An intelligent building intelligent air environment maintenance system and method
By constructing a multi-dimensional environmental perception network and airflow field coupling analysis, combined with equipment health assessment and dynamic adaptive maintenance strategies, the problems of multi-space collaborative control and insufficient equipment wear perception in intelligent building air environment maintenance systems have been solved. This has achieved global stability of air quality and improved equipment reliability, while reducing operating energy consumption.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HENGTAI LIANYING (TIANJIN) TECH CO LTD
- Filing Date
- 2026-05-09
- Publication Date
- 2026-06-12
AI Technical Summary
Existing intelligent building air environment maintenance systems have shortcomings in multi-space collaborative control, insufficient equipment wear and tear detection, weak dynamic response capabilities, and a lack of forward-looking early warning mechanisms, resulting in poor air environment regulation effects and delayed operation and maintenance.
A multi-dimensional environmental perception network is constructed, which collects data in real time through a multi-source sensor array. Combined with airflow field coupling analysis and equipment health assessment, a dynamic adaptive maintenance strategy is generated using a long short-term memory neural network model. A closed-loop control and early warning feedback mechanism is adopted to achieve cross-regional air quality prediction and preventive maintenance of equipment.
It effectively addresses the diffusion of pollutants caused by air convection, improves the overall stability of air quality, extends equipment failure intervals, reduces unplanned downtime maintenance costs, enhances the dynamic response capability and physiological comfort of environmental regulation, and achieves long-term system operation and energy saving.
Smart Images
Figure CN122191700A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent building technology, and in particular to an intelligent building intelligent air environment maintenance system and method. Background Technology
[0002] As a core direction of modern urban architectural development, intelligent buildings integrate advanced information technology, automatic control technology, and Internet of Things (IoT) technology to achieve intelligent management and efficient resource allocation of the building's internal environment. Among these, the maintenance and optimization of the indoor air environment is a crucial cornerstone for ensuring the health and comfort of building spaces. With people's increasing demands for indoor microclimate, air environment maintenance technology has evolved from single-indicator monitoring to a complex systems engineering encompassing multi-source sensing, logical processing, and execution feedback. Its operational efficiency directly impacts the physiological health and work efficiency of people inside the building.
[0003] The intelligent building air environment maintenance system primarily collects environmental parameters in real time through multi-dimensional sensor nodes deployed in various areas of the building. Combined with control algorithms, it drives the operation of ventilation, filtration, and temperature and humidity regulation equipment. Its core objective is to construct a closed-loop control system capable of automatically sensing air quality fluctuations and executing corresponding purification and regulation commands. This system aims to achieve refined control of key indicators such as carbon dioxide concentration, particulate matter content, and volatile organic compounds while meeting energy-saving requirements.
[0004] In existing technologies, intelligent building air environment maintenance solutions still face many challenges in practical applications: First, the system has significant shortcomings in multi-space collaborative control, making it difficult to effectively address air convection and pollutant diffusion caused by interconnected interior spaces, resulting in local adjustment strategies failing to achieve optimal global effects. Second, the system lacks sufficient perception of the physical wear and performance degradation trends of air handling equipment, and lacks a mechanism for deeply integrating equipment health with environmental control strategies, leading to maintenance work often lagging behind equipment failures and making full lifecycle management difficult. Third, the dynamic response capability of the control strategy is weak; when faced with complex conditions such as instantaneous changes in population density and sudden environmental pollution, the fixed threshold-based control mode is unable to maintain the stability of the air environment. Finally, existing monitoring and maintenance logic lacks a forward-looking early warning mechanism and proactive closed-loop self-maintenance capability, mainly relying on manual intervention and passive response. These deficiencies limit the precision and long-term operational efficiency of intelligent building environmental maintenance. Summary of the Invention
[0005] The purpose of this invention is to provide an intelligent building air environment maintenance system and method to solve the problems of insufficient multi-space collaborative control, lack of equipment wear and tear perception, and weak dynamic response capability in existing systems.
[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:
[0007] A smart building intelligent air environment maintenance system and method, comprising:
[0008] Step 1: Construct a multi-dimensional environmental perception network: Deploy a multi-source sensor array in the area to be monitored in the smart building to collect environmental parameters in real time, including carbon dioxide concentration, fine particulate matter concentration, total volatile organic compound concentration, temperature and humidity. Then, aggregate the data to the edge computing node through a wireless sensor network and use the median filtering algorithm to remove abnormal noise from the original data to obtain a standardized environmental dataset.
[0009] Step 2: Perform cross-regional airflow field coupling analysis: Based on the building information model of the intelligent building, obtain the geometric topological relationship of each monitoring space, combine the opening status of doors and windows and the wind speed of ventilation ducts, calculate the pressure gradient and air exchange between different spaces, establish a cross-regional pollutant diffusion model, and determine the coupling influence coefficient of local environmental fluctuations on adjacent areas.
[0010] Step 3: Assess the health of the air handling unit: Monitor the operating current, fan speed and vibration frequency of the air handling unit in real time, extract the equipment operation feature vector, compare it with the preset equipment health benchmark model, calculate the wear coefficient and performance degradation rate of each component, and determine the preventive maintenance cycle of the equipment.
[0011] Step 4: Generate dynamic adaptive maintenance strategy: Input the standardized environmental dataset, coupling influence coefficient and equipment health status into the pre-trained long short-term memory neural network model to predict the air quality evolution trend within a predetermined time period, and dynamically adjust the ventilation frequency, filtration level and temperature and humidity setpoints according to the prediction results and equipment status to generate the optimal environmental regulation instructions.
[0012] Step 5: Execute closed-loop control and early warning feedback: The control system drives the air handling equipment to operate according to the adjustment command and continuously monitors the feedback of the adjusted environmental parameters. If the measured value deviates from the target value by more than the first preset threshold, the secondary compensation adjustment logic is automatically started. At the same time, when the equipment wear coefficient exceeds the second preset threshold, a preventive maintenance early warning is pushed to the operation and maintenance terminal.
[0013] In step 1, the deployment density of the multi-source sensor array is dynamically determined according to the spatial functional attributes. For office areas, at least a predetermined number of sensor nodes are deployed in each preset first area. For corridor areas, at least a predetermined number of sensor nodes are deployed in each preset second area. The sampling frequency of the sensor nodes is set to a preset sampling frequency to ensure the real-time performance and accuracy of data acquisition.
[0014] In step 1, the preprocessing of the standardized environmental dataset includes data normalization, which maps environmental parameters of different dimensions to a preset numerical range. The normalization formula is: the difference between the observed value and the minimum value divided by the difference between the maximum value and the minimum value, thereby eliminating the order-of-magnitude differences between different indicators and improving the convergence speed of subsequent algorithms.
[0015] In step 2, the pollutant cross-regional diffusion model uses the laws of conservation of mass and momentum as constraints. Its calculation logic takes into account the chimney effect and wind pressure inside the building. The coupling influence coefficient λ is calculated as follows: the influence coefficient of space A on space B is equal to the cross-sectional area of the two spaces multiplied by the air velocity and then divided by the total volume of space B. This coefficient is used to quantify the degree of environmental correlation between spaces.
[0016] In step 2, the calculation accuracy of the pressure gradient is required to reach a preset pressure accuracy threshold. The calculation model is corrected in real time by setting differential pressure sensors on both sides of the main partition wall. When the differential pressure exceeds the preset differential pressure threshold, the system automatically determines that there is a strong convection trend and increases the weight coefficient of the coupling influence coefficient to a preset multiple.
[0017] In step 3, the equipment operation feature vector includes current harmonic distortion, bearing vibration acceleration RMS value, and motor winding temperature. By performing a fast Fourier transform on the vibration signal, the feature frequency component is extracted. If the amplitude at the feature frequency exceeds the standard deviation of a preset multiple under normal operating conditions, the component is determined to have entered the accelerated wear stage.
[0018] In step 3, the calculation model for the wear coefficient k is: k equals the current cumulative operating time divided by the rated design life, and is corrected by the load factor. The load factor is determined based on the proportion of time the equipment operates above the rated power. If the equipment is in an overloaded state for a long time, the load factor is set to a preset correction coefficient, thereby achieving accurate prediction of the remaining life of the equipment.
[0019] In step 4, the long short-term memory neural network model includes a preset number of input layers, hidden layers, and output layers. Each hidden layer contains a preset number of neuron units. It adopts the hyperbolic tangent activation function and learns from historical environmental data within a preset historical period to capture the periodic patterns and sudden trends of air quality changes.
[0020] In step 4, the generation logic of the dynamic adjustment strategy is based on a multi-objective optimization algorithm. Its objective functions include minimizing the environmental comfort offset, minimizing the system operating energy consumption, and minimizing the equipment wear rate. Different operating modes are switched by setting different weight factors. In the energy-saving mode, the proportion of energy consumption weight is set to the first preset weight ratio.
[0021] In step 5, the closed-loop control process adopts an improved proportional-integral-derivative control algorithm. Its proportional coefficient, integral coefficient and derivative coefficient are adaptively adjusted in real time according to the severity of environmental fluctuations. When the rate of increase of carbon dioxide concentration exceeds the preset rate of change threshold, the proportional coefficient automatically increases by the preset ratio to accelerate the system response speed.
[0022] In step 5, the preventive maintenance early warning system predicts the fault location based on the growth slope of the wear coefficient. If the predicted fault occurs within the preset early warning time threshold, the system will automatically lock the operating frequency of the actuator below the preset safe operating ratio to reduce the risk of sudden faults, and simultaneously generate a work order containing the faulty component number, a list of recommended replacement parts, and maintenance operation procedures.
[0023] The intelligent building air environment maintenance system also includes a global data management platform. This platform stores environmental operation data and maintenance records within a preset time period through a high-performance database, and uses big data analysis technology to explore the deep correlation between environmental changes and building energy consumption, providing data support for building energy-saving renovation.
[0024] When performing environmental regulation, the system also considers the intervention of outdoor meteorological data. By accessing real-time temperature, humidity, wind direction and atmospheric pressure data released by the meteorological station, it predicts the impact of the outdoor environment on the building's heat load and air quality. When the outdoor air quality is better than the indoor air quality and the temperature and humidity are suitable, the system prioritizes the use of natural ventilation mode.
[0025] The wireless sensor network in step 1 adopts low-power wide-area network communication technology. Each sensor node is connected to the gateway through a star topology. The communication link has an automatic retransmission and error correction mechanism to ensure that the packet loss rate of data transmission is lower than the preset packet loss rate threshold in the complex electromagnetic environment of the building.
[0026] The prediction results in step 4 also incorporate real-time sensing data of indoor personnel density. The number and distribution information of people in the area are obtained through infrared thermal imaging sensors, and the carbon dioxide load generated by people's breathing is used as a dynamic input variable to further revise the air quality evolution trend prediction model.
[0027] The air handling equipment includes a variable frequency drive (VFD) fan, a return fan, an electronic dust removal and purification unit, a surface cooler, and a humidifier. Each device communicates with the central controller via a bus protocol, and the controller controls each actuator to a predetermined response accuracy according to adjustment commands.
[0028] The system's self-maintenance logic also includes zero-point drift calibration of the sensors themselves. Every time a preset calibration cycle is run, the system automatically controls the solenoid valve to switch to the standard calibration gas flow path to perform benchmark calibration on the carbon dioxide and volatile organic compound sensors, ensuring the stability and consistency of long-term monitoring data.
[0029] Compared with the prior art, the beneficial technical effects of the present invention are as follows:
[0030] This invention breaks through the limitations of traditional systems that can only make local adjustments by establishing a cross-regional pollutant diffusion model and calculating the coupling influence coefficient. It can effectively deal with the problem of pollutant diffusion caused by air convection inside buildings. By quantifying the pressure gradient and air exchange between different spaces, the system can predict pollution trends and initiate linkage adjustment of adjacent areas in advance, which significantly improves the global stability of indoor air quality and effectively eliminates local air quality dead zones.
[0031] This invention deeply integrates environmental control strategies with equipment health status by combining equipment operation feature vectors and wear coefficient calculation models. The system no longer relies on fixed maintenance cycles or passive fault repair, but instead performs preventative maintenance based on real-time monitored current, vibration, and temperature data. This condition-based operation and maintenance mode significantly extends the mean time between failures of air handling equipment, while effectively reducing unplanned downtime maintenance costs, thus achieving long-term effectiveness and reliability of system operation.
[0032] This invention utilizes a long short-term memory neural network model to predict air quality trends and combines it with a multi-objective optimization algorithm to dynamically generate maintenance strategies. This enables the system to flexibly cope with complex operating conditions such as sudden changes in population density, sudden pollution, and drastic changes in outdoor weather. Compared with the traditional adjustment mode based on static thresholds, the dynamic adaptive mechanism controls the fluctuation deviation of environmental parameters within a very small range, significantly improving the physiological comfort of people inside the building. Furthermore, while meeting environmental quality requirements, it achieves a certain proportion of energy consumption reduction by optimizing the operating frequency of the actuators. Attached Figure Description
[0033] Figure 1 This is a schematic diagram of the overall technical solution architecture of the intelligent building intelligent air environment maintenance system proposed in this invention;
[0034] Figure 2 This is a schematic diagram of the core principle framework for generating the dynamic adaptive maintenance strategy in this invention. Detailed Implementation
[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0036] Example 1
[0037] This embodiment provides an intelligent building air environment maintenance system for large commercial complexes. The system employs a layered distributed structure in its hardware architecture, aiming to achieve in-depth control over complex, large-space environments through a high-density sensing network and precise actuators.
[0038] In terms of system architecture, the hardware components of this embodiment include a sensing layer, a transmission layer, a processing layer, and an execution layer. The sensing layer deploys a multi-source sensor array, which consists of a carbon dioxide sensor, a particle sensor for particles smaller than 2.5 micrometers, a total volatile organic compound (TVOC) sensor, a high-precision thermistor temperature sensor, and a capacitive humidity sensor. The carbon dioxide sensor employs non-dispersive infrared detection, with a measurement range set from 0 to 5000 parts per million (ppm), and a measurement accuracy error controlled within ±30 ppm. The particle sensor for particles smaller than 2.5 micrometers utilizes laser scattering principles and can identify particles with diameters between 0.3 and 2.5 micrometers, with a minimum resolvable particle size of 0.3 micrometers. The TVOC sensor is based on metal oxide semiconductor technology and exhibits extremely high sensitivity to organic gaseous substances such as formaldehyde, benzene, and toluene.
[0039] In terms of spatial deployment, the system is divided into office areas, shop areas, corridor areas, and underground parking garages based on the functional attributes of the commercial complex. The sensor deployment density in the office area is set at one set per 20 square meters, installed at a breathing zone height of 1.5 meters above the ground, avoiding air conditioning vents and the edges of doors and windows to ensure the representativeness of the collected data. In the corridor area, due to more frequent airflow and shorter dwell time for people, the deployment density is adjusted to one set per 50 square meters. All sensor nodes integrate a 12-bit analog-to-digital converter to convert analog voltage signals into digital sequences.
[0040] The transport layer employs a wireless sensor network based on a low-power wide-area network protocol. Each sensor node establishes a communication link with the edge computing nodes through its built-in wireless communication module. The wireless communication frequency operates in the 433 MHz band, using a star topology to ensure strong penetration within reinforced concrete structures. The edge computing nodes utilize ARM architecture processors with high-performance floating-point arithmetic capabilities, configured with 1GB of random access memory and 8GB of flash memory, and are responsible for initial data aggregation and cleaning.
[0041] The core of the processing layer is a central control server, equipped with an industrial-grade processor and a large-capacity database, running a long short-term memory neural network prediction model and multi-objective optimization algorithms. The execution layer consists of air handling units, variable frequency fans, return fans, electronic dust removal and purification modules, surface coolers, and electrode humidifiers. The air handling units are connected to the central controller via a Modbus-RTU bus, with a baud rate set to 9600 bits per second, supporting millisecond-level adjustment of fan frequency, valve opening, and purification power.
[0042] Under this system architecture, the workflow of this embodiment is as follows:
[0043] Step 1: Construct a multi-dimensional environmental sensing network. A multi-source sensor array in the sensing layer collects environmental parameters in real time at a sampling frequency of 0.2 Hz. The microcontroller inside each sensor node amplifies and filters the raw analog signals, then outputs digital signals via a 12-bit analog-to-digital converter. Data packets are sent to edge computing nodes via a wireless sensor network. Each data packet contains a node number, timestamp, sensor type code, measured value, and checksum. Upon receiving the data, the edge computing node first executes a median filtering algorithm. This algorithm effectively removes transient abnormal noise caused by electromagnetic interference by sorting five consecutive sampling points and taking the median value. Subsequently, the system performs data normalization processing, mapping environmental parameters of different dimensions to the range of 0 to 1. The mathematical expression is:
[0044]
[0045] Where x is the measured value of the sensor, This is the historical minimum value of this type of parameter. This represents the historical maximum value of this type of parameter. The normalized, standardized environment dataset is stored in a temporary buffer on the edge nodes and periodically synchronized to the central control server.
[0046] Step 2: Perform cross-regional airflow field coupling analysis. The central control server calls the pre-stored building information model to obtain the geometric topology, wall thermal resistance coefficient, and physical layout of ventilation ducts for each monitored space. The system obtains the opening status of magnetostrictive switch sensors installed at the main doors and windows, and combines this with real-time wind speed measurements from Pitot tube anemometers inside the ventilation ducts. Based on the laws of conservation of mass and momentum, the system calculates the pressure gradient between different spaces. When the pressure difference between two adjacent spaces exceeds 0.1 Pascals, the system determines that airflow field coupling exists. At this point, the system establishes a cross-regional pollutant diffusion model and calculates the coupling influence coefficient λ. The calculation logic for this coefficient is as follows:
[0047]
[0048] Where S is the effective cross-sectional area of the connecting part between the two spaces, v is the average air velocity at the connecting part, and V is the total physical volume of the affected space. This coefficient λ is used to quantify the contribution of local environmental fluctuations to the pollution diffusion in adjacent areas. If the value of λ exceeds a preset critical threshold, the system will automatically upgrade the warning level of the air handling equipment in the adjacent area.
[0049] Step 3: Assess the operational health of the air handling unit. During operation, the system monitors the operating current in real time using a Hall current sensor integrated into the motor control circuit, acquires the vibration frequency using a triaxial vibration accelerometer mounted on the fan bearing housing, and monitors the motor winding temperature using an infrared temperature sensor. The DSP chip acquires the vibration signal in real time at a sampling rate of 10 kHz and performs a fast Fourier transform to extract characteristic frequency components. If the amplitude at the fan blade passing frequency and its harmonics exceeds three times the standard deviation of the normal reference value, it is determined that the fan has a dynamic imbalance or bearing wear risk. The system further calculates the wear coefficient k:
[0050]
[0051] in, This represents the current cumulative operating time of the equipment. For the rated design life, This refers to the load factor. When the equipment operates continuously at a full-load power frequency of 50 Hz, Set to 1.25; when running below 30 Hz, The value of k is set to 0.8. By monitoring the k value in real time, the system can accurately determine the preventive maintenance cycle of the equipment.
[0052] Step 4: Generate a dynamic adaptive maintenance strategy. The central control server uses the standardized environmental dataset, coupling influence coefficient λ, and equipment wear coefficient k as input feature vectors, feeding them into a pre-trained long short-term memory neural network model. This model consists of one input layer, three hidden layers, and one output layer. By recursively processing historical data from the past 72 hours, it predicts the evolution trends of carbon dioxide concentration and particulate matter concentration below 2.5 micrometers within the next 1 to 4 hours. Based on the prediction results, the system initiates a multi-objective optimization algorithm. The objective function aims to minimize the environmental comfort shift, minimize operating energy consumption, and minimize equipment wear rate. During peak shopping hours, the system automatically switches to a comfort-first mode, increasing the frequency of fresh air exchange; during off-peak hours, the system switches to an energy-first mode, utilizing off-peak electricity prices for environmental pretreatment.
[0053] Step 5: Execute closed-loop control and early warning feedback. The central controller drives the supply and return air fans via frequency converters according to the generated maintenance strategy. The control process employs an improved proportional-integral-derivative (PID) control algorithm, with the proportional coefficient adjusted in real-time according to the rate of increase in carbon dioxide concentration. If the measured environmental parameters deviate from the target setpoint by more than 5%, the system initiates secondary compensation adjustment logic to increase the output power of the actuators. Simultaneously, when the wear coefficient k exceeds 0.85, the system determines that the equipment has entered a high-risk operating period and immediately pushes a work order containing fault prediction analysis, a list of recommended replacement parts, and maintenance operating procedures to the maintenance terminal. It also automatically locks the equipment's operating frequency below 70% of its rated value, extending the equipment's lifespan through derating until maintenance is completed.
[0054] Example 2
[0055] This embodiment, based on Embodiment 1, has been deeply optimized for the scenario of a conference center where personnel density fluctuates drastically. Conference centers are characterized by large spaces and rapid changes in personnel density, which places higher demands on the response speed and accuracy of the air environment maintenance system.
[0056] In terms of system architecture, this embodiment adds an infrared thermal imaging sensor array to the perception layer. This sensor array is installed on the ceiling of the conference hall, arranged in a matrix, and can cover the entire venue area. The infrared thermal imaging sensors have a thermal resolution of 320 x 240 pixels, capturing the infrared radiation emitted by the human body to generate a real-time thermal map of the indoor population distribution. The edge computing nodes have a built-in people counting plugin based on a convolutional neural network, which can accurately identify the number of people and their dynamic movement within the room.
[0057] Furthermore, this embodiment incorporates a natural ventilation linkage mechanism in its execution layer. Electric skylights and louvers are installed on the building facade and roof, driven by high-torque synchronous motors. The system also integrates data from an outdoor weather station to obtain real-time outdoor temperature, humidity, wind direction, and atmospheric pressure.
[0058] In terms of workflow derivation, in the data acquisition stage of step 1, this embodiment simultaneously acquires personnel density data provided by infrared thermal imaging sensors. The carbon dioxide load generated by personnel respiration is introduced into the environmental dataset as a dynamic input variable. In the strategy generation stage of step 4, the long short-term memory neural network model uses the number of personnel as an important exogenous variable. When the system detects that the number of people in the conference hall increases by more than 50 within 10 minutes, the prediction model will predict that the carbon dioxide concentration will exceed 1000 parts per million within the next 15 minutes, thereby proactively driving the air handling unit into a high-load operation mode, realizing the transformation from passive response to proactive pre-regulation.
[0059] When implementing environmental regulation, the system prioritizes the feasibility of natural ventilation. If the outdoor air quality index is better than the indoor air quality index, and the outdoor temperature and humidity are within the comfortable range of 22 to 26 degrees Celsius and 40% to 60% humidity, the central controller will prioritize instructing the electric sunroof to open and simultaneously reduce the operating frequency of the air supply fan. Through the dynamic coupling of natural and mechanical ventilation, this embodiment further reduces the operating energy consumption of the air conditioning system while ensuring air quality.
[0060] For sensor maintenance, this embodiment introduces automatic zero-point calibration logic. Since carbon dioxide sensors are prone to zero-point drift during long-term operation, the system automatically switches the solenoid valve to the standard calibration gas path every 720 hours of operation. The calibration gas is 99.99% pure nitrogen, and the system resets the zero-point reference based on the sensor output value under nitrogen conditions. For the total volatile organic compound (TVOC) sensor, reference calibration is performed during the early morning hours when there are few people and air quality is most stable, ensuring the long-term stability and consistency of the monitoring data.
[0061] Example 3
[0062] This embodiment focuses on the full lifecycle management of equipment and big data operation and maintenance support for intelligent air environment maintenance systems. In terms of hardware architecture, this embodiment introduces a global data management platform based on a cloud architecture.
[0063] In terms of system architecture, the global data management platform is interconnected with the central control server within each building via high-performance gigabit Ethernet. The core of the platform consists of a distributed database cluster capable of storing at least five years of environmental operation data, equipment feature vectors, and maintenance work order records. The platform also integrates a big data analytics engine, utilizing massive amounts of historical data to deeply analyze environmental evolution patterns under different seasons and climatic conditions.
[0064] In terms of workflow implementation, this embodiment introduces fault diagnosis logic based on big data comparison in the equipment health assessment step 3. The central control server uploads the extracted equipment operation feature vectors to the cloud platform in real time. By performing pattern matching between the current feature vectors and thousands of fault case features stored in the database, the platform can identify specific fault symptoms such as dust accumulation on fan blades, loose belts, and clogged filters.
[0065] When the system detects a continuous rise in the differential pressure sensor readings on both sides of the filter, and the fan operating current exhibits abnormal fluctuations at the same speed, the big data analytics engine will determine that the filter is severely clogged. At this point, the system will not only generate an early warning in step 5, but will also call upon a big data model to calculate the impact of this clogging state on the system's total energy consumption. If the increase in energy consumption exceeds the depreciation cost of replacing the filter, the system will automatically prioritize maintenance to ensure optimal allocation of operational resources.
[0066] In this embodiment, an emergency mode for extreme pollution weather is added to the closed-loop control logic. When the outdoor concentration of particulate matter smaller than 2.5 micrometers exceeds 200 micrograms per cubic meter, the system automatically shuts down all natural ventilation mechanisms, switches the air handling unit to full return air operation mode, and activates the three-stage high-efficiency filtration system. The electric field voltage of the electronic dust removal and purification unit is automatically increased by 20% to enhance its ability to capture fine particulate matter.
[0067] In addition, the global data management platform also has energy efficiency auditing capabilities. By correlating and analyzing the operational data of the air environment maintenance system with the building's total electricity meter data, the system can automatically calculate the energy consumed to improve air quality by one unit. By comparing the energy efficiency ratios under different operating strategies, the system continuously iterates and optimizes the weighting factors in the multi-objective algorithm, gradually optimizing the energy efficiency weight ratio in energy-saving mode from 0.6 to 0.75, achieving an optimal balance between environmental quality and operating costs.
[0068] Example 4
[0069] This embodiment details the micro-control logic and communication guarantee mechanism of the system at the actuator level.
[0070] In terms of system architecture, to ensure communication reliability in complex building electromagnetic environments, the wireless sensor network in this embodiment employs a communication link with automatic retransmission and error correction mechanisms. Each data frame contains a cyclic redundancy check (CRC) code, and the receiver automatically initiates a retransmission request when a check error is detected. Frequency hopping communication is supported between edge computing nodes and sensor nodes, automatically avoiding interference-prone frequency bands.
[0071] The frequency converters in the execution layer are connected via shielded twisted-pair cables, with a 120-ohm resistor at the end to suppress signal reflection. Each air handling unit integrates an independent logic controller as a slave station. In the extreme case of loss of contact with the central server, the slave controller can operate independently according to the locally preset emergency strategy to ensure the basic environmental safety of the building.
[0072] In terms of workflow derivation, during the closed-loop control process in step 5, the improved proportional-integral-derivative (PID) control algorithm employs anti-saturation logic. When the actuator reaches its maximum output limit, the integral term will stop accumulating to prevent severe overshoot during system adjustment.
[0073] The system employs segmented adjustment logic to control carbon dioxide concentration. When the concentration is below 600 parts per million, the system is in a low-power maintenance mode; when the concentration is between 600 and 1000 parts per million, the proportional gain increases linearly with the concentration difference; when the concentration exceeds 1000 parts per million, the system enters a fast response mode, the proportional gain automatically increases by 20%, and the auxiliary exhaust fan is activated simultaneously.
[0074] In terms of equipment maintenance, this embodiment introduces virtual maintenance simulation logic. Maintenance personnel can view the operating parameters and health status distribution of each device in real time through the 3D visualization interface provided by the global data management platform. Once the system generates a maintenance work order, maintenance personnel can use augmented reality technology to scan the device's QR code with a mobile terminal to obtain the device's internal structure diagram, historical maintenance records, and real-time sensor readings. This combined virtual and real maintenance method significantly shortens fault location time and improves the accuracy of maintenance operations.
[0075] The system's self-learning mechanism is also reflected in its compensation for environmental response lag. Because air conditioning in large spaces has a significant time constant, the system analyzes historical adjustment curves to automatically calculate the delay time between the activation of the actuator and changes in environmental parameters. When generating maintenance strategies, the system issues adjustment commands in advance based on this delay time, achieving proactive compensation for environmental fluctuations and strictly controlling the fluctuation deviation of indoor environmental parameters within a preset threshold of 5%.
[0076] In summary, this invention constructs a multi-dimensional environmental sensing network, combines cross-regional airflow field coupling analysis with equipment operational health assessment, utilizes a long short-term memory neural network model to generate dynamic adaptive maintenance strategies, and ultimately achieves refined maintenance of the air environment in intelligent buildings through closed-loop control and early warning feedback mechanisms. The system not only solves the challenge of multi-space coordinated control but also constructs a state-based preventative operation and maintenance system, significantly improving environmental stability and equipment reliability while reducing operating energy consumption, demonstrating extremely high practical value and technological advancement.
[0077] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.
Claims
1. An intelligent building intelligent air environment maintenance system, characterized in that, include: The multi-dimensional environmental perception module is used to collect environmental parameters in real time, including carbon dioxide concentration, fine particulate matter concentration, total volatile organic compound concentration, temperature and humidity, through a multi-source sensor array deployed in the monitoring area of the smart building. It also uses edge computing nodes to construct a standardized environmental dataset through a median filtering algorithm. The cross-regional flow field analysis module is used to obtain the geometric topological relationship of the monitoring space based on the building information model, calculate the pressure gradient and air exchange volume by combining the opening status of doors and windows and the wind speed of ventilation ducts, establish a cross-regional pollutant diffusion model and determine the coupling influence coefficient. The equipment health assessment module is used to monitor the operating current, fan speed and vibration frequency of the air handling unit in real time, extract the equipment operating feature vector, and calculate the wear coefficient of each actuator by comparing it with the preset equipment health benchmark model, and determine the preventive maintenance cycle. The dynamic strategy generation module is used to input the standardized environmental dataset, the coupling influence coefficient, and the wear coefficient into a pre-trained long short-term memory neural network model to predict the air quality evolution trend within a predetermined time period and generate environmental regulation instructions. The closed-loop control and feedback module is used to drive the air handling equipment to operate, and to activate the secondary compensation adjustment logic when the measured environmental parameters deviate from the target value by more than the first preset threshold, and to push the preventive maintenance warning when the wear coefficient exceeds the second preset threshold. The air handling equipment includes a variable frequency drive (VFD) fan, a return fan, an electronic dust removal and purification unit, a surface cooler, and a humidifier. Each device communicates with the central controller via a bus protocol. The system also includes a global data management platform for storing environmental operation data and maintenance records for a preset time period through a distributed database.
2. A method for maintaining an intelligent air environment in a smart building, applied to the intelligent air environment maintenance system of claim 1, characterized in that, Includes the following steps: Step 1: Real-time collection of environmental parameters including carbon dioxide concentration, fine particulate matter concentration, total volatile organic compound concentration, temperature and humidity is carried out by a multi-source sensor array deployed in the monitoring area of the smart building. The environmental parameters are aggregated to the edge computing node using a wireless sensor network, and the raw data of the environmental parameters are denoised by a median filtering algorithm to construct a standardized environmental dataset. Step 2: Based on the building information model of the intelligent building, obtain the geometric topological relationship of each monitoring space, combine the opening status of doors and windows and the wind speed of ventilation ducts, calculate the pressure gradient and air exchange between different spaces, establish a cross-regional diffusion model of pollutants, and determine the coupling influence coefficient of local environmental fluctuations on adjacent areas. Step 3: Monitor the operating current, fan speed and vibration frequency of the air handling unit in real time, extract the equipment operating feature vector, compare the equipment operating feature vector with the preset equipment health benchmark model, calculate the wear coefficient and performance degradation rate of each actuator, and determine the preventive maintenance cycle of the equipment. Step 4: Input the standardized environmental dataset, the coupling influence coefficient, and the wear coefficient into the pre-trained long short-term memory neural network model to predict the air quality evolution trend within a predetermined time period, and dynamically adjust the ventilation frequency, filtration level, and temperature and humidity setpoints based on the prediction results and equipment status to generate environmental regulation instructions. Step 5: The control system drives the air handling equipment to operate according to the environmental adjustment command and continuously monitors the feedback of the adjusted environmental parameters. If the deviation of the measured value from the target value exceeds the first preset threshold, the secondary compensation adjustment logic is activated, and when the wear coefficient exceeds the second preset threshold, a preventive maintenance warning is pushed to the operation and maintenance terminal.
3. The intelligent building intelligent air environment maintenance method according to claim 2, characterized in that, In step 1, the deployment density of the multi-source sensor array is dynamically determined based on the spatial functional attributes. For office areas, a predetermined number of sensor nodes are deployed in each first preset area, and the installation height of the sensor nodes is within a preset breathing zone height range. For corridor areas, a predetermined number of sensor nodes are deployed in each second preset area. The sensor nodes integrate analog-to-digital converters to convert analog voltage signals into digital sequences and send data packets containing node numbers, timestamps, sensor type codes, measured values, and check codes to the edge computing nodes at a preset sampling frequency.
4. The intelligent building intelligent air environment maintenance method according to claim 2, characterized in that, In step 1, the process of constructing a standardized environmental dataset includes performing normalization processing on environmental parameters. The logic of the normalization processing is as follows: calculate the difference between the measured value of the sensor and the historical minimum value of the parameter, and divide the difference by the range between the historical maximum value and the historical minimum value of the parameter, so as to map environmental parameters of different dimensions to a preset numerical range. The standardized environmental dataset is stored in the temporary buffer of the edge computing node and synchronized to the central control server according to a preset synchronization period.
5. The intelligent building intelligent air environment maintenance method according to claim 2, characterized in that, In step 2, the pollutant cross-regional diffusion model uses the laws of conservation of mass and momentum as physical constraints, and takes into account the chimney effect and wind pressure inside the building. The calculation logic of the coupling influence coefficient is as follows: multiply the effective cross-sectional area of the two connected parts by the average air velocity of the connected parts to obtain the air exchange volume, and then divide the air exchange volume by the total physical volume of the affected space to obtain the coupling influence coefficient. If the coupling influence coefficient exceeds the preset critical threshold, the warning level of the air handling equipment in the adjacent area is increased.
6. The intelligent building intelligent air environment maintenance method according to claim 2, characterized in that, Step 2 further includes real-time correction of the pollutant cross-regional diffusion model by differential pressure sensors installed on both sides of the main partition wall; when the pressure gradient between two adjacent spaces exceeds a preset differential pressure threshold, it is determined that there is an airflow field coupling phenomenon, and the weight coefficient of the coupling influence coefficient in the subsequent strategy generation is increased to a preset multiple; the calculation accuracy of the pressure gradient is dynamically compensated by the real-time feedback of the differential pressure sensor.
7. The intelligent building intelligent air environment maintenance method according to claim 2, characterized in that, In step 3, the equipment operating characteristic vector includes current harmonic distortion, effective value of bearing vibration acceleration, and motor winding temperature. The process of extracting the equipment operation feature vector includes: performing a fast Fourier transform on the vibration frequency signal, extracting the feature frequency components, and comparing them with the reference value under normal operating conditions; If the amplitude at the preset characteristic frequency exceeds the reference value by a preset multiple of the standard deviation, the actuator is determined to have entered the accelerated wear stage.
8. The intelligent building intelligent air environment maintenance method according to claim 2, characterized in that, In step 3, the calculation model for the wear coefficient is as follows: divide the current cumulative operating time of the equipment by the rated design life to obtain the basic wear value, and then use the load factor to correct the basic wear value; the load factor is determined based on the proportion of the equipment's operating time above the rated power; if the average operating frequency of the equipment within the preset operating cycle exceeds the first preset frequency threshold, then the load factor is set as the first preset correction coefficient. If the average operating frequency is lower than the second preset frequency threshold, the load factor is set to the second preset correction coefficient.
9. The intelligent building intelligent air environment maintenance method according to claim 2, characterized in that, In step 4, the long short-term memory neural network model includes a preset number of input layers, hidden layers, and output layers. The hidden layers are recursively connected to capture the time-series features of environmental parameters. The long short-term memory neural network model uses a hyperbolic tangent activation function and learns from a standardized environmental dataset within a preset historical period to output the carbon dioxide concentration evolution curve and the fine particulate matter concentration evolution curve within a predetermined future prediction period. The generation logic of the dynamic adjustment strategy is based on a multi-objective optimization algorithm. The objective functions of the multi-objective optimization algorithm include minimizing the environmental comfort offset, minimizing system operating energy consumption, and minimizing equipment wear rate.
10. The intelligent building intelligent air environment maintenance method according to claim 2, characterized in that, In step 5, the closed-loop control process employs an improved proportional-integral-derivative (PID) control algorithm. The proportional coefficient, integral coefficient, and derivative coefficient of the PID control algorithm are adaptively adjusted in real time according to the severity of fluctuations in environmental parameters. When the rate of increase of carbon dioxide concentration exceeds a preset rate of change threshold, the proportional coefficient increases by a preset percentage. When the wear coefficient exceeds a second preset threshold, the control system automatically locks the operating frequency of the actuator below a preset safe operating ratio and simultaneously generates a work order containing the faulty component number, a list of recommended replacement parts, and maintenance operation procedures.