CN118870623A - Building intelligent illumination control method and system based on BIM and digital twin technology - Google Patents

Abstract

The invention discloses a building intelligent illumination control method and system based on BIM and digital twin technology, which relate to the technical field of intelligent illumination and comprise the steps of obtaining a BIM space model diagram comprising a building space structure and illumination equipment distribution; dividing the building space into different types of areas according to lighting conditions; selecting a control mode to calculate opening values of lighting equipment in different areas; and adjusting the on-off states of the lighting equipment in different areas according to the difference value of the opening value and the set value to realize intelligent illumination control of the building. The intelligent illumination control system for the building based on the BIM+ digital twin technology comprises a data acquisition module, a central processing module, an RFID tag module and a calling module. The invention utilizes the building BIM model to truly reflect the distribution position and the spatial distribution condition of each lighting device in the building area, realizes the spatial point position of each lighting device, and combines the RFID technology to realize the double consideration of the brightness adjustment and the energy-saving lighting of the high-efficiency lighting device.

Description

Building intelligent illumination control method and system based on BIM and digital twin technology

Technical Field

The invention relates to the technical field of intelligent illumination, in particular to a building intelligent illumination control method and system based on BIM and digital twin technology.

Background

At present, indoor illumination of buildings generally comprises a plurality of illumination lamps, wherein the illumination lamps are arranged at different positions of the same room so as to illuminate the whole indoor area. However, when no person or only one to two persons exist in the building, unnecessary waste is caused if all the illuminating lamps are all lighted; if the functions of turning on and off the light by people and distributing the illumination brightness according to the needs are automatically realized by combining the internet of things, expensive infrared sensors are required to be installed at all indoor positions so as to detect whether people exist at the corresponding positions, and therefore the problems of difficult installation and extremely high hardware cost exist. Meanwhile, in an office area part in a building, due to the fact that the concentration of people is large and various use states exist, the conventional illumination adjustment method cannot adapt to the light brightness adjustment requirement of the office area, and in addition, the existing indoor illumination control technology cannot balance between the illumination requirement and the illumination energy consumption, so that the energy efficiency is difficult to improve.

Chinese patent document CN202311423512.4 discloses "a method, apparatus, device and storage medium for controlling intelligent lighting in building room". The method comprises the steps of acquiring field video data acquired by a building indoor monitoring camera in real time; according to the field video data, tracking each indoor person in the building room in real time by adopting a multi-target tracking algorithm to obtain an indoor person tracking result; determining the recent activity areas of all indoor personnel in real time according to the indoor personnel tracking result, and taking the union of the recent activity areas of all indoor personnel as a current illumination target area; extracting video frame images from the live video data in real time, and converting the video frame images into images to be processed, wherein the color space of the images is HSV; and carrying out binarization processing on the image to be processed according to a preset target brightness threshold value to obtain a binarized image containing at least one highlight image area, wherein the highlight image area is an image area with brightness values of all pixel points in the area being greater than or equal to the target brightness threshold value. However, the patent cannot solve the illumination adjustment requirement of the office area in the building, cannot meet the illumination change of personnel in the office area under different states, and cannot realize the aim of energy-saving illumination.

Disclosure of Invention

The invention mainly solves the technical problems of illumination requirements and illumination energy consumption in the original building office area, and provides a building intelligent illumination control method and system based on BIM and digital twin technology.

The technical problems of the invention are mainly solved by the following technical proposal: the invention includes

S1, acquiring a BIM space model diagram comprising a building space structure and lighting equipment distribution;

S2, dividing the building space into different types of areas according to lighting conditions;

s3, selecting a control mode to calculate opening values of lighting equipment in different areas;

s4, adjusting the on-off states of the lighting equipment in different areas according to the difference value of the opening value and the set value to realize intelligent illumination control of the building.

Preferably, the control mode comprises a daytime mode, and the brightness of the lighting equipment is adjusted according to the lighting condition of the divided areas to realize brightness gradient adjustment; in the noon break mode, the sunshade curtain and the lighting equipment of the office area are closed in the set noon break time; the night shift mode uniformly adjusts all areas to set brightness, and if overtime staff exists, the digital twin technology and the human body detector are combined to monitor and accurately start the lighting equipment of the part where the overtime staff is located; patrol mode, interval turns on corridor lights and turns off lighting of the rest of the area.

Preferably, when different types of areas are divided, the station distribution condition is considered, the space in the building is divided in a regular graph, and the types of the different areas are classified according to the space positions of the different areas and the lighting states acquired by the acquisition equipment.

Preferably, the region type comprises a region with better illumination, and the illumination brightness in the region is larger than a set value x; the illumination area is poor in illumination, and the illumination brightness in the area is smaller than a set value x; the window area comprises a window, and brightness change can be regulated by linking a roller shutter arranged on the window when brightness in the area is regulated.

Preferably, in the step S4, the illumination devices in the divided areas are labeled with RFID tags according to the location and type of the areas and the type of the devices, and when the intelligent illuminance control of the building is performed, the illumination devices corresponding to the illumination devices are adjusted by reading the RFID tags.

Preferably, the opening value is calculated by:

s31, acquiring average illumination and design illumination in the region p through acquisition equipment, and comparing the difference value to calculate illumination supported by illumination equipment required in the region p;

S32, acquiring the area of the region p and the number of the lighting equipment in the region through the BIM space model diagram;

s33 calculates the opening value of the lighting device from the calculated and counted related parameters within the region p.

Preferably, the opening value is calculated as follows:

ΔEp=E-Ep，

wherein E is the design illuminance, and Ep is the average illuminance in the region p;

Ep=（Ea+Eb+Ec+Ed）/4

ea, eb, ec and Ed are real-time parameters of points a, b, c and d in the acquisition area of the acquisition equipment;

，

The coordinate of the acquisition point a is a (x/y/z), the coordinate of the acquisition point b is b (x/y/z), the coordinate of the acquisition point c is c (x/y/z), the coordinate of the acquisition point d is d (x/y/z), and the acquisition point b is the area A of the region p;

，

Wherein the method comprises the steps of The aperture value of the lamp in the region p is phi, the luminous flux of the lighting equipment, U is the utilization coefficient of the lamp, and K is the maintenance coefficient of the lamp.

Preferably, the daytime mode is adjusted as follows:

Wherein E is the average illuminance,

Calculating to obtain real-time average illuminance, comparing a preset illuminance value with the area average illuminance of a fixed area, calculating the opening value of the brightness of the area lamp on the premise of setting stable quantity of the area lamp, and regulating the brightness according to the illumination requirement of the area in daytime.

The intelligent illumination control system for the building based on the BIM+ digital twin technology comprises a data acquisition module, a plurality of illumination acquisition and transmission devices and a central processing module, wherein the plurality of illumination acquisition and transmission devices are connected with the central processing module; the central processing module is used for carrying out centralized processing on parameters related to illuminance value calculation in different areas and confirming an adjustment scheme according to a processing result; the RFID tag module is used for setting different lighting equipment to characterize RFID tags and assisting the central processing module in calling; and the calling module is used for calling the corresponding lighting equipment to adjust according to the adjustment scheme of the central processing module and the marking of the RFID tag module.

Preferably, the RFID tag module comprises a plurality of RFID tags, different RFID tags are correspondingly bound with the lighting equipment components in the BIM space model, and unique IDs are arranged on the different lighting equipment components of the BIM space model; the cloud computing server is used for retrieving real-time energy consumption information of the lighting equipment in the current time period from a relational database in the regional controller, inputting the real-time energy consumption information, the building body attribute information and the real-time environment information in the building into the lighting equipment energy consumption prediction model together, and predicting an energy consumption predicted value of the lighting equipment in the next time period.

The beneficial effects of the invention are as follows: the invention is based on BIM and digital twin technology, utilizes the building BIM model to truly reflect the distribution position and the spatial distribution condition of each lighting device in the building area, realizes the spatial point position of each lighting device, combines the illumination requirement and the brightness change of the area to assist a manager to quickly adjust the brightness of each area, and improves the brightness adjustment efficiency and the brightness adjustment accuracy in the building. In addition, the invention also combines the RFID technology, automatically updates the related information of the lighting equipment in the BIM model in real time through the RFID radio frequency technology, displays the running condition of the real-time lighting equipment and effectively tracks the energy data by combining the BIM model, improves the prediction precision and the control effect of the building energy consumption on the basis of comprehensively considering the attribute information of the building body and the real-time environment information in the building, and realizes the double consideration of the brightness adjustment and the energy-saving lighting of the high-efficiency lighting equipment.

Drawings

Fig. 1 is a flow chart of an illumination control mode selection according to the present invention.

Fig. 2 is a flow chart of a lamp brightness adjustment according to the present invention.

Fig. 3 is a flow chart of brightness adjustment in night shift mode of the present invention.

Detailed Description

The embodiment 1 details a building intelligent illumination control method and system based on BIM and digital twin technology. The inventive method aims to solve the problems existing in the existing building lighting control system and provides a more refined and intelligent lighting management solution.

Limitations of existing architectural lighting control. Traditionally, architectural lighting control has been largely divided into two ways: manual control and automatic control. Manual control relies on a worker manually switching the lighting circuits of different areas through an area control panel or operator station. Although flexible, the method is low in efficiency and is easily influenced by human factors, and continuous optimal control cannot be ensured. On the other hand, automatic control typically employs a time controller to time switch the lighting loop. While this approach increases the degree of automation, it ignores the complexity of the building interior structure and the differences in lighting conditions and time of use in different areas. The unified management mode can not meet the personalized illumination requirements of users in different areas, so that energy waste and user experience are reduced. Aiming at the problems, the embodiment provides an intelligent illumination control method based on BIM and digital twinning technology. The method is characterized in that the interior space of the building is finely divided and characterized so as to realize more accurate and personalized illumination control. The first step in the implementation is to conduct a comprehensive and detailed on-site survey of the target building. The importance of this step is not neglected as it directly affects the accuracy and reliability of the subsequent BIM model. The investigation process involves not only the measurement and recording of the physical structure of the building, but also the in-depth analysis of the lighting environment. In particular, survey teams need to install light harvesting devices in different areas of the building that can monitor and record the light intensity changes at various locations in real time. Meanwhile, in order to better understand the space use condition, the team also needs to arrange the human body detector according to the distribution of office positions STRATEGICALLY. Typically 2-3 detectors are configured per space, and this redundancy arrangement can effectively prevent data loss or misinterpretation caused by a single detector failure. After the field survey is completed, the next step is to construct a BIM space model of the target building. This process requires combining the actual measurement data obtained in the survey with BIM techniques to create a highly accurate virtual building model. The model not only contains geometric information of a building, but also integrates key information such as distribution of lighting equipment, window positions, installation positions of various sensors and the like. After BIM model construction is completed, the system can intelligently partition the inner space of the building by combining real-time illumination data obtained by the acquisition equipment. This zoning process takes into account a number of factors including natural illumination intensity, artificial lighting distribution, window position, etc. Specifically, the system divides the area with the illumination brightness greater than the preset threshold value x into a "better illumination area", and divides the area with the illumination brightness less than the threshold value x into a "worse illumination area". For areas close to the window, the system will be particularly marked as "critical areas" because the lighting conditions of these areas are not only affected by the external environment, but also vary as a function of the open and closed state of the window and the use of roller blinds. To achieve finer control, the system would divide the building interior space further into multiple small area units. The refined dividing method can remarkably improve the accuracy of illumination control. During the dividing process, the system uses a regular geometry as much as possible and ensures that the user's position is at the center of the dividing area. The layout is convenient for subsequent area calculation and lamp quantity statistics, and can ensure uniformity and comfort of lighting effect. Furthermore, the system is also particularly concerned with the location of the placement of the acquisition devices. To accurately reflect the natural illumination of each region, the acquisition device is typically placed at a boundary point of the divided region, not at a central position. This is because if the collection device is located in the center of the area, it is easily affected by artificial illumination, and thus the natural illumination condition of the area cannot be accurately reflected. If the acquisition device is found to be in the center of the region during the partitioning process, the system may optimize the accuracy of the data acquisition by adjusting the region partitioning or rearranging the acquisition device. The intelligent illumination control system based on the BIM and the digital twin technology not only can realize more accurate illumination management, but also can dynamically adjust the illumination strategy according to real-time data. For example, in a window area, the system can assist in adjusting indoor brightness by controlling curtains or roller blinds in a linked manner, so that the glare problem caused by direct irradiation of natural light is avoided, and meanwhile, natural light resources are utilized to the maximum extent. In addition, the system has strong expandability and adaptability. With the continuous development of the Internet of things technology and artificial intelligence algorithms, the system can integrate more types of sensor data such as temperature, humidity, air quality and the like, so that more comprehensive indoor environment control is realized. Meanwhile, through a machine learning algorithm, the system can continuously optimize the control strategy, learn the preference and behavior mode of the user, and provide more personalized and intelligent lighting experience.

As shown in fig. 1 and fig. 2, the technical solution in this embodiment shows an advanced intelligent lighting control system, which not only considers multiple control modes, but also deeply analyzes scene requirements and scene lighting conditions, and accordingly designs corresponding scene control and main control programs. The comprehensive method provides a comprehensive and flexible solution for intelligent illumination control of the building. Let us discuss various aspects of this solution in detail.

Control mode classification:

The present embodiment proposes three main control modes, namely a daytime mode, a noon break mode and a night shift mode. The advantage of this multi-mode design is that it can adapt to the lighting requirements of the building in different time periods and usage scenarios, thereby achieving maximization of energy efficiency and optimization of user experience.

Daytime mode:

the daytime mode is an intelligent illumination control strategy, fully utilizes natural light, and adjusts the brightness of the lamplight according to natural lighting conditions of different areas to realize brightness gradient adjustment. This approach not only provides a balanced lighting effect, but also minimizes unnecessary energy consumption.

The control algorithm for daytime mode is based on the following principle: the illuminance of an individual lamp is closely related to the brightness adjustment of the lighting device. Therefore, when the adjustment is performed, the system compares the preset illuminance values in different areas with the average illuminance actually acquired and calculated, so as to calculate the opening value of the brightness of the lamp in the area.

The calculation formula of the opening value is as follows:

E = N × f(x) × Φ × U × K / A，

Wherein:

e: the average illuminance, i.e. the intensity of illumination that is desired to be achieved at design time, is in lux (lx),

N: the number of lamps, i.e. the number of lamps in the area,

Φ: the luminous flux of the luminaire, i.e. the total amount of light emitted by a single lighting device in the area, is expressed in lumens (lm),

F (x): the opening degree of the light-adjustable lamp can be adjusted,

U: the utilization coefficient of the lamp refers to the proportion of the luminous flux emitted by the lighting equipment, which is effectively irradiated on the working surface,

K: the maintenance factor of the luminaire, considering the problem of reduced efficiency of the lighting device after a period of use, is typically taken to be 0.7 to 0.8 in this embodiment,

A: the area of the area, i.e. the area of the area to be illuminated, is in square meters (m),

The system obtains a difference value by comparing the calculation result of the average illuminance with the actual illumination condition, and judges whether the difference value exceeds the expected setting. This process enables the system to confirm whether different partitioned areas require illumination adjustment, thereby featuring, personalized, accurate brightness adjustment of the lighting devices in the different areas.

The specific steps of the opening degree calculation of the lighting equipment are as follows:

Step 1: the computing area 1 requires the illumination provided by the luminaire,

ΔE1 = E - E1，

Wherein E is the design illuminance and E1 is the average illuminance in region 1.

Step 2: collecting real-time collection parameters of collection equipment on four boundary points a, b, c, d in a partitioned area, taking the average to obtain the actual illuminance in the area,

E1 = (Ea + Eb + Ec + Ed) / 4，

Wherein Ea, eb, ec, ed is the illuminance value acquired by the acquisition device at four boundary points in real time.

Step 3: calculating the area A of the area 1 using the geographical information of the sensors in the BIM model

In calculating the area, coordinates a (x/y/z), b (x/y/z), c (x/y/z), d (x/y/z) of the acquisition device at four boundary points are used.

Step 4: the number of luminaires N1 in region 1 is counted using the geographical information of the sensors in the BIM model.

Step 5: and calculating the opening degree of the lamp in the area 1.

In noon break mode, during noon break time, the system takes into account that most staff needs to rest, so the following measures are taken:

The control part in the linkage adjacent window area closes the sunshade curtain on the window, closes the office area lamplight or adjusts the lamps at intervals to be a designated brightness value (such as 5%), and closes other lamps.

In the night shift mode, as shown in fig. 3, no natural lighting exists in all areas. The system adopts corresponding control strategies according to different conditions:

In the first case, when the working hours are not reached, staff work, unified control is carried out on lamps in the office area, illumination of one area is sampled, the opening degree of the lamps is calculated, and the calculation result is sent to all illumination equipment in the office area.

And secondly, when the working hours are up, only part of staff works, a human body detector is used for detecting staff, if the office staff is detected, the BIM technology is utilized for calculating lamps within 2 meters around the detector, the digital twin technology is utilized for binding the corresponding relation between staff and stations and lamps, the overtime staff can add an overtime illumination report flow and select a time period through a mobile phone, the system submits the staff information according to the flow to obtain the lamps corresponding to the staff stations, and the overtime illumination report flow has higher priority. When the human body detector detects that all workers leave an office area, the system enters a patrol mode, all lighting equipment in the office area in a building is closed, corridor down lamps are turned on at intervals, safety of the patrol workers during patrol is facilitated, and a patrol scene specific algorithm is as follows: through this scene mode of time triggering or property personnel switching, close all lamps and lanterns in office area, divide into a, b two sets of corridor down lamp, turn on a, b two sets of down lamp in turn when triggering the scene mode of patrol to extension lamps and lanterns life, special time control: in a special time period, according to the special activity date and holidays, the system can perform self-defined control on the lighting equipment in the building to realize full-open or full-close.

Manual control intervention: to cope with errors that may occur in intelligent control, the system is also equipped with the following measures: the field control panel is equipped in different buildings, the manual control intervention of the operation station or app is realized through the digital twin technology, the required scene mode is allowed to be selected according to the manual demand, if no personnel participate in the control, the system can automatically trigger different scene modes according to the office time, the manual control priority is higher than the automatic control, the intervention can be performed at any time, the intelligent illumination control system based on multiple modes and multiple scenes not only considers the energy efficiency and the user comfort, but also fuses advanced technologies such as BIM and digital twinning and the like, and realizes highly intelligent and personalized illumination control. The method provides a comprehensive and flexible lighting solution for modern intelligent buildings, and is expected to bring remarkable benefits in various aspects of improving energy efficiency, improving working environment, reducing operation cost and the like.

The embodiment provides an illumination equipment energy consumption optimization control method based on a Building Information Model (BIM) and a Radio Frequency Identification (RFID) technology, and aims to realize efficient energy-saving management of illumination equipment through an intelligent means. The specific implementation steps of the control method are as follows: firstly, three-dimensional data of a building is obtained by using a BIM space model, wherein the three-dimensional data comprise the structure, layout, internal space and space distribution conditions of various devices of the building. In this stage, the BIM model not only provides geometric information of the building, but also can show specific positions and numbers of lighting devices in the building in detail, and provides basic data support for subsequent energy consumption analysis and control.

Next, by installing an RFID tag module on each lighting device, all lighting devices in the BIM model are assigned a unique identification ID. The RFID tags are not only bound with the corresponding IDs, but also can track and acquire the energy consumption information of the lighting equipment in real time. The implementation of this process means that the status and energy consumption data of each lighting device can be monitored and recorded in real time, thereby providing support for subsequent data analysis. The real-time performance is incomparable with the traditional means, and the running state of the equipment can be reflected in time.

In terms of data storage, all lighting device energy consumption information acquired by the RFID tag and the corresponding lighting device ID will be stored by the system efficiently into the relational database. The selection of the relational database provides a structured storage mode for data management, and can organize the data in a form of a table so as to ensure that the relation between the data is clearly expressed. For example, the energy consumption data may be stored in a table named "EnergyConsumption" containing fields such as "DeviceID" (device ID), "Timestamp" (time stamp), "EnergyUsed" (energy consumption), etc. Each record represents energy consumption information of one lighting device over a specific period of time. Further, in order to better manage attribute information of the device, another table "Devices" may store detailed information of the device, including "DeviceID", "DEVICETYPE" (device type), "Location" (installation Location), "Status" (device Status), and the like. The main foreign key relation is established between the two tables through the DeviceID, so that the energy consumption data and the equipment attribute information can be conveniently associated during inquiry. This design of relational databases allows data integrity to be guaranteed, preventing data redundancy and inconsistencies through constraints such as primary keys, foreign keys, and uniqueness constraints. The relational database also supports the standard SQL query language, and users can realize complex data retrieval through simple query sentences. For example, the system can quickly screen out the energy consumption data of a specific type of lighting device in a specific time period through SQL sentences, or combine the energy consumption information with the environmental parameters (such as temperature, humidity and illumination intensity) of the building through a joint operation to form a comprehensive analysis view. In the control time period, the system can rapidly call the real-time energy consumption information of the required lighting equipment from the relational database, and input the information, the attribute information of the building body and the real-time environment information inside the building into the lighting equipment energy consumption prediction model. This ability to integrate data enables the model to more accurately analyze energy consumption patterns, thereby predicting future energy consumption trends and providing decision support for intelligent lighting control. Through the efficient management of the relational database, the system not only realizes the centralized storage and management of the data, but also provides a solid foundation for the subsequent energy consumption analysis and optimization, and ensures the reliability and traceability of the data.

The illumination device energy consumption prediction model adopts a neural network technology, and the selection is based on the advantage of the neural network in the aspect of processing complex nonlinear data. In the model training phase, environment information, building body attribute information and historical energy consumption data of the lighting equipment are taken as input parameters. In the process, the parameters in the neural network are learned and trained by using the existing training method until the preset performance standard is reached. Successful implementation of the stage greatly improves the prediction accuracy of the model on future energy consumption, and provides reliable basis for subsequent intelligent control.

In some embodiments, a training process of a neural network-based illumination device energy consumption prediction model, a data acquisition and feature extraction method, a data normalization process, a selection of a loss function, and a model optimization strategy are described in detail.

1. The successful establishment of the model depends on high-quality and accurate input data. Thus, the acquisition of data is the first step in the overall model building process. In this embodiment, we use a variety of sensors and intelligent devices to obtain comprehensive illumination energy consumption data.

And (3) collecting environmental information: we have deployed a variety of sensors inside and outside the building to monitor environmental factors in real time. These sensors include temperature and humidity sensors, light intensity sensors, and weather monitors. The temperature and humidity sensor can effectively monitor the temperature and humidity changes inside and outside the room, and the illumination intensity sensor is used for acquiring real-time data of natural illumination intensity. Weather monitors provide weather related data such as precipitation, wind speed, etc. These data are transmitted in real-time to a central database through IoT (internet of things) technology.

Collection of building characteristic data: the design drawing and BIM (building information model) data of the building provide important basis for obtaining the attribute of the building body. We have extracted information about the type of building (e.g., office building, mall, house), building area, number of floors, window orientation, building materials, etc. These properties affect not only the energy consumption of the lighting device, but also the accuracy of the energy consumption prediction by interacting with environmental factors.

Acquisition of historical energy consumption data of lighting equipment: in order to accurately grasp the energy consumption characteristics of the lighting devices, a smart meter is used for monitoring the electricity consumption of each lighting device. The intelligent electric meters can record electricity consumption data of each device in different time periods to form detailed historical energy consumption records. Through the data, the energy consumption characteristics of the equipment under different use scenes can be analyzed, and rich input information is provided for the model.

2. The feature extraction and the data preprocessing are performed, and the quality of the data directly influences the prediction effect of the model, so that after the data acquisition is completed, the data needs to be preprocessed and the feature extraction is performed, so that the subsequent model training is facilitated.

Feature selection: from the collected data we first make feature selection, focusing on features related to illumination energy consumption. Environmental factors (e.g., temperature, humidity, illumination intensity), architectural characteristics (e.g., architectural area, window orientation), and historical electricity usage data are all potential influencing factors. By correlation analysis we can identify features that are closely related to energy consumption, which will be input to model training.

The feature engineering is an important link of data preprocessing, and after feature selection is completed, the data is subjected to systematic feature engineering processing so as to improve the performance and prediction capability of the model. For classification features, such as building type and window orientation, we have employed one-hot coding methods to convert these class features into numerical features. Specifically, one-hot encoding is accomplished by creating a new binary feature for each category, representing the value of each category as either 0 or 1, so that the model can understand these category information numerically. For example, assuming that the building type contains three categories, "residential", "commercial" and "industrial", by one-hot encoding we will generate three new feature columns, one for each building type, with only columns for the respective category being 1 and the remaining columns being 0. This approach not only effectively eliminates sequential relationships between class features, but also avoids misunderstanding of the model when processing these features. In addition, the one-hot coding ensures that the data input by the model are numerical while retaining the category information, and further enhances the learning capacity and prediction accuracy of the model. Through the feature engineering processing, clearer and more effective feature representation can be provided for the subsequent modeling stage, and the overall performance of the model is further improved.

In some embodiments, when the dimension of the feature vector of the training data is too large, we can perform dimension reduction processing on the feature of the training data to improve the training efficiency and performance of the model. High dimensional data may not only lead to increased computational complexity, but may also cause "curse-dimensional" problems, making it difficult for the model to find effective decision boundaries during training. To solve this problem, we can use various dimension reduction techniques. For example, principal Component Analysis (PCA) is a commonly used linear dimension reduction method, which extracts the most important principal components in data by linearly transforming features, thereby reducing dimensions while preserving variance of the data as much as possible. In addition, linear Discriminant Analysis (LDA) is also an effective dimension reduction technique, particularly suited for classification tasks, which enhances the separability between classes by finding the optimal projection direction, thereby reducing the feature dimension. Methods such as t-distributed random neighbor embedding (t-SNE) and self-encoder (Autoencoders) are also popular in processing nonlinear data. the t-SNE effectively maps high-dimensional data into low-dimensional space in a manner of maintaining a local structure, is suitable for visualization, and the self-encoder is a neural network structure, compresses an original input into a low-dimensional representation through the encoder, and can capture complex nonlinear characteristics through reconstruction of a decoder. In addition, factor Analysis (Factor Analysis) can also be used for dimension reduction, which simplifies the data structure by identifying potential variables (factors) to interpret the relationships between multiple observed variables. Also, feature selection algorithms, such as L1 regularization (Lasso) and tree-based models (e.g., random forests), are used to evaluate the importance of features to select the most representative features, and are also an indirect dimension reduction method. Through the diversified dimension reduction technologies, the feature space of training data can be effectively simplified, the complexity of a model is reduced, and meanwhile, the training efficiency and the prediction accuracy of the model are improved.

Extracting time sequence features: since the illumination energy consumption data has time series characteristics, we need to extract time-dependent characteristics. For example, we can extract information on hours, days of the week, holidays, etc. in the time stamp. These features can help the model capture the law of energy consumption over time.

Data normalization: after the feature processing is completed, we normalize the data to ensure that the magnitude of each feature value is comparable. Normalization can avoid that certain features have too much impact on model training. By means of the Min-Max normalization method, all characteristic values are scaled to be within the interval of [0, 1], so that all input characteristics can be trained under the same scale, and the convergence speed and prediction accuracy of the model are improved. 3. The selection of a loss function is a very important link in constructing a neural network model. The loss function is used for measuring the difference between the predicted value and the actual value of the model and guiding the optimization process of the model.

Selection of a loss function: in this embodiment, we choose the Mean Square Error (MSE) as the loss function. The MSE calculation formula is:

，

Wherein, As a result of the fact that the value,Is a predicted value; n is the number of samples. The MSE has good mathematical properties, and can effectively reflect the square sum of the prediction value errors, so that the model is easier to optimize in the training process.

N is the number of samples. The MSE has good mathematical properties, and can effectively reflect the square sum of the prediction value errors, so that the model is easier to optimize in the training process.

Optimization of the loss function is a key step in improving model training efficiency and performance. Throughout the training process, we calculate the value of the loss function periodically to monitor the performance of the model in each training period (epoch). The loss function is a metric used for evaluating the difference between the predicted output of the model and the actual label, and by analyzing the change condition of the loss function, the learning process of the model can be adjusted. Specifically, in the early stage of training, the learning rate may be set relatively large. The purpose of this strategy is to allow the model to explore the parameter space quickly in order to converge quickly to a better region in the initial phase. At this stage, the greater learning rate can promote the sensitivity of the model to different parameter settings, so that the model can search in a wider range, thereby increasing the training speed. As training progresses, we observe that the value of the loss function gradually decreases and gradually tends to converge. At this stage, we will gradually decrease the learning rate in order to ensure that the model can be more finely tuned as it approaches the optimal solution. Through reducing the learning rate, the model can carry out finer parameter updating when approaching to the optimal solution, and the vibration phenomenon caused by overlarge learning rate is prevented. This method of dynamically adjusting the learning rate may employ various strategies such as learning rate decay (LEARNING RATE DECAY), i.e., a proportional decrease in learning rate after each training period, or adaptive learning rate algorithms such as Adam, RMSprop, etc., which can adaptively adjust the learning rate based on the sum of squares of the historical gradients. Through the optimization strategy, the convergence speed of the model can be increased, and the accuracy and the robustness of the final model can be improved. The periodic monitoring of the change in the loss function and the adjustment of the learning rate according to its trend is an important means to ensure that the model remains efficient and effective during training. The validity of the methodology is verified in a plurality of practical applications, and the development and application of the deep learning model are further promoted.

4. Model training process, after data preparation and loss function selection, model training process formally starts. We used the following steps for model training and validation.

Model initialization: prior to training, we initialize parameters of the neural network, including weights and biases. The weights may be randomly initialized to avoid symmetry problems, while accelerating model convergence.

Training set and verification set partitioning: we randomly divided the collected data into training (70%) and validation (30%). The training set is used for training the model, and the verification set is used for evaluating the performance of the model, so that the generalization capability of the model is ensured.

Forward propagation: forward propagation is performed using the training set, and a predicted value for each sample is calculated. In the process, the input layer receives the characteristic data, and the output layer finally generates a predicted value through multiple nonlinear transformation of the hidden layer.

Loss calculation: the loss between the predicted value and the actual value is calculated using a loss function. This process can help us to understand the behavior of the model under the current parameters.

Back propagation and parameter update: gradients were calculated by a back-propagation algorithm and model parameters were updated using Adam optimization algorithm. The Adam algorithm combines a momentum method and an adaptive learning rate, and can effectively cope with sparse gradient and noise data. And when updating each time, the model is gradually optimized according to the calculated gradient and the current learning rate adjustment parameter.

Verification set evaluation: at the end of each epoch we evaluate the model performance on a validation set, calculate validation losses and observe their trend. If the verification loss fails to continuously decrease after a plurality of iterations, triggering an early-stopping mechanism, stopping training and storing the current optimal model.

Super-parameter adjustment: in the training process, in order to further improve the prediction capability of the model, systematic tuning of the super-parameters is performed by using methods such as grid Search (GRID SEARCH) or Random Search (Random Search). Hyper-parameters refer to parameters that need to be set prior to model training, which have an important impact on the learning process and final performance of the model. Specifically, we will optimize a number of key hyper-parameters, including learning rate, batch size (batch size), regularization coefficients, etc. First, the learning rate is an important super-parameter that affects the model convergence rate and final performance. Through grid search, a grid containing a plurality of learning values, such as 0.001, 0.01, 0.1 and the like, can be set, a model is trained one by one, and the change condition of a loss function under different learning rates is observed, so that the optimal learning rate is selected. The random search is to randomly extract values in a preset learning rate range for training, and the method can more effectively explore super-parameter combinations in a high-dimensional space, so that the calculation time is saved.

Second, batch size (batch size) is also an important hyper-parameter that determines the number of samples used for each model update. By adjusting the batch size, we can influence the convergence speed and memory usage efficiency of the model. Smaller batch sizes generally provide more stable gradient estimates, but increase training time; larger batch sizes may accelerate the training process, but may result in the model staying too long in the locally optimal solution. Thus, we will explore different batch sizes, such as 32, 64 and 128, in a grid search or random search to find the best balance point. Finally, the regularization coefficient is also a key hyper-parameter, which is used to control the complexity of the model and prevent the over-fitting phenomenon. By adjusting the value of the regularization coefficient, such as the intensity of L1 or L2 regularization, the model weight can be effectively limited, so that the generalization capability of the model on unseen data is improved. In the tuning process, we combine different regularization coefficients into the search space for optimization along with other super parameters.

5. After model optimization and iteration and preliminary training, the model is evaluated in detail to judge whether the predicted result reaches the expected effect. If the predictive performance of the model is found to be unsatisfactory, we can take a series of optimization strategies to promote its performance. Among these, feature optimization is an important step that directly affects the learning ability and prediction accuracy of the model. The feature optimization process first involves a deep analysis of existing features. This analysis aims to identify those features that contribute significantly to the model predictions, as well as those redundant or extraneous features that may lead to poor model performance. By using techniques such as feature importance analysis, correlation matrix or Principal Component Analysis (PCA), we can quantify the impact of each feature on model predictions to determine which features need to be retained, deleted or replaced. After the evaluation of the existing features is completed, we will try to add features that are more energy-consumption dependent. This may be achieved by consulting field literature, consulting expert opinion, or using data mining techniques. For example, considering external factors such as the introduction of weather data (e.g., temperature, humidity, wind speed, etc.) or building usage patterns (e.g., personnel density, equipment usage time, etc.), such information may have a significant impact on energy consumption predictions. In addition, new features can be created by means of feature combination to strengthen the expression capacity of the model. This process may involve performing a mathematical operation or a logical operation on a plurality of existing features to generate new features. For example, a new "window area" feature may be generated by multiplying the building area by the number of windows, or energy consumption data for different time periods may be summed to create a "total energy consumption" feature. This combination not only captures potential interactions between features, but also allows the model to better adapt to complex patterns in the data. By implementing these feature optimization strategies, we hope to significantly improve the performance of the model, thereby improving its prediction accuracy and robustness. The process is not only beneficial to the improvement of the performance of the model on training data, but also can enhance the generalization capability of the model on unseen data, and provides more reliable prediction results for practical application.

Activation functions play a critical role in deep learning models because they introduce nonlinearities that enable neural networks to learn complex features and patterns. Common activation functions include ReLU (modified linear units), leak ReLU, tanh (hyperbolic tangent functions), sigmoid, and the like, each of which has unique advantages and disadvantages. ReLU is widely used because of its high computational efficiency, has linear characteristics in the positive interval, and can effectively accelerate convergence. However, reLU may cause a "dead neuron" problem in some cases, i.e., the output of some neurons is always zero during training, and no effective learning can be performed. To solve this problem, a leak ReLU has been developed. The leak ReLU introduces a small slope in the negative interval, so that negative values can also transmit some gradients, thereby improving the learning ability of the model and reducing the occurrence of the phenomenon of dead neurons. The output range of the Tanh activation function is between-1 and 1, and the Tanh activation function has a good normalization effect and is suitable for processing data with strong negative correlation characteristics. Compared to Sigmoid, tanh is more resistant to the gradient vanishing problem and therefore performs better in deep networks. However, the gradient of the Tanh function becomes very small as the output approaches-1 or 1, resulting in a slow training speed. The Sigmoid function limits the output to between 0 and 1, which is often used for the output layer of the two-class problem, but the problem of gradient extinction easily occurs in deep networks, especially in multi-layer neural networks, resulting in difficulty in training. Thus, care must be taken when using in the hidden layer. Through experiments on different activation functions, we can better understand their performance in a specific task, and thus select the most suitable activation function. Comprehensively considering the data characteristics and the model structure, reasonably selecting and adjusting the activation function can obviously improve the performance and training efficiency of the model. And (3) ensemble learning: to further improve the accuracy of the model, we can consider integrating multiple pre-trained models to obtain the final prediction result by voting or weighted averaging. The integrated learning method can effectively reduce the deviation of a single model and improve the reliability of overall prediction.

Training data augmentation is an important strategy for improving the performance of a deep learning model, and particularly when the model is not well behaved under certain specific conditions, more training samples are generated through a data enhancement technology, so that the generalization capability and adaptability of the model can be remarkably improved. In particular, data augmentation may augment the training data set in a variety of ways. First, the image data may be transformed by rotation, translation, scaling, flipping, cropping, and other geometric transformations to generate new samples. For example, in an image classification task, the original image may be rotated randomly through a certain angle, or scaled while maintaining a scale, to simulate scenes at different perspectives and distances. Second, color conversion is also an efficient way to adjust brightness, contrast, saturation, etc., so that the model can adapt to images under various lighting conditions. In addition, the interference in the actual environment can be simulated by adding noise or fuzzy processing, so that the robustness of the model is further improved.

Through the continuous iteration and optimization of the steps, a deep learning model capable of accurately predicting the energy consumption of the lighting equipment is finally formed, and solid technical support is provided for an intelligent control system of a building. With the continued optimization and application of models, we believe that this technique will play an important role in energy efficiency management of buildings.

The output of the predictive model will include the predicted values of the energy consumption of the individual luminaires over a period of time in the future and the corresponding luminaire adjustments. This information is of great importance for building manager decision making, and can effectively guide the on-off control strategy of the lighting equipment. Specifically, based on the predicted result, the system can automatically adjust the on-off state of the non-air-conditioning electric equipment, so that the operation efficiency of the lighting equipment is optimized, and unnecessary energy consumption is reduced.

In the specific implementation process, the intelligent control capability of the system can flexibly carry out energy-saving management on the lighting equipment in the building on the basis of not damaging the existing intelligent illumination control mode of the building. The intelligent management not only improves the overall operation efficiency of the lighting system, but also provides technical assurance for realizing high-efficiency and energy-saving building illuminance adjustment control. In addition, through continuous energy consumption monitoring and data analysis, a building manager can timely master the running condition of the lighting equipment and find potential energy saving opportunities, so that the lighting management strategy is further optimized.

Example 3: the embodiment provides a building intelligent illumination control system based on a Building Information Model (BIM) and a digital twinning technology. The design of the system aims at realizing the efficient management and optimization of the lighting equipment in the building through real-time data acquisition and intelligent processing, thereby improving the energy efficiency and the comfort level of indoor lighting. Specific embodiments and components thereof are as follows: first, the system includes a data acquisition module that is responsible for acquiring the illumination intensity data of each area within the building in real time. The illumination acquisition and transmission devices are distributed at different positions of the building, can accurately measure the current natural illumination condition, and transmit acquired data to the central processing module. The design of the data acquisition module enables the system to comprehensively monitor illumination changes of different areas, and ensures that the lighting equipment can respond in time under different environmental conditions.

The central processing module is a core part of the system, and the main function of the central processing module is to perform centralized processing on parameters related to illuminance value calculation in different areas. The module analyzes the current illumination condition according to the collected illumination data, the attribute information of the building body and the user requirements, and confirms the adjustment scheme according to the processing result. For example, if the illumination intensity of a certain area is lower than a preset standard, the central processing module will generate a corresponding adjustment scheme to instruct the relevant lighting device to turn on or adjust the brightness.

In order to achieve efficient device management, the system also integrates an RFID tag module, which plays a vital role in the overall lighting management system. The RFID (radio frequency identification) tag module is responsible for assigning each lighting device a characteristic RFID tag, which not only contains basic information of the device, such as model number, power, date of production, etc., but also integrates real-time data related to the running state of the device, such as the use duration of the device, maintenance records, fault alarm information, etc. This integration of information makes device management more efficient and intelligent. In practical application, the RFID tag and the central processing module are closely cooperated, so that real-time calling and state monitoring of the equipment can be realized. When the system needs to operate a certain lighting device, the central processing module can quickly identify the device through the RFID reader and acquire the position and state information of the device, so that accurate control is implemented. For example, when the natural light intensity of a certain area changes, the system can instantly identify the lighting equipment in the area and adjust according to a preset intelligent dimming strategy to ensure the efficient use of energy.

Each RFID tag is bound to a lighting fixture component in a Building Information Modeling (BIM) space model, a design that ensures that the system can accurately identify the specific location and status of each fixture. By establishing the corresponding relation between the RFID tag and the BIM model, a manager can intuitively check the distribution condition and the running state of each lighting device in the BIM model, so that the convenience and the efficiency of device management are greatly improved. In the BIM model, all lighting equipment components are provided with unique IDs, which not only helps to avoid confusion and repeated management of equipment, but also provides accurate basis for subsequent equipment maintenance, updating and fault investigation. In addition, the use of RFID tags also provides important support for asset management and lifecycle management of the devices. By periodically scanning and recording the RFID tag information, the system can generate detailed equipment operation reports, help management personnel to timely master the service condition and maintenance requirement of the equipment, thereby making a reasonable maintenance plan, prolonging the service life of the equipment and reducing the operation cost. In a word, the integrated RFID tag module provides powerful technical support for efficient management of lighting equipment, and promotes comprehensive upgrading and optimization of an intelligent lighting system.

The calling module is used for calling corresponding lighting equipment to adjust according to the adjustment scheme of the central processing module and the marking information of the RFID tag module. Through the cooperation with the RFID tag, the calling module can quickly and accurately find the illumination equipment to be adjusted, ensure to respond to the instruction of the central processing module in time, and effectively improve the illumination efficiency and accuracy in the area.

In terms of data storage and processing, the system adopts a cloud computing server. The main function of the cloud computing server is to call the real-time energy consumption information of the lighting equipment in the current time period from a relational database in the regional controller. The energy consumption data and the attribute information of the building body and the real-time environment information in the building are input into the illumination equipment energy consumption prediction model together. By analyzing the historical data and the real-time data, the system can estimate the predicted value of the energy consumption of the lighting equipment in the next time period.

In order to improve the accuracy of the lighting device energy consumption prediction model, the system gradually optimizes the model by comparing the difference values of the actual energy consumption values of the corresponding lighting devices in a future time period. After each adjustment, the system evaluates the prediction result and continuously adjusts the model parameters according to the feedback information so as to improve the prediction accuracy.

It is worth mentioning that, by using the illumination device energy consumption prediction model, the system not only can monitor the energy consumption condition of the illumination device in real time, but also can predict the predicted value of the brightness collected in the inner chamber in the future time period. The realization of the function depends on advanced data analysis technology and machine learning algorithm, and the system can comprehensively consider a plurality of factors, such as historical energy consumption data, meteorological conditions, time variation, indoor and outdoor illumination intensity and the like to generate accurate daylighting brightness prediction. Through the predicted values, the central processing module can fit a detailed brightness curve. This luminance profile reflects the change in indoor lighting demand at different points in time over the future time period, from which the system forms specific lighting device control instructions. The control method based on prediction not only can improve the intelligent level of the lighting system, but also can realize more efficient energy management. For example, suppose that the system predicts that an area will be affected by natural illumination during the afternoon hours by analysis, and predicts that the indoor brightness will increase. In this case, the central processing module will automatically generate instructions instructing the associated lighting devices to reduce their luminance output, so as to avoid wasting energy in the case of good lighting conditions. Through the dynamic adjustment, the system can furthest reduce the power consumption on the premise of ensuring the comfort level of the indoor environment, and achieves the energy-saving effect. Furthermore, this combination of prediction and control provides a guarantee for flexibility and adaptability of the lighting system. Along with the change of seasons, weather conditions and indoor activity modes, the system can adjust the brightness curve in real time, so that the illumination equipment is always kept in an optimal working state, the requirements of users are met, and unnecessary energy consumption is reduced. Through the intelligent management mode, the system not only improves the lighting effect, but also creates a more comfortable and efficient indoor environment for users, and further promotes the targets of green building and sustainable development.

The characterization function of the RFID tag enables the system to accurately correlate the operating characteristic information of the lighting fixture component acquired in real-time with the lighting fixture component ID and store it to the relational database. The process ensures that the system can call the running state and the energy consumption condition of each lighting device at any time, and provides comprehensive data support for the central processing module. Meanwhile, the judging and processing mechanism arranged in the cloud computing server can judge and analyze the normal operation characteristics of the lighting equipment prestored in the database, and the equipment is ensured to operate in an optimal state.

In different areas of the building, all lighting devices are provided with a characterized RFID tag. This design allows the system to accurately locate different lighting devices by RFID tags when adjusting the zone brightness. When the brightness of a certain area needs to be adjusted, the calling module can rapidly schedule the illumination equipment with the accurate RFID tag to be correspondingly adjusted. This capability not only improves the efficiency and accuracy of the brightness adjustment within the area, but also reduces errors and delays that may be caused by manual operation.

In addition, the intelligent illumination control system for the building in the embodiment also has good expandability. With the expansion of the building scale, the system can easily integrate more illumination acquisition and transmission devices and lighting equipment, and ensure that the system can adapt to buildings with different scales and functions. Meanwhile, the system design also considers the updating and iteration of future technology, and can be seamlessly connected with other intelligent building management systems to realize a more comprehensive building management scheme.

In practical application, the system can remarkably improve the lighting management efficiency in the building. Through real-time data acquisition, intelligent processing and control, the system can automatically adjust the running state of the lighting equipment according to actual use requirements and environmental changes. The indoor lighting comfort level can be improved, unnecessary energy consumption can be effectively reduced, and the building is promoted to be in a green and intelligent development direction.

Aiming at a plurality of defects existing in the existing building interior illuminance adjustment, the invention further discusses challenges of the current illumination control system in practical application. These challenges not only affect the energy consumption efficiency within a building, but also are directly related to the comfort and work efficiency of the staff. By analyzing the prior art, the defects of the current lighting control system can be clarified, so that a foundation is laid for the innovative solution of the invention.

First, the conventional lighting control method is too single, and mainly depends on the unified switching control of the lighting device. The control mode cannot be intelligently adjusted aiming at natural lighting conditions in different areas. In fact, there are significant differences in the natural lighting conditions inside a building, especially in different weather conditions and time periods, the intensity and direction of the illumination changes. For example, a southbound window may produce intense illumination when direct sunlight, while a northbound window is relatively dark. In this case, a single control manner cannot meet the lighting requirements of different areas and different time periods, so that part of the areas are too bright and other areas are dim, which affects the whole indoor environment.

Second, another important drawback of conventional lighting control is that it is not considered enough for the user's needs. Different users have great differences in the requirements for illumination in the actual use process. For example, some employees may prefer a bright environment to increase work efficiency, while others may prefer softer lights to reduce visual fatigue. The unified control mode cannot meet the individual requirements, and the illumination effect is often not ideal due to lower control precision. Such a situation is particularly prominent in modern office environments, where the office typically needs to accommodate multiple users, and traditional lighting control approaches have been plagued by elbows.

In complex building environments, the complexity of lighting control is not negligible. Since modern buildings are usually composed of multiple floors and areas, the illumination requirements of the areas are different, and it is obvious that the requirements in actual use cannot be met by only relying on time timing control. The traditional management mode often depends on manual intervention, but the workload of management staff is increased, a large amount of human resources are consumed, and the efficiency is low. As building scales are increased, the number of lighting devices that the manager needs to face is also increasing, which puts higher demands on manual management, resulting in greater working pressure.

Further analyzing the limitations of conventional lighting control systems, it can be found that they are deficient in time division. The existing control mode generally divides the working mode of one day into a daytime working mode, a noon break mode, a overtime mode, a patrol mode and the like. However, such time-divided control schemes often fail to carefully cope with variations in illumination intensity and range over different time periods. For example, during daytime hours of operation, the intensity and position of sunlight may change over time, and thus it is difficult to accommodate such changes with only a fixed control mode. In addition, the lighting requirements of different areas are also affected by factors such as indoor layout, window orientation, and shielding of surrounding buildings, which are often ignored in conventional control schemes.

The energy consumption problem of conventional lighting control systems is also not small, a phenomenon which is common in many sites and affects far. Firstly, because the control granularity of the systems is larger, a fine dimming function cannot be realized, so that the illumination intensity cannot be dynamically adjusted according to actual requirements. Such limitations directly lead to waste of energy, since in many cases the lighting device remains on even if not needed, resulting in a significant loss of energy consumption. For example, during daylight sunny periods, the indoor lighting devices still remain fully lit, resulting in unnecessary power consumption, increasing operating costs.

In addition, the traditional lighting control mode generally lacks of a corresponding intelligent control algorithm, and the mode of relying on manual control greatly limits the flexibility and efficiency of the system. Manual control often suffers from untimely or inaccurate control, which may further result in waste of energy. For example, if the natural lighting conditions of an area are improved, the lighting system still uses a fixed lighting intensity without timely adjustment, which may cause unnecessary power consumption. The situation is particularly obvious in large buildings or public places, especially in open meeting rooms, hallways, parking lots and other areas, and the lighting system cannot respond intelligently according to the use condition, so that the energy waste is more serious.

Aiming at the defects of the prior art, the invention provides an innovative solution, and the comprehensive monitoring of the illuminance in an office area is realized by combining an illumination sensor and a human body detection sensor through the LED lamp with the adjustable light and communication function and the matched illumination controller. The sensors can be reasonably designed and arranged according to indoor lighting conditions and office positions to cover the whole office area, and the illuminance condition of each area can be monitored in real time. Meanwhile, the sensor can monitor personnel conditions in an office area in an overtime period, so that intelligent control of the lighting equipment is realized.

Through the internet of things technology and the digital twin technology, building parameters, illumination parameters and equipment parameters of an office area are mapped to upper computer monitoring software, so that the system can accurately control a single lamp and adjust the brightness of the single lamp. The processing mode can monitor the illuminance conditions of all areas of the office environment, can know the distribution conditions of personnel in real time, and provides a solid material basis for the design of an intelligent illuminance control algorithm. In addition, building information and lighting equipment information of an office building are modeled and managed by adopting the BIM technology, and data support can be provided for subsequent lighting control in the early design stage.

In a specific implementation, according to the coordinate information of the illumination sensor in the BIM model, average illumination in a specific area is calculated through sampling, and the number of lamps required in the area is calculated according to the average illumination. Similarly, the coordinate information of the human body detection sensor can also help to calculate nearby lamps, and the lamps and sensor parameters acquired in real time are combined to provide necessary data support for the design of an intelligent illuminance control algorithm. By the method, the system can adjust the illumination requirements in real time in different time periods and different areas, and the illumination of each area can be guaranteed to reach the optimal state.

When the illuminance control algorithm conforming to different scenes is designed, the illuminance control method fully considers the illuminance requirements of office staff in different time periods. For example, in areas where natural light is sufficient, the system may automatically decrease the brightness of the light fixtures, and during overtime hours, if no personnel are in a certain area, the system may choose to turn off the corresponding light fixtures. Through these intelligent control strategies, can satisfy office staff's user demand, can effectively practice thrift the electric energy again, improve the economic nature of system.

By combining BIM and digital twin technology, the invention can realize the accurate management of the lighting equipment in the building. The BIM model feeds back the distribution position of the lighting equipment in the building and the space distribution condition of the building structure in real time, so that a manager can quickly know the illumination requirements and the brightness changes of each area, and efficient brightness adjustment is realized. The integrated management mode not only improves the adjustment efficiency of the illumination in the building, but also ensures the accuracy of illumination control and meets the high requirement of modern office environment on the illumination.

In addition, the invention also combines RFID technology, and each lighting device can carry unique ID by attaching the characteristic RFID label to all lighting devices in the building. The design not only improves the convenience of equipment management, but also provides a foundation for real-time monitoring and data analysis. The information stored in each RFID tag contains basic attributes of the device such as model, power, installation location, and other relevant parameters, which enable the system to accurately identify the specific situation of each device. Through RFID technology, the system can display the running state of the lighting equipment in real time, including the information such as the current working mode, brightness level, fault alarm and the like. The real-time monitoring function enables a manager to master the running condition of the lighting equipment at any time and discover and process potential problems in time, so that excessive use or energy consumption waste of the equipment is avoided. In addition, the system can also effectively track the energy data, record the energy consumption condition of each device and generate a detailed operation report. The data provides important basis for subsequent maintenance and optimization, and helps the manager make more scientific decisions.

On the basis of comprehensively considering the attribute information of the building body and the real-time environment information in the building, the system can obviously improve the accuracy and control effect of the building energy consumption prediction. For example, when the system predicts energy consumption, the structural characteristics of the building (such as the orientation of the window, the heat conduction performance of the building material, etc.) and the current real-time data of ambient light, temperature, humidity, etc. are combined to form a multi-dimensional analysis model. The comprehensive analysis capability enables the system to not only predict future energy consumption trend, but also identify key factors influencing energy consumption, thereby better controlling and optimizing. Based on the data and analysis, the system can realize efficient brightness adjustment of the lighting equipment and automatically adjust the lighting intensity of each area so as to adapt to different use requirements and environmental changes. For example, the system can automatically reduce the brightness of the lighting device in the case of sufficient natural light, and increase the brightness in the case of insufficient illumination, so as to ensure the comfort and safety of the indoor environment. Through intelligent regulation, the system can find the best balance point between energy-saving illumination and use requirement, effectively reduces unnecessary energy consumption, promotes the whole energy efficiency level of building. The intelligent lighting management scheme not only accords with the concept of sustainable development, but also creates more comfortable and efficient working and living environments for users.

In summary, the building intelligent illumination control system based on the BIM and digital twin technology provided by the invention provides an efficient and intelligent solution for the defects of the existing illumination control scheme. Through real-time monitoring and intelligent regulation of illumination conditions, the system can meet illumination requirements of different users, meanwhile, building energy consumption is effectively reduced, and the aim of sustainable development is fulfilled.

Claims (10)

1. The intelligent illumination control method for the building based on the BIM and digital twin technology is characterized by comprising the following steps of:

s1, acquiring a BIM space model diagram comprising a building space structure and lighting equipment distribution;

S2, dividing the building space into different types of areas according to lighting conditions;

s3, selecting a control mode to calculate opening values of lighting equipment in different areas;

s4, adjusting the on-off states of the lighting equipment in different areas according to the difference value of the opening value and the set value to realize intelligent illumination control of the building.

2. The building intelligent illumination control method based on the BIM and digital twin technology according to claim 1, wherein the control mode comprises a daytime mode, and the brightness of the lighting equipment is adjusted according to the lighting condition of the divided areas, so that the brightness gradient adjustment is realized; in the noon break mode, the sunshade curtain and the lighting equipment of the office area are closed in the set noon break time; the night shift mode uniformly adjusts all areas to set brightness, and if overtime staff exists, the digital twin technology and the human body detector are combined to monitor and accurately start the lighting equipment of the part where the overtime staff is located; patrol mode, interval turns on corridor lights and turns off lighting of the rest of the area.

3. The intelligent illumination control method for the building based on the BIM and digital twin technology according to any one of claims 1 or 2, wherein when different types of areas are divided, the space in the building is divided in a regular pattern in consideration of station distribution conditions, and the types of the different areas are classified according to the space positions of the different areas and the lighting states acquired by acquisition equipment.

4. The intelligent illumination control method for the building based on the BIM and the digital twin technology according to claim 3, wherein the region type comprises a region with better illumination, and the illumination brightness in the region is larger than a set value x; the illumination area is poor in illumination, and the illumination brightness in the area is smaller than a set value x; the window area comprises a window, and brightness change can be regulated by linking a roller shutter arranged on the window when brightness in the area is regulated.

5. The building intelligent illuminance control method based on the BIM and digital twin technology according to claim 3, wherein in the step S4, the illumination devices in the divided areas are labeled with RFID tags according to the location and type of the areas and the type of the devices, and each illumination device is provided with an independent RFID tag, and when the building intelligent illuminance control is performed, the corresponding illumination device is adjusted by reading the RFID tag.

6. The intelligent illuminance control method of building based on BIM and digital twin technique according to claim 4, wherein the opening value is calculated by:

s31, acquiring average illumination and design illumination in the region p through acquisition equipment, and comparing the difference value to calculate illumination supported by illumination equipment required in the region p;

S32, acquiring the area of the region p and the number of the lighting equipment in the region through the BIM space model diagram;

s33 calculates the opening value of the lighting device from the calculated and counted related parameters within the region p.

7. The intelligent illuminance control method of building based on BIM and digital twin technology according to claim 6, wherein the opening value is specifically calculated as follows: Δep=e-Ep, where E is the design illuminance and Ep is the average illuminance within the region p; ep= (ea+eb+ ec+Ed)/4

Ea, eb, ec and Ed are real-time parameters of points a, b, c and d in the acquisition area of the acquisition equipment;

，

wherein the coordinate of the acquisition point a is a (x/y/z), the coordinate of the acquisition point b is b (x/y/z), the coordinate of the acquisition point c is c (x/y/z), the coordinate of the acquisition point d is d (x/y/z), Area a, which is region p;

，

，

Wherein the method comprises the steps of The aperture value of the lamp in the region p is phi, the luminous flux of the lighting equipment, U is the utilization coefficient of the lamp, K is the maintenance coefficient of the lamp, and f (x) is the aperture of the dimmable lamp.

8. The intelligent illuminance control method of building based on BIM and digital twin technology according to claim 7, wherein the adjustment mode of the daytime mode is as follows:

Wherein E is the average illuminance,

Calculating to obtain real-time average illuminance, comparing a preset illuminance value with the area average illuminance of a fixed area, calculating the opening value of the brightness of the area lamp on the premise of setting stable quantity of the area lamp, and regulating the brightness according to the illumination requirement of the area in daytime.

9. The building intelligent illumination control system based on the BIM and digital twin technology is applicable to the building intelligent illumination control method based on the BIM and digital twin technology according to any one of claims 1 to 8, and is characterized by comprising a data acquisition module, a plurality of illumination acquisition and transmission devices and a central processing module, wherein the plurality of illumination acquisition and transmission devices are connected with the central processing module; the central processing module is used for carrying out centralized processing on parameters related to illuminance value calculation in different areas and confirming an adjustment scheme according to a processing result; the RFID tag module is used for setting different lighting equipment to characterize RFID tags and assisting the central processing module in calling; and the calling module is used for calling the corresponding lighting equipment to adjust according to the adjustment scheme of the central processing module and the marking of the RFID tag module.

10. The building intelligent illumination control system based on the BIM and digital twinning technology according to claim 9, wherein the RFID tag module comprises a plurality of RFID tags, different RFID tags are correspondingly bound with lighting equipment components in a BIM space model, and unique IDs are set on different lighting equipment components of the BIM space model; the cloud computing server is used for retrieving real-time energy consumption information of the lighting equipment in the current time period from a relational database in the regional controller, inputting the real-time energy consumption information, the building body attribute information and the real-time environment information in the building into the lighting equipment energy consumption prediction model together, and predicting an energy consumption predicted value of the lighting equipment in the next time period.