value. If there is a deviation, adaptively adjust the model parameters or control strategy in subsequent control cycles, including updating data fusion weights, optimizing algorithm parameters or fuzzy rules, so that the system can adapt to new environmental characteristics.

[0142] The adaptive adjustment in step S5 includes: when the actual indoor illuminance deviates from the target value for a long period of time, adjusting the gain parameter of the illuminance control loop or relearning the lighting energy consumption model; when the actual temperature control overshoot or lag is detected, adjusting the PID parameter of the air conditioning control or the heat transfer coefficient of the prediction model; when environmental or load mode changes cause the original fuzzy rules to no longer be applicable, automatically activating the previously trained and stored rule set for the new mode or using a fuzzy adaptive algorithm to generate new control rules, thereby ensuring that the system can maintain efficient and stable control performance under different environmental conditions.

[0143] S6. Repeat steps S1 to S5 to form a cyclical control process to continuously optimize indoor lighting utilization and energy consumption.

[0144] Example

[0145] A 20-square-meter office with a large south-facing window (with adjustable motorized blinds), equipped with four LED fluorescent lights, and central air conditioning to maintain the temperature. During daytime working hours, the required indoor work surface illuminance is no less than 500 lx, and the temperature is maintained within a comfortable range of 22–26℃, while minimizing energy consumption. We optimized the calculations for the scenario of abundant sunlight at midday.

[0146] The current indoor illuminance L is obtained through multi-source data fusion. in =300lx (lower due to partial obstruction by curtains), outdoor horizontal illuminance (L) out =80000lx (noon on a sunny day), indoor temperature T in =24.0℃, occupied, air conditioner temperature setting T set =24℃, at this moment the air conditioner's cooling power P HVAC =1.5kW, lighting power P L = 0.8kW (approximately half of the lights are on).

[0147] The system's co-optimization objective is to achieve an illuminance ≥ 500 lx and a temperature within a comfortable range, while simultaneously reducing the total power P. total =2.3kW. Optimization variables include venetian blind opening θ (0 fully closed, 1 fully open) and lighting dimming ratio u. L (0-1), and the air conditioning temperature setting △T can be slightly adjusted within the range of ±1℃. Define the objective function:

[0148] J = γ1·P L +γ2·P HVAC

[0149] Here, the energy consumption of γ1 and γ2 is normalized to economic cost or carbon emission weight, and is tentatively set to be equal to 1, because comfort is a hard constraint: illuminance ≥500lx, temperature 2226℃.

[0150] The energy synergy optimization module uses models to predict and evaluate the effects of different combinations of actions on P. L and P HVAC Impact:

[0151] If natural lighting is increased: increasing θ from the current 0.5 (half-open) to 0.8 (with curtains open more), the expected indoor illuminance (L) will be [value missing]. in It can be increased to over 500 lx, thus illuminating P L It can be reduced to 0.4kW (turning off half the lights). However, at the same time, solar radiation heat increases, and P is estimated to be higher. HVAC It will rise to 1.8kW to maintain the temperature, then P total Approximately 2.2kW, slightly lower than before.

[0152] If you keep the curtains closed, increase the lighting: maintain θ at 0.5, u L Increasing the current value from 0.5 to 0.8 (with all lights on) will ensure the required illuminance. At this point, P... L Increased to 1.6kW, while P remains unchanged due to the unchanged shading. HVAC Maintaining 1.5kW, the total power is 3.1kW, which is significantly higher and not optimal.

[0153] If the curtains are fully open and the temperature is appropriately increased (θ = 1.0, maximum daylight), the indoor illuminance can theoretically reach 600 lx, eliminating the need for artificial lighting (P). L ≈0). At this time, the solar heat increase is at its maximum, and the room temperature may rise. The air conditioner needs to increase its power to about 2.5kW to maintain 24℃, with a total power of 2.5kW. However, if the air conditioner temperature setting is raised to 25℃ (△T=+1℃ within the allowable range), the air conditioner power can be reduced to about 2.0kW. The room temperature will rise slightly but remain within the comfortable range, with a total power of about 2.0kW, which is the lowest option.

[0154] like Figure 3 The diagram shows the effect of shading opening on light and heat gain. The horizontal axis represents the shading device opening θ (0-1), and the vertical axis represents the normalized natural light utilization and the relative amount of solar heat gain. Figure 3 This visually demonstrates the coupling trend between daylighting benefits and heat load at the same louver angle, illustrating why coordinated optimization is needed between shading, lighting, and air conditioning.

[0155] In summary, the combination of "increasing natural light + slightly raising the temperature" has the lowest total energy consumption and acceptable comfort among all options. Based on this, the optimization algorithm arrives at an approximate optimal solution: open the blinds to over 80%, turn off most lights, leaving only a small amount of light to compensate for shadow areas, and simultaneously raise the air conditioning temperature by 1°C to 25°C. This maintains an illuminance of approximately 500–550 lx and a room temperature of 25°C. The new P0 value is then calculated. L ≈0.2kW (almost zero lighting energy consumption), P HVACThe power output is approximately 1.9kW, and the total power is approximately 2.1kW, which is about 9% lower than the initial state.

[0156] The control decision module sends commands to the intelligent lighting module to open the curtains to 80%, dim the lights to 20%, and set the temperature to 25℃ to the HVAC system. Five minutes after execution, the sensors report an indoor illuminance of 510 lx and a temperature of 25.0℃, with no discomfort reported by occupants. The system detects that the performance meets expectations and records the control strategy and its effects for the adaptive module to learn from. If clouds suddenly block the sun, causing an illuminance drop of less than 500 lx, the environmental prediction module will update its forecast, and the collaborative optimization module will increase the light output in the next control cycle to compensate and maintain stable illuminance.

[0157] This example demonstrates that the control algorithm of this invention dynamically seeks the optimal balance point by optimizing the combination of natural light utilization, artificial light sources, and air conditioning settings, thereby maximizing energy efficiency. It simultaneously meets the control requirements for illuminance and temperature, achieving a true "dual optimization" goal.

[0158] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A green building intelligent lighting and energy synergistic optimization system based on multi-source data fusion, characterized in that, include: The multi-source data acquisition and fusion module is used to acquire multiple data sources from inside and outside the building, and to fuse the data to generate comprehensive environmental status information for control decisions. The multiple data sources include at least indoor illuminance, outdoor illuminance and meteorological data, indoor occupancy information, indoor environmental parameters and electricity price data. The intelligent lighting control module is used to adjust the building's lighting devices and artificial lighting equipment according to the comprehensive environmental status information to achieve intelligent control of indoor illuminance. The intelligent lighting control module establishes a fuzzy variable and fuzzy rule base that includes indoor brightness, outside light, and glare risk. It fuzzifies the sensor data in real time and inputs it into the rule base to infer a preliminary adjustment scheme. The fuzzy control output is combined with PID dimming. The energy-coordinated optimization control module is used to optimize the control of the building's lighting and HVAC systems by integrating the comprehensive environmental state information. It establishes a lighting-air conditioning coordinated energy model, which includes a lighting energy consumption sub-model and an air conditioning energy consumption sub-model. The air conditioning energy consumption sub-model is based on the building heat balance equation and uses solar radiation heat entering through windows as a coupling variable with daylighting control, reflecting the direct impact of changes in daylighting device opening degree on air conditioning load. The energy-coordinated optimization control module, based on a predetermined dual optimization objective function, weights and combines indoor illuminance deviation, indoor temperature deviation, and energy consumption. The system comprehensively calculates the output of lighting control and air conditioning control to balance the energy consumption of lighting and air conditioning under indoor illuminance requirements and temperature comfort requirements, so as to minimize the total energy consumption or operating cost. The energy collaborative optimization control module adopts a model predictive control algorithm. Based on the lighting-air conditioning collaborative energy model and the indoor environment state equation, it predicts the evolution of the environmental state and the energy consumption accumulation within the next H time range at each control moment, using the current state as the initial condition. It performs rolling time-domain optimization solution for the control sequence of lighting dimming, shading and air conditioning settings over the entire prediction time domain, and only executes the control action at the current moment and then rolls forward to repeat the calculation. The environmental prediction and adaptation module is used to predict the future environmental state based on historical data and external prediction information and dynamically adjust the control strategy parameters. The environmental prediction and adaptation module uses outdoor weather forecast data and indoor sensor historical data to use machine learning models to predict the outdoor light intensity, indoor cooling load demand or personnel occupancy within a predetermined time interval in the future. By monitoring the deviation between the actual control effect and the prediction, it automatically corrects the parameters of the energy consumption model or fuzzy control rules. The control decision and execution module is used to receive lighting control and air conditioning control commands output by the energy collaborative optimization control module and send them to the execution devices to adjust the working status of indoor lighting fixtures and air conditioning equipment. It also receives execution feedback in real time to ensure that the commands are executed in place and to perform closed-loop verification. Indoor illuminance and temperature are controlled in conjunction with the optimal control decision.

2. The green building intelligent lighting and energy synergistic optimization system based on multi-source data fusion according to claim 1, characterized in that, The multi-source data acquisition and fusion module includes an indoor illuminance sensor, an outdoor light sensor, an indoor occupancy sensor, temperature and humidity sensors, and a data interface for acquiring weather forecasts and electricity price information. The module filters, synchronizes, and normalizes data from different sources, fuses multi-sensor readings using Kalman filtering to obtain smooth illuminance and temperature values, and incorporates occupancy information and comfort requirements into environmental indicators using fuzzy logic to obtain a fuzzy comprehensive evaluation value describing the current light and heat environment. The intelligent lighting control module connects adjustable lighting equipment and lighting fixtures. By controlling the on / off state of the adjustable lighting equipment and the output brightness of the lighting fixtures, it prioritizes the use of natural light and supplements artificial lighting when natural light is insufficient to maintain the target illuminance, while avoiding glare.

3. The green building intelligent lighting and energy synergistic optimization system based on multi-source data fusion according to claim 2, characterized in that, The control decision and execution module includes a lighting execution unit connected to adjustable lighting equipment and lighting fixtures, and an air conditioning execution unit connected to the HVAC system. It controls indoor illuminance and temperature in conjunction with the optimal control decision. The lighting execution unit controls the output of dimmable lighting fixtures through DALI or 0-10V interface and controls the opening and closing of electric curtains or adjustable blinds through motor drive. The air conditioning execution unit sends temperature setting, valve opening or fan speed commands to the HVAC system through BACnet or MODBUS interface.

4. A method for intelligent daylighting and energy synergistic optimization in green buildings based on multi-source data fusion, characterized in that, Includes the following steps: S1. Collect multi-source data on indoor and outdoor lighting, room occupancy, weather and energy, preprocess and time-align the collected data, and use data fusion algorithms to generate comprehensive characterization parameters of the current environmental state. S2. Based on the comprehensive characterization parameters of the current environmental state and the obtained external prediction information, predict the changes in outdoor light intensity, indoor thermal environment, or room occupancy within a predetermined period in the future. S3. Establish a coordinated energy consumption model for lighting and air conditioning and optimize the objective function. Based on the comprehensive characterization parameters and prediction results of the current environmental state, use the optimization algorithm to calculate the optimal combination decision of lighting control and air conditioning control under the constraints of indoor illuminance and thermal comfort, including the opening degree of adjustable daylighting equipment, the dimming level of lighting fixtures and the set control amount of air conditioning temperature. The optimization algorithm in step S3 adopts the model predictive control method, which specifically includes: establishing a prediction model with the current state as the initial state in each control cycle, discretizing candidate lighting and air conditioning control sequences in the prediction time domain, calculating the overall optimization target value by simulating the changes in indoor illuminance and temperature and the accumulation of energy consumption under the action of these control sequences, selecting the control sequence that minimizes the target value as the optimal solution, and only executing the control action of the current cycle in the controlled sequence, and repeatedly optimizing the calculation in the next cycle using new state feedback and prediction information. S4. Send the optimal combination decision to the corresponding lighting and air conditioning actuators to adjust the operating status of the indoor lighting and air conditioning systems, thereby achieving coordinated control of the indoor light and thermal environment. S5. Monitor the actual indoor illuminance and temperature feedback information after step S4 is executed, and compare it with the predicted value or set value. If there is a deviation, adaptively adjust the model parameters or control strategy in subsequent control cycles, including updating data fusion weights, optimizing algorithm parameters or fuzzy rules. The adaptive adjustment in step S5 includes: when the actual indoor illuminance deviates from the target value for a long period of time, adjusting the gain parameter of the illuminance control loop or relearning the lighting energy consumption model; when the actual temperature control overshoot or lag is detected, adjusting the PID parameter of the air conditioning control or the heat transfer coefficient of the prediction model; when environmental or load mode changes cause the original fuzzy rules to no longer be applicable, automatically activating the previously trained and stored rule set for the new mode or using a fuzzy adaptive algorithm to generate new control rules to ensure that the system can maintain efficient and stable control performance under different environmental conditions. S6. Repeat steps S1 to S5 to form a cyclical control process to continuously optimize indoor lighting utilization and energy consumption.

Images