Central air conditioner installation method and system based on intelligent sensing mechanism

Abstract

The invention provides a central air conditioner installation method and system based on an intelligent sensing mechanism, and the method mainly comprises an installation scheme set acquisition step, a heat source prediction model construction step, an installation scheme dynamic adjustment step and a scheme verification and calibration step. Cooperative optimization of a central air conditioner installation scheme on the aspects of thermal comfort, balance, energy consumption and response speed is achieved, and an efficient solution is provided for accurate deployment of an intelligent air conditioner system of a building.

Description

Technical Field

[0001] The present invention relates to the field of air conditioning installation, and in particular to a central air conditioning installation method and system based on an intelligent sensing mechanism. Background Art

[0002] With the increasing demand for intelligent buildings and energy conservation and emission reduction, the installation scheme of central air-conditioning systems has an increasingly significant impact on indoor thermal environment control and energy consumption. The existing central air-conditioning installation methods have the following main shortcomings: Design solutions rely on experience and lack adaptability. Traditional installation solutions often rely on engineers' experience to determine the location and number of indoor units and air vents, lacking precise alignment with building structure and thermal characteristics. For example, in a mixed space consisting of open office areas and enclosed meeting rooms, a uniform installation model can easily lead to localized temperatures being too high or too low, making it difficult to prioritize thermal comfort in key areas such as workstations. Heat source sensing is static and lacks dynamic adaptability. Existing technologies often make static assumptions about indoor heat sources such as personnel flow, equipment heat dissipation, and solar radiation, lacking dynamic prediction mechanisms that adapt over time and in different scenarios. When heat source distribution suddenly changes, such as when an impromptu meeting results in a large crowd or equipment suddenly operates at high load, the air conditioning system cannot adjust its air supply strategy in a timely manner, resulting in a decrease in temperature uniformity and even the emergence of localized "heat islands" or "cold islands." Traditional optimization solutions often focus on a single objective, such as minimizing energy consumption, while ignoring the collaborative needs of thermal comfort, temperature uniformity, and rapid response. For example, simply pursuing energy reduction can lead to the concentration of air vents in non-key areas, resulting in a delayed cooling rate in office areas and an inability to meet real-time thermal comfort requirements. The lack of a closed-loop verification mechanism limits solution reliability. After installation, deviations between system performance and simulation predictions are difficult to feed back into solution optimization. Deviations caused by factors such as the actual thermal parameters of the building envelope and equipment performance degradation cannot be calibrated, leading to increased energy consumption and decreased comfort levels over long-term operation.

[0003] Therefore, in order to solve the problems existing in the prior art, the present invention proposes a central air-conditioning installation method and system based on an intelligent sensing mechanism. Summary of the Invention

[0004] In view of the deficiencies in the prior art, the present invention aims to provide a central air conditioning installation method and system based on an intelligent sensing mechanism.

[0005] A central air conditioning installation method based on an intelligent sensing mechanism comprises the following steps: An installation solution set acquisition step is performed, wherein the geometric structure and thermal parameters of the installation area are acquired based on the three-dimensional model of the installation area to construct a digital twin, and air conditioning installation data of the same scenario as the digital twin is retrieved from the historical operation database to generate an initial candidate installation solution set including indoor unit location, air outlet direction, and quantity; a heat source prediction model construction step, establishing a dynamic heat source distribution prediction model based on the three-dimensional model of the installation area and combining historical heat source data, and outputting the heat source prediction spatial position, heat source prediction intensity and prediction coverage data of the current installation area as heat source spatial distribution data; The step of dynamically adjusting the installation plan includes simulating the deployment of environmental perception sensors to collect real-time temperature, humidity, and heat source information based on the initial candidate installation plan set and the heat source spatial distribution data, and using a multi-objective optimization algorithm to iterate the candidate plans and output an optimized installation plan set; The solution verification and calibration step is to actually install the air-conditioning system according to the optimized installation solution set, obtain operating data through the actual installation sensor network, compare the deviation value between the measured operating data and the simulation results, correct the digital twin parameters and generate an updated installation solution.

[0006] As a further improvement of the present invention, the step of obtaining the installation solution set includes extracting wall dimensions and door and window coordinates as the installation area set structure through the building information model and mapping the physical structure of the installation area in the digital twin, and extracting the thermal conductivity of the material as a thermal parameter; retrieving a case data set in the historical operation database that is similar to the current installation area structure and purpose, and generating an initial candidate installation solution set based on the indoor unit installation coordinates, air outlet layout angle and quantity combination in the case data set.

[0007] As a further improvement of the present invention, the heat source prediction model construction step includes inputting the thermal parameters, seasonal time variables and equipment operation schedule and loading infrared thermal map samples of the same scene in the historical heat source database, extracting the location intensity characteristics of densely populated areas and equipment clusters, fusing the solar radiation flux data monitored in real time by the light intensity sensor, calculating the theoretical temperature rise curve of each area, and outputting the heat source spatial distribution data with timestamps. The heat source spatial distribution data includes the core coordinates of the heat source as the predicted spatial position, the temperature intensity gradient as the heat source predicted intensity, and the radiation range boundary as the predicted coverage range data.

[0008] As a further improvement of the present invention, the step of dynamically adjusting the installation plan includes marking key monitoring areas in the digital twin according to the heat source core coordinates and radiation boundaries in the heat source distribution data, densely deploying temperature and humidity sensors in the office area with temperature balance as a constraint condition, deploying infrared thermal imagers at the heat source radiation boundary, and collecting air flow velocity, heat source intensity and local temperature and humidity data to drive the digital twin to dynamically update the thermal field distribution.

[0009] As a further improvement of the present invention, the dynamic adjustment step of the installation scheme also includes inputting the initial candidate installation scheme set and the heat source spatial distribution data into the optimization engine, setting the regional temperature priority weight coefficient, the temperature difference threshold between key areas, and the air conditioning thermal efficiency evaluation index, generating a three-dimensional temperature field cloud map through computational fluid dynamics simulation based on the initial candidate installation scheme set, calculating the cooling rate based on the spatial relationship between the air supply path and the heat source core area; calculating the weighted average temperature based on the regional temperature priority weight coefficient, and then calculating the temperature balance index by statistically analyzing the weighted sum of the squares of the temperature differences at the monitoring points, and calculating the optimized installation scheme set through a non-dominated sorting genetic algorithm.

[0010] As a further improvement of the present invention, the temperature balance index calculation includes parsing the regional weight coefficient definition table to obtain the priority weight values ​​of the office area and the open area; extracting the average temperature of each partition in the simulated temperature field, and calculating the global weighted average temperature according to the weight; for each temperature and humidity monitoring point data, calculating its instantaneous difference with the weighted average temperature; multiplying the difference of the office area monitoring point by the weight coefficient and taking the square operation, and directly taking the square operation for the open area monitoring point; finally, calculating the arithmetic mean of all calculation results and outputting it as a quantitative indicator of temperature balance.

[0011] As a further improvement of the present invention, the scheme verification and calibration step includes, after the air-conditioning system is installed and operated according to the optimized installation scheme set, collecting actual temperature and humidity distribution, equipment power and air outlet wind speed data through the sensor network; calculating the measured thermal efficiency index, temperature balance index and energy consumption index; differentiating the measured value and the simulation value item by item, and if the absolute value of the difference of any indicator exceeds the preset tolerance threshold, initiating the deviation tracing analysis: detecting the thermal capacity parameter deviation of the building envelope structure, the heat source intensity mapping error or the equipment air volume-power consumption curve offset; correcting the corresponding parameters in the digital twin through the back propagation algorithm, generating a model correction parameter table and outputting an updated installation scheme; the air-conditioning system includes an air-conditioning indoor unit, an air outlet and a sensor network for real-time data detection, and the sensor network includes a temperature and humidity sensor, a light intensity sensor, an infrared thermal imager, an air volume sensor and an equipment power sensor.

[0012] As a further improvement of the present invention, the generation of the model correction parameter table includes: when the temperature balance index deviation exceeds the limit, the heat source radiation range parameter is reversely adjusted according to the weight ratio of the office area and the open area; when the thermal efficiency index deviation exceeds the limit, the turbulence coefficient of the airflow organization model is corrected according to the spatial topological relationship between the air supply angle and the heat source area; when the energy consumption index deviation exceeds the limit, the slope of the energy efficiency curve is updated based on the ratio of the actual power of the equipment to the simulated air volume; all correction parameters are normalized by the error distribution model of the historical database to form a calibration parameter comparison table that matches the current building scene.

[0013] A central air-conditioning installation system based on an intelligent sensing mechanism, comprising: The installation solution set acquisition module obtains the geometric structure and thermal parameters of the installation area based on the three-dimensional model of the installation area to construct a digital twin. It then calls the air conditioning installation data of the same scenario as the digital twin in the historical operation database to generate an initial candidate installation solution set including the location of the indoor unit, the direction and number of the air outlet; a heat source prediction model building module, which establishes a dynamic heat source distribution prediction model based on the three-dimensional model of the installation area and historical heat source data, and outputs the heat source prediction spatial position, heat source prediction intensity and prediction coverage data of the current installation area as heat source spatial distribution data; The installation plan dynamic adjustment module simulates the deployment of environmental perception sensors to collect real-time temperature, humidity and heat source information based on the initial candidate installation plan set and the heat source spatial distribution data, and uses a multi-objective optimization algorithm to simulate and iterate the candidate plans and output an optimized installation plan set; The solution verification and calibration module actually installs the air-conditioning system according to the optimized installation solution set, obtains operating data through the actual installation sensor network, compares the deviation values ​​of the measured operating data with the simulation results, corrects the digital twin parameters and generates an updated installation solution.

[0014] As a further improvement of the present invention, the environmental perception sensor includes a temperature and humidity sensor, an infrared thermal imager, a light intensity sensor, a wind volume sensor and an equipment power sensor.

[0015] The beneficial effects of the present invention are: Achieving efficient thermal comfort, prioritizing the needs of key areas, the digital twin accurately maps building structure and thermal parameters, combining historical case data to generate initial plans. This ensures that the layout of air conditioning equipment is compatible with the spaces in key areas, such as offices. A dynamic heat source prediction model tracks changes in heat sources such as personnel, equipment, and solar radiation in real time, enabling installation plans to tailor airflow direction and air supply intensity, improving thermal comfort compliance in key areas.

[0016] 2. Improve temperature balance and minimize regional temperature differences. Using a multi-objective optimization algorithm, with temperature balance as a hard constraint, computational fluid dynamics simulation generates a three-dimensional temperature field cloud map to quantitatively evaluate the temperature distribution uniformity of different solutions. The temperature balance index calculation method uses differentiated weighting to prioritize reducing the contribution of temperature differences in key areas, thereby lowering the overall standard deviation of indoor temperature.

[0017] Optimize energy efficiency and reduce system operating energy consumption. While ensuring thermal comfort and temperature balance, a multi-objective optimization algorithm balances the spatial relationship between the air supply path and the core heat source area, reducing ineffective air supply energy consumption. Combining real-time data from equipment power sensors with energy efficiency curve corrections, system operating energy consumption is lower than traditional solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a flow chart of a central air-conditioning installation method based on an intelligent sensing mechanism of the present invention; Figure 2 This is a flow chart of temperature balance index calculation of the present invention; Figure 3 It is a flow chart for constructing the model correction parameter table of the present invention; Figure 4 This is a flow chart of the verification and calibration steps of the present invention; Figure 5 This is a block diagram of a central air-conditioning installation system based on an intelligent sensing mechanism of the present invention. DETAILED DESCRIPTION

[0019] The present invention will be described in further detail below with reference to the accompanying drawings and embodiments. Identical components are denoted by the same reference numerals. It should be noted that the terms "front," "rear," "left," "right," "upper," and "lower" used in the following description refer to directions in the accompanying drawings, and the terms "bottom," "top," "inner," and "outer" refer to directions toward or away from the geometric center of a particular component, respectively.

[0020] The central air-conditioning installation method based on the intelligent perception mechanism disclosed in the embodiment of the present invention realizes the intelligent and precise dynamic optimization of the central air-conditioning installation plan by constructing a full-process technical framework of "digital twin modeling-dynamic heat source prediction-multi-objective optimization iteration-actual measurement closed-loop calibration".

[0021] A central air conditioning installation method based on intelligent sensing mechanism, such as Figures 1 to 4 As shown, the following steps are included: An installation solution set acquisition step is performed, wherein the geometric structure and thermal parameters of the installation area are acquired based on the three-dimensional model of the installation area to construct a digital twin, and air conditioning installation data of the same scenario as the digital twin is retrieved from the historical operation database to generate an initial candidate installation solution set including indoor unit location, air outlet direction, and quantity; In the step of obtaining the installation plan set, the digital twin constructed based on the three-dimensional model of the installation area can accurately map the physical structure and thermal characteristics of the building, solving the problem of insufficient structural adaptability caused by the traditional installation plan's reliance on experience; calling similar scene data in the historical operation database to generate the initial plan set reduces the blindness of the plan design and improves the matching degree between the initial plan and the scene.

[0022] a heat source prediction model construction step, establishing a dynamic heat source distribution prediction model based on the three-dimensional model of the installation area and combining historical heat source data, and outputting the heat source prediction spatial position, heat source prediction intensity and prediction coverage data of the current installation area as heat source spatial distribution data; The heat source prediction model construction step breaks through the limitations of the static heat source assumption by integrating historical heat source data with real-time environmental parameters. It can dynamically output heat source spatial distribution data for different time periods, reducing the heat source position prediction error and intensity prediction deviation, and providing an accurate heat source benchmark for subsequent scheme optimization.

[0023] The step of dynamically adjusting the installation plan includes simulating the deployment of environmental perception sensors to collect real-time temperature, humidity, and heat source information based on the initial candidate installation plan set and the heat source spatial distribution data, and using a multi-objective optimization algorithm to iterate the candidate plans and output an optimized installation plan set; The combination of simulated deployment of environmental perception sensors and multi-objective optimization algorithms enables the iterative evolution of solutions: three-dimensional temperature field cloud maps are generated through computational fluid dynamics simulation to intuitively present the temperature control effects of different solutions; the application of non-dominated sorting genetic algorithms can balance thermal efficiency, temperature balance and energy consumption targets while meeting constraints such as regional temperature priority weights and temperature difference thresholds, thereby increasing the speed at which key areas meet temperature standards and reducing regional temperature differences.

[0024] The solution verification and calibration step is to actually install the air-conditioning system according to the optimized installation solution set, obtain operating data through the actual installation sensor network, compare the deviation value between the measured operating data and the simulation results, correct the digital twin parameters and generate an updated installation solution.

[0025] The solution verification and calibration steps established a dynamic correction mechanism through comparative analysis of measured data and simulation results. When any indicator deviation exceeds the limit, the digital twin parameters can be corrected through the backpropagation algorithm to continuously improve the model prediction accuracy. After several iterations, the degree of consistency between the simulation and measured data gradually increases.

[0026] Specifically, such as Figures 1 to 4 As shown, the steps of obtaining the installation solution set include extracting the wall dimensions and door and window coordinates as the installation area set structure through the building information model and mapping the physical structure of the installation area in the digital twin, and extracting the thermal conductivity of the material as a thermal parameter; retrieving a case data set in the historical operation database that is similar to the current installation area structure and purpose, and generating an initial candidate installation solution set based on the indoor unit installation coordinates, air outlet layout angle and quantity combination in the case data set.

[0027] During the building information extraction phase, the BIM model accurately extracts geometric parameters such as wall dimensions (e.g., office partition wall thickness of 300mm, load-bearing wall thickness of 500mm), and door and window coordinates (e.g., south-facing floor-to-ceiling windows with coordinates X=10.2m, Y=3.5m, and dimensions of 2.4m x 1.8m). These parameters are then directly mapped to the digital twin, ensuring minimal error in the digital restoration of the physical structure. Furthermore, extracted thermal parameters such as material thermal conductivity (1.74W / m・K for reinforced concrete and 2.8W / m・K for glass curtain walls) provide key physical property support for subsequent thermal field simulations. Specifically, these parameters include: A building information model processing engine that supports the IFC format is used to import a high-precision BIM model of the target installation area. This engine automatically identifies enclosure components such as walls, doors, and windows by parsing the BIM model's hierarchical structure, from buildings to floors to rooms and then to components, and extracts their basic attribute information, such as component ID, type, and material association code. For example, in a BIM model generated by Revit, the property sets of the "Wall" and "Door / Window" classes are called through the API interface to filter out non-virtual components and exclude temporary reference lines, annotations, etc., to ensure that the extracted object is a physical entity structure; for the refined extraction of wall geometry parameters, key geometric parameters for wall components are obtained through the following sub-steps: The size parameters are extracted by calling the "Width", "Height" and "Length" attributes of the wall in the BIM model. For non-standard straight walls such as curved walls and inclined walls, the axis coordinates are discretized and sampled at intervals of 0.1m to calculate the curve length and curvature radius. For composite walls such as sandwich insulation walls, the thickness of each material layer is extracted in layers, such as 300mm for the outer concrete layer, 50mm for the insulation layer, and 15mm for the inner gypsum board layer.

[0028] Spatial coordinate positioning uses a fixed reference point on the first floor of the building, such as the axis of the southwest corner column, as the origin to establish a three-dimensional coordinate system with the X axis as the east-west direction, the Y axis as the north-south direction, and the Z axis as the height direction. The three-dimensional coordinates of the starting and ending points of the wall axis are extracted. For example, the axis of the partition wall of an office area is 5.0m, 3.0m, 0.0m to 5.0m, 8.0m, 3.5m. The coordinates of the connection nodes between the wall and the ground and ceiling are recorded to ensure that the spatial position is unambiguous.

[0029] Structural feature marking is performed by identifying whether the wall is load-bearing through material strength properties. For example, a concrete wall is marked as load-bearing, and whether it contains reserved openings such as air duct holes. The center coordinates and dimensions of the openings are recorded, such as the center coordinates of a φ0.3m opening are 5.0m, 5.0m, and 2.5m.

[0030] To accurately extract the geometric parameters of doors and windows, perform the following operations on the door and window components: Coordinate and dimension extraction: By extracting the three-dimensional boundary coordinates of door and window openings, for example, the lower left corner coordinates of a south-facing window are 10.2m, 3.5m, 0.8m, and the upper right corner coordinates are 12.6m, 3.5m, 2.6m, and the width of the door and window is calculated to be 2.4m and the height is 1.8m. For special types of doors such as sliding doors and rotating windows, the opening direction, such as the sliding direction of the sliding door along the X-axis, and the opening angle range, such as the maximum opening of a rotating window is 30°, are additionally extracted.

[0031] Installation attribute association records the installation height of doors and windows, such as the window sill 0.8m from the ground, the frame material such as aluminum alloy, and the glass type such as double-layer insulating glass, and associates them with the corresponding wall ID to clarify their spatial attachment relationship in the building structure.

[0032] Standardization and error checking of geometric parameters automatically convert any imperial units, such as inches, that may appear in the BIM model into international units, such as meters, retaining three decimal places, e.g., 0.254m instead of 10 inches. Collision detection algorithms verify the spatial relationship between walls and doors and windows, for example, checking that door and window openings are completely contained within the corresponding walls, with a deviation of no more than 0.01m, to avoid geometric contradictions such as "floating doors and windows" or "wall penetration." For complex spaces such as polygonal rooms, the angles and closure of each wall are calculated to ensure that the room outline is free of gaps. Three to five key structures, such as the thickness of the main wall and the width of the floor-to-ceiling window, are randomly selected for on-site laser ranging with an accuracy of ±0.005m. If the BIM model-derived value deviates from the measured value by more than 2%, the parameters of these components are calibrated as a whole using the linear correction formula: Correction value = measured value × model value / measured value.

[0033] The associated extraction of thermal parameters uses the material association codes of components in the BIM model to call the built-in building material thermal performance database, which contains the thermal conductivity λ, specific heat capacity c, and density ρ of more than 500 common building materials. It automatically matches and extracts the thermal conductivity of each wall material layer, such as reinforced concrete λ=1.74W / m·K, aerated concrete λ=0.18W / m·K, the thermal transfer coefficient K value of door and window glass, such as double-layer insulating glass K=2.7W / m²·K, and the linear heat transfer coefficient ψ value of frame material, such as aluminum alloy window frame ψ=0.06W / m·K.

[0034] Geometric and thermal property mapping of digital twins, Based on the processed coordinates and size parameters of walls, doors and windows, solid models are generated using parametric modeling tools on digital twin platforms such as Unity and Unreal Engine. Walls are stretched along the height direction to form three-dimensional entities using stretching modeling. Doors and windows use Boolean operations to subtract the opening geometry from the corresponding wall positions and are assigned spatial coordinates consistent with the BIM model, ensuring that the spatial mapping error between the digital twin and the actual building is ≤0.05m.

[0035] In the digital twin's physics engine module, the extracted parameters, such as thermal conductivity and heat transfer coefficient, are associated with the material property sets of the corresponding geometric components. For example, the material node of the "south exterior wall" component is assigned a value of λ=1.74W / m·K, and the material node of the "office area window" is assigned a value of K=2.7W / m²·K, giving the digital twin a physical basis for heat conduction simulation.

[0036] Structured attribute tags such as "wall: thickness 0.3m, material concrete, thermal conductivity 1.74" and "WD-02# window: coordinates 10.2, 3.5, 0.8-12.6, 3.5, 2.6, glass type double-layer hollow" are embedded in the digital twin to provide queryable structural feature data for subsequent historical case retrieval and solution generation.

[0037] Through the above steps, the digital twin can accurately reproduce the physical structure and thermal characteristics of the installation area, and its geometric parameters are significantly more consistent with the actual building, providing a high-fidelity digital foundation for the generation of initial candidate installation plans.

[0038] The historical case retrieval process utilizes a three-level similarity matching mechanism: first-level matching is based on building structure similarity, such as room outline overlap ≥80%; second-level matching is based on functional consistency, such as open-plan office areas; and third-level matching is based on heat source feature similarity, such as equipment density and peak occupancy deviation ≤20%. Based on the matching results, the 5-10 best cases were selected from the historical database. Their indoor unit installation coordinates, such as 2.5m from the west wall and 3.0m from the north wall, air outlet layout angles, such as 15° horizontal outlet deflection and 10° vertical outlet pitch, and quantity combinations, such as two 3-horsepower indoor units with eight outlets, were extracted. Through parameterized combinations, an initial set of candidate installation solutions, consisting of 20-30 solutions, was generated, significantly reducing the trial-and-error cost of solution design.

[0039] Specifically, such as Figures 1 to 4As shown, the heat source prediction model construction steps include inputting the thermal parameters, seasonal time variables and equipment operation schedule and loading infrared thermal map samples of the same scene in the historical heat source database, extracting the location intensity characteristics of densely populated areas and equipment clusters, fusing the solar radiation flux data monitored in real time by the light intensity sensor, calculating the theoretical temperature rise curve of each area, and outputting the heat source spatial distribution data with timestamps. The heat source spatial distribution data includes the core coordinates of the heat source as the predicted spatial position, the temperature intensity gradient as the heat source predicted intensity, and the radiation range boundary as the predicted coverage range data.

[0040] The heat source prediction model construction step achieves accurate prediction of heat source distribution through multi-dimensional parameter fusion and dynamic feature extraction, providing a dynamic benchmark for installation solution optimization.

[0041] The input parameters for this step encompass both static and dynamic categories. Static parameters include building thermal parameters such as the heat transfer coefficient of the building envelope and spatial geometric features such as room depth and window opening ratio. Dynamic parameters include seasonal variables such as the sun's altitude angle from 10:00 AM to 4:00 PM on a typical summer day, equipment operation schedules such as full load operation in the server room from 8:00 AM to 8:00 PM, and real-time solar radiation flux collected by light intensity sensors in W / m². By loading over 500 infrared thermogram samples of the same scene from a historical heat source database, a deep learning algorithm was used to extract the elliptical distribution characteristics of densely populated areas, such as office workstation clusters, and the location and intensity characteristics of rectangular high-temperature areas in equipment clusters, such as server cabinets, to form a heat source feature library.

[0042] During the calculation process, the heat gain of the window area is first calculated based on solar radiation flux data. For example, when each square meter of glass receives 1000W of solar radiation at noon, the local temperature rise rate is 0.8°C / h. The theoretical heating power is calculated by combining equipment power (e.g., 200W for a single computer, 300W for a printer), and 100W for each person. This is then integrated with historical temperature rise curves for the same period to generate theoretical temperature rise curves for each area. The final output is time-stamped heat source spatial distribution data, including the core location of the heat source, represented by coordinates (e.g., X=5.2m, Y=3.8m), the predicted heat source intensity (e.g., a temperature gradient of 5°C / m), and the predicted coverage area (e.g., a 3m diameter circular area), defined by a 3°C temperature difference boundary. This allows subsequent solution optimization to accurately match the dynamic changes of the heat source.

[0043] Specifically, such as Figures 1 to 4 As shown, the dynamic adjustment step of the installation plan includes marking key monitoring areas in the digital twin according to the heat source core coordinates and radiation boundaries in the heat source distribution data, densely deploying temperature and humidity sensors in the office area with temperature balance as a constraint condition, deploying infrared thermal imagers at the heat source radiation boundary, and collecting air flow velocity, heat source intensity and local temperature and humidity data to drive the digital twin to dynamically update the thermal field distribution.

[0044] The sensor deployment strategy in the dynamic adjustment step of the installation plan provides real-time data support for plan optimization by accurately covering key areas and dynamically driving digital twin updates.

[0045] Based on the core coordinates and radiation boundaries of the heat source in the heat source distribution data, key monitoring areas are first marked in the digital twin. For key areas such as office areas, temperature and humidity sensors are densely deployed at a density of one per 2-3 square meters to ensure that the distance between monitoring points is ≤1.5m, and there is at least one sensor near each workstation; infrared thermal imagers are deployed at the radiation boundaries of the heat source, such as the edge of the equipment heat dissipation area, and a scanning frequency of 15 minutes / time is used to obtain the regional thermal distribution. The resolution reaches 640×512 pixels and can identify temperature differences of 0.5°C.

[0046] Data collected by the sensors includes real-time temperature accuracy of ±0.3°C and humidity accuracy of ±2%RH from the temperature and humidity sensors, heat source intensity error of ≤1°C from the infrared thermal imager, and wind speed data from the air velocity sensor with a range of 0.1-10 m / s and an accuracy of ±0.05 m / s. This data is uploaded to the digital twin system in real time via a wireless transmission module, driving dynamic updates of the thermal field distribution. The three-dimensional temperature field model is refreshed every five minutes, keeping the deviation between the simulated environment and the actual environment within 3%, providing a high-fidelity computational foundation for subsequent optimization algorithms. Furthermore, redundant sensor network deployment, with sensors deployed in key areas, ensures continuous data collection and avoids optimization interruptions caused by single point failures.

[0047] Specifically, such as Figures 1 to 4 As shown, the dynamic adjustment step of the installation scheme also includes inputting the initial candidate installation scheme set and the heat source spatial distribution data into the optimization engine, setting the regional temperature priority weight coefficient, the temperature difference threshold between key areas, and the air conditioning thermal efficiency evaluation index, generating a three-dimensional temperature field cloud map through computational fluid dynamics simulation based on the initial candidate installation scheme set, calculating the air conditioning thermal efficiency evaluation index according to the spatial relationship between the air supply path and the heat source core area; calculating the weighted average temperature according to the regional temperature priority weight coefficient, and then calculating the temperature balance index by statistically analyzing the weighted sum of the squares of the temperature differences at the monitoring points, and calculating the optimized installation scheme set through the non-dominated sorting genetic algorithm.

[0048] The multi-objective optimization engine in the dynamic adjustment step of the installation plan achieves the global optimal solution of the installation plan by constructing a multi-dimensional evaluation system and intelligent algorithm iteration.

[0049] The optimization engine's input parameters include an initial set of 20-30 candidate solutions, heat source spatial distribution data with timestamps and constraints, regional temperature priority weights (e.g., 0.7 for offices, 0.2 for corridors, and 0.1 for restrooms), a temperature difference threshold of ≤2°C between key areas, and air conditioning thermal efficiency evaluation metrics (e.g., a cooling rate ≥1°C / 10 minutes). During the calculation process, each candidate solution is first simulated using computational fluid dynamics (CFD) simulation, setting a supply air temperature of 16°C and an initial wind speed of 3m / s. The simulation lasts one hour, generating a three-dimensional temperature field cloud map containing over 5,000 monitoring points. The cloud map has a color temperature resolution of 0.1°C, clearly showing temperature stratification and airflow blind spots.

[0050] The thermal efficiency evaluation index is calculated based on the spatial overlap between the airflow path and the heat source core area. When the airflow covers more than 70% of the heat source core area, the cooling rate is calculated according to the actual simulation value, such as 1.2°C / 10 minutes. When the coverage is less than 30%, the cooling rate is reduced by 50%. The temperature balance index is calculated using the weighted square difference method to ensure that the temperature difference in key areas is given higher weight.

[0051] The application of a non-dominated sorting genetic algorithm achieved multi-objective optimization. The algorithm population size was set to 100, with a crossover probability of 0.8 and a mutation probability of 0.05. After 50 iterations, the top five solutions in the Pareto optimal solution set were selected as the optimized installation solution set. This process was 80% more efficient than traditional trial-and-error methods and simultaneously met thermal comfort, balance, and energy efficiency objectives.

[0052] The implementation process of the multi-objective optimization engine is the core of the dynamic adjustment step of the installation plan. Through the closed-loop process of "parameter input-simulation calculation-indicator quantification-algorithm iteration-plan output", the global optimal screening of candidate plans is achieved. The specific steps are as follows: The input parameters of the optimization engine are divided into three categories: basic data, constraints, and target weights. The specific sources and setting rules are as follows: The initial candidate installation plan set includes 20-30 groups of plans. Each group of plans clearly defines the three-dimensional coordinates of the indoor unit installation, such as x=5m, y=5m, z=3m, the number of air outlets, such as 4 air outlets + 2 return air outlets, and the air outlet direction parameters, horizontal deflection angle ±30° and vertical pitch angle 5°-15°. The data comes from the historical case matching and parameter combination described in Claim 2.

[0053] Heat source spatial distribution data, heat source core coordinates with timestamps such as 8m, 6m, 0m in densely populated areas, 15m, 12m, 0m in equipment areas, temperature intensity gradients such as 2°C / m, and radiation range boundaries such as a circular area with a diameter of 2m surrounded by a 26°C isotherm, are output by the dynamic heat source prediction model described in claim 3.

[0054] Air conditioning equipment performance parameters, call the air volume-static pressure curve in the equipment database such as air volume 300-800 m³ / h corresponding to static pressure 50-200 Pa, refrigerating capacity curve such as refrigerating capacity 3.5 kW when the ambient temperature is 25℃, and energy efficiency ratio EER data, as the basis for energy consumption calculation.

[0055] Regional temperature priority weight coefficient, determined by analyzing the user's preset "regional weight definition table", such as office area key area weight 0.7, corridor non-key area weight 0.2, and toilet weight 0.1, to ensure that the optimization process is inclined to key areas.

[0056] Key area temperature difference threshold, set a hard constraint according to the functional requirements of the building, such as ΔT_max=2℃, if the temperature difference between the office area and the conference room in the simulation exceeds 2℃, the scheme is directly determined as unfeasible.

[0057] Thermal efficiency evaluation benchmark, define the minimum cooling rate of the heat source core area, such as ≥1℃ / 10min, schemes below this value are eliminated in the preliminary screening.

[0058] Computational fluid dynamics (CFD) simulation modeling and running, for each set of schemes in the initial candidate scheme set, simulate the indoor thermal environment and air distribution after air conditioning operation through CFD simulation, the specific process is as follows: Simulation boundary condition setting, the space boundary is based on the digital twin built in claim 2, import the geometric parameters of walls, doors and windows such as wall thickness 300mm, glass window size 2.4m×1.8m and thermal parameters such as wall thermal conductivity 1.74W / m·K, set the building envelope as an adiabatic boundary to simplify the influence of environmental heat transfer in short-term simulation. The device operation parameters are uniformly set, such as supply air temperature 16℃, initial air speed 3m / s, air outlet size fixed according to device model such as 0.3m×0.3m, and return air outlet as natural return air pressure 0Pa. The heat source parameters convert the heat source space distribution data into heat boundary conditions in simulation, such as setting the heat dissipation of personnel area as 100W / person and the constant heat flux density of equipment area as 800W / table.

[0059] Structural grid division is performed on the installation area, the grid size of key areas such as office area and heat source core area is encrypted to 0.1m×0.1m×0.1m, the grid size of non-key areas is relaxed to 0.5m×0.5m×0.5m, the total number of grids is controlled within 500-1000, to ensure the balance between calculation accuracy and efficiency.

[0060] The finite volume method is used to solve the Navier-Stokes equation and the energy equation, the iteration step is set to 5000 steps to ensure the convergence of the flow field and temperature field, and the simulation time is 1 hour to cover the whole period from air conditioning start to stable operation.

[0061] Output result: Generate a three-dimensional temperature field cloud map containing 5000+ monitoring point temperature data, resolution 0.1℃, airflow velocity vector diagram to show the air supply path and vortex area, temperature change curve of heat source area over time, provide data support for subsequent index calculation.

[0062] Multi-dimensional evaluation index quantitative calculation, based on CFD simulation results, calculate the heat efficiency, temperature balance, and energy consumption of each candidate scheme, the specific calculation method is as follows: Heat efficiency index Eff_heat: Take the cooling effect of air conditioner air on the heat source core area as the quantitative basis, the calculation steps are as follows: Extract the temperature change curve of the heat source core area such as the diameter 2m range, determine the temperature drop value ΔT in 10 minutes such as from 30℃ to 28.8℃, ΔT=1.2℃.

[0063] Calculate the spatial overlap of the air supply path and the heat source core area: through vector analysis, it is determined that the angle between the air outlet airflow direction and the heat source core point is ≤30° for high coverage, ≥60° for low coverage, the higher the coverage, the greater the weight coefficient of the cooling rate, such as coverage 80%, weight coefficient 0.8.

[0064] The final heat efficiency index = ΔT × coverage weight coefficient, such as 1.2℃ × 0.8 = 0.96, the higher the value, the better the cooling effect.

[0065] Temperature balance index ΔT_balanced: Analyze the area weight table to determine the office area weight 0.7 and the open area weight 0.3.

[0066] Extract the average temperature of each subarea in the simulation temperature field, such as office area 25.2℃, corridor 26.8℃, calculate the global weighted average temperature T_weighted=25.2×0.7+26.8×0.3=25.7℃.

[0067] Calculate the instantaneous difference between each monitoring point and T_weighted, multiply the difference value of the office area monitoring point by the weight 0.7 and take the square, and directly take the square of the open area, such as the difference value of an office area point -0.5℃, calculated as -0.5×0.7²=0.1225; the difference value of an open area point +1.1℃, calculated as 1.21.

[0068] Take the arithmetic mean of the square values of all monitoring points to get the temperature balance index, such as 0.28, the smaller the value, the more balanced the temperature distribution.

[0069] Energy consumption index E_rel: Based on the air volume data in the simulation, such as the average air volume of 500 m³ / h and the equipment power curve, such as the air volume of 500 m³ / h corresponding to the power of 1.2 kW, the theoretical hourly power consumption of the calculation scheme is 1.2 kW × 1h = 1.2 kWh.

[0070] Set the hourly power consumption of the benchmark plan, such as the plan with the lowest energy consumption in the historical case, as E_base, such as 1.5kWh, then the relative energy consumption E_rel=1.2 / 1.5=0.8. The smaller the value, the more energy-saving.

[0071] Iterative Optimization of Non-Dominated Sorting Genetic Algorithm The non-dominated sorting genetic algorithm NSGA-Ⅱ is used to perform multi-objective optimization on candidate solutions. The optimal solution is selected through the iterative process of "population initialization-non-dominated sorting-selection-crossover-mutation". The specific process is as follows: Algorithm parameter settings: Population size, 100. Each iteration processes 100 sets of solutions simultaneously, including initial candidate solutions and iteratively generated solutions.

[0072] Genetic operator, with a crossover probability of 0.8, randomly selects the equipment position and tuyere angle parameters of two groups of schemes for crossover combination, and with a mutation probability of 0.05, randomly changes the number or direction of tuyere of a scheme by ±5°.

[0073] The termination condition is that after 50 generations of iteration, if the optimal solution of 5 consecutive generations has no significant change and the fluctuation index is ≤5%, the iteration is stopped.

[0074] The iterative process involves randomly selecting 100 supplementary schemes, including duplicate schemes, from 20-30 initial candidate schemes, and calculating the thermal efficiency, temperature balance index, and energy consumption index of each group. Schemes are graded according to the "Pareto optimality" principle. That is, if all indicators of scheme A are better than those of scheme B, then A dominates B and B is eliminated; the undominated schemes enter the next level of sorting until all schemes are graded. Using the tournament selection method, schemes are randomly selected from each level for comparison, and the schemes with better indicators are retained for the crossover phase. The selected schemes are subjected to parameter reorganization, such as combining the indoor unit position of scheme A with the air outlet direction of scheme B, and random mutation, such as adjusting a certain air outlet angle, to generate new schemes to supplement the population. The sorting-selection-crossover-mutation process is repeated, retaining the top 20% of the best schemes in each generation, and gradually improving the overall performance of the population.

[0075] Optimal solution screening: After the iteration is terminated, 5-8 groups of solutions are usually selected from the Pareto optimal solution set of the final population. The weight coefficients can be adjusted according to user needs based on the principle of "thermal efficiency first, taking into account temperature balance and energy consumption", such as W_heat=0.4, W_balance=0.3, W_energy=0.3. The comprehensive score is calculated and the 5 groups of solutions with the highest scores are selected as the output of the optimized installation solution set.

[0076] The output optimized installation solution set contains the following information: Equipment parameters: three-dimensional coordinates of indoor unit installation, number of air outlets (e.g. 8 air outlets for 2 indoor units), horizontal deflection angle of 15° and pitch angle of 10°.

[0077] The thermal efficiency of each set of solutions is 0.92, the temperature balance index is 0.22, and the relative energy consumption is 0.85.

[0078] The three-dimensional temperature field cloud map and airflow path simulation animation intuitively demonstrate the temperature control effect of the solution.

[0079] Through this process, the optimization engine can achieve the comprehensive optimization goals of a 30% increase in thermal efficiency, a 40% reduction in the temperature balance index, and a 20% reduction in energy consumption while meeting the hard constraint of a temperature difference of ≤2°C in key areas. This improves the adaptability of the solution by more than 50% compared to traditional single-objective optimization methods.

[0080] Specifically, such as Figures 1 to 4 As shown, the temperature balance index calculation includes parsing the regional weight coefficient definition table to obtain the priority weight values ​​of the office area and the open area; extracting the average temperature of each partition in the simulated temperature field, and calculating the global weighted average temperature according to the weight; for each temperature and humidity monitoring point data, calculating its instantaneous difference with the weighted average temperature; multiplying the difference of the office area monitoring point by the weight coefficient and taking the square operation, and directly taking the square operation for the open area monitoring point; finally, calculating the arithmetic mean of all calculation results and outputting it as the temperature balance quantitative index.

[0081] The quantitative calculation method of the temperature balance index achieves an objective evaluation of indoor temperature balance through differentiated weight design and precise data processing.

[0082] The calculation process is divided into five steps: the first step is to analyze the regional weight coefficient definition table to clarify that the priority weight values ​​of key areas such as office areas and conference rooms are 0.6-0.8, and the priority weight values ​​of open areas such as corridors and storage rooms are 0.2-0.4; the second step is to divide the simulated temperature field into 50cm×50cm grids and extract the average temperature of each partition, such as the average temperature of a grid in the office area is 24.5℃ and the average temperature of a grid in the corridor is 26.2℃; the third step is to calculate the global weighted average temperature according to the weight, such as 24.5×0.7+26.2×0.3=25.01 ℃; the fourth step is to calculate the instantaneous difference between each monitoring point and the weighted average temperature, such as 24.8℃ at a point in the office area, the difference is +0.21℃; 26.0℃ at a point in the corridor, the difference is +0.99℃; the fifth step is to multiply the difference of the monitoring points in the office area by the weight coefficient 0.7 and take the square, such as 0.21×0.7²=0.022, and directly take the square of the monitoring points in the open area, such as 0.99²=0.98. Finally, the arithmetic mean of all calculation results is calculated, such as 0.022+0.98 / total number of monitoring points, to obtain the temperature balance index, such as 0.35.

[0083] The smaller the index value, the better the temperature balance. When the index is ≤0.5, it is determined to meet the temperature balance requirements. Through differentiated weight design, the temperature deviation in key areas such as office areas has a greater impact on the index, which meets the invention goal of "prioritizing key areas."

[0084] Specifically, such as Figures 1 to 4 As shown, the scheme verification and calibration steps include, after the air-conditioning system is installed and operated according to the optimized installation scheme set, collecting actual temperature and humidity distribution, equipment power and air outlet wind speed data through the sensor network; calculating the measured thermal efficiency index, temperature balance index and energy consumption index; differentiating the measured value and the simulation value item by item, and if the absolute value of the difference of any indicator exceeds the preset tolerance threshold, starting the deviation tracing analysis: detecting the deviation of the thermal capacity parameter of the building envelope structure, the heat source intensity mapping error or the equipment air volume-power consumption curve offset; correcting the corresponding parameters in the digital twin through the back propagation algorithm, generating a model correction parameter table and outputting an updated installation scheme; the air-conditioning system includes an air-conditioning indoor unit, an air outlet and a sensor network for real-time detection of data, and the sensor network includes a temperature and humidity sensor, a light intensity sensor, an infrared thermal imager, an air volume sensor and an equipment power sensor.

[0085] The solution verification and calibration steps achieve continuous optimization of the digital twin and dynamic updating of the installation solution by building a closed loop of "measurement-comparison-traceability-correction".

[0086] During system operation, a sensor network consisting of temperature and humidity, infrared thermal imaging, power, and air volume sensors collects data every 30 seconds. Actual temperature and humidity distribution is presented as a heat map, with device power accuracy to ±5W and air velocity at the air outlet measured within a range of 0-10m / s with an accuracy of ±0.1m / s. Three core metrics are calculated based on this data: thermal efficiency (the ratio of the actual cooling rate to the target value); temperature balance (calculated using the method described in Claim 6); and energy consumption (daily power consumption per unit area, in kWh / m²).

[0087] When the absolute value of the difference in any indicator exceeds the preset tolerance threshold of thermal efficiency ±10%, temperature balance index ±0.1, and energy consumption ±15%, the deviation tracing analysis is initiated: by comparing the measured thermal capacity parameters of the building envelope structure, such as the actual thermal capacity of the wall is 12% higher than the model value, the heat source intensity mapping error, such as the actual heat generation of the equipment is 150W higher than the predicted value, and the offset of the equipment air volume-power consumption curve, such as the power consumption is 80W higher than the model value when the actual air volume is 300m³ / h, the source of the deviation is located.

[0088] The correction process utilizes a backpropagation algorithm. For thermal capacity parameter deviations, the corresponding parameters in the digital twin are adjusted based on the ratio of the measured value to the model value, for example, 1.12. For heat source intensity errors, a correction factor, such as 1.15, is added to the heat source prediction model. For equipment curve offsets, the air volume-power consumption function is refitted, for example, from y = 0.5x + 100 to y = 0.6x + 120. After the corrections, a model correction parameter table is generated, and the installation plan is updated based on the new parameter output, improving the system's adaptability in complex environments by over 30%.

[0089] Specifically, such as Figures 1 to 4 As shown, the generation of the model correction parameter table includes: when the temperature balance index deviation exceeds the limit, the heat source radiation range parameter is reversely adjusted according to the weight ratio of the office area and the open area; when the thermal efficiency index deviation exceeds the limit, the turbulence coefficient of the airflow organization model is corrected according to the spatial topological relationship between the air supply angle and the heat source area; when the energy consumption index deviation exceeds the limit, the energy efficiency curve slope is updated based on the ratio of the actual power of the equipment to the simulated air volume; all correction parameters are normalized by the error distribution model of the historical database to form a calibration parameter comparison table that matches the current building scene.

[0090] The process of generating the model correction parameter table achieves accurate correction of digital twin parameters through sub-indicator deviation processing and normalization calibration.

[0091] When the temperature balance index deviation exceeds the limit, the heat source radiation range parameter is adjusted in reverse according to the weight ratio of the office area and the open area, for example 7:3: if the actual temperature difference in the office area is too large, the heat source radiation range will be expanded by 10%×0.7 based on the original prediction, and the radiation range of the open area will be reduced by 5%×0.3, so that the heat source coverage of key areas is more accurate.

[0092] When the thermal efficiency index deviation exceeds the limit, the turbulence coefficient of the airflow organization model is corrected according to the spatial topological relationship between the air supply angle and the heat source area: when the angle between the air supply direction and the normal of the heat source area exceeds 30°, the turbulence coefficient is adjusted from 0.025 to 0.03, an increase of 120%, to enhance the airflow diffusion capacity; when the angle is less than 15°, the turbulence coefficient is reduced to 0.02, reducing airflow loss.

[0093] If the energy consumption index deviation exceeds the limit, the slope of the energy efficiency curve is updated based on the ratio of the actual power of the equipment to the simulated air volume. For example, if the actual power / simulated air volume = 0.8kW / 1000m³ / h, and the model value is 0.6kW / 1000m³ / h, the slope of the energy efficiency curve is corrected from 0.6 to 0.8.

[0094] All correction parameters must be normalized using the error distribution model of the historical database, such as the normal distribution model, with a confidence interval of 95%, to ensure that the correction value is within the allowable error range of similar scenarios. Ultimately, a calibration parameter comparison table matching the current building scenario is formed, containing parameter name, original value, correction value, and correction basis, providing a quantitative basis for subsequent program updates.

[0095] A central air conditioning installation system based on intelligent sensing mechanism, such as Figure 5 As shown, including: The installation solution set acquisition module obtains the geometric structure and thermal parameters of the installation area based on the three-dimensional model of the installation area to construct a digital twin. It then calls the air conditioning installation data of the same scenario as the digital twin in the historical operation database to generate an initial candidate installation solution set including the location of the indoor unit, the direction and number of the air outlet; a heat source prediction model building module, which establishes a dynamic heat source distribution prediction model based on the three-dimensional model of the installation area and historical heat source data, and outputs the heat source prediction spatial position, heat source prediction intensity and prediction coverage data of the current installation area as heat source spatial distribution data; The installation plan dynamic adjustment module simulates the deployment of environmental perception sensors to collect real-time temperature, humidity and heat source information based on the initial candidate installation plan set and the heat source spatial distribution data, and uses a multi-objective optimization algorithm to simulate and iterate the candidate plans and output an optimized installation plan set; The solution verification and calibration module actually installs the air-conditioning system according to the optimized installation solution set, obtains operating data through the actual installation sensor network, compares the deviation values ​​of the measured operating data with the simulation results, corrects the digital twin parameters and generates an updated installation solution.

[0096] Specifically, such as Figure 5 As shown, the environmental perception sensors include temperature and humidity sensors, infrared thermal imagers, light intensity sensors, wind volume sensors and equipment power sensors.

[0097] The above shows and describes the basic features, principles, and advantages of the present invention. It should be noted that the present invention is not limited to the above embodiments, which are only some embodiments. Without departing from the spirit and scope of the present invention, various improvements and supplements made are considered to be within the scope of protection of the present invention.

Claims

1. A central air conditioning installation method based on intelligent perception mechanism, characterized in that: The steps include: An installation solution set acquisition step is performed, wherein the geometric structure and thermal parameters of the installation area are acquired based on the three-dimensional model of the installation area to construct a digital twin, and air conditioning installation data of the same scenario as the digital twin is retrieved from the historical operation database to generate an initial candidate installation solution set including indoor unit location, air outlet direction, and quantity; a heat source prediction model construction step, establishing a dynamic heat source distribution prediction model based on the three-dimensional model of the installation area and combining historical heat source data, and outputting the heat source prediction spatial position, heat source prediction intensity and prediction coverage data of the current installation area as heat source spatial distribution data; The step of dynamically adjusting the installation plan includes simulating the deployment of environmental perception sensors to collect real-time temperature, humidity, and heat source information based on the initial candidate installation plan set and the heat source spatial distribution data, and using a multi-objective optimization algorithm to iterate the candidate plans and output an optimized installation plan set; The solution verification and calibration step is to actually install the air-conditioning system according to the optimized installation solution set, obtain operating data through the actual installation sensor network, compare the deviation value between the measured operating data and the simulation results, correct the digital twin parameters and generate an updated installation solution.

2. A central air conditioning installation method based on intelligent perception mechanism according to claim 1, characterized in that: The step of obtaining the installation solution set includes extracting wall dimensions and door and window coordinates as the installation area set structure through the building information model, mapping the physical structure of the installation area in the digital twin, and extracting the thermal conductivity of the material as a thermal parameter; retrieving a case data set in the historical operation database that is similar to the current installation area structure and purpose, and generating an initial candidate installation solution set based on the indoor unit installation coordinates, air outlet layout angles and quantity combinations in the case data set.

3. A central air conditioning installation method based on intelligent perception mechanism according to claim 1, characterized in that: The heat source prediction model construction step includes inputting the thermal parameters, seasonal time variables and equipment operation schedule and loading infrared thermal map samples of the same scene in the historical heat source database, extracting the location intensity characteristics of densely populated areas and equipment clusters, integrating the solar radiation flux data monitored in real time by the light intensity sensor, calculating the theoretical temperature rise curve of each area, and outputting the heat source spatial distribution data with timestamps. The heat source spatial distribution data includes the core coordinates of the heat source as the predicted spatial position, the temperature intensity gradient as the heat source predicted intensity, and the radiation range boundary as the predicted coverage range data.

4. A central air conditioning installation method based on intelligent perception mechanism according to claim 1 or 3, characterized in that: The dynamic adjustment step of the installation plan includes marking key monitoring areas in the digital twin based on the heat source core coordinates and radiation boundaries in the heat source distribution data, densely deploying temperature and humidity sensors in the office area with temperature balance as a constraint condition, deploying infrared thermal imagers at the heat source radiation boundary, and collecting airflow velocity, heat source intensity and local temperature and humidity data to drive the digital twin to dynamically update the thermal field distribution.

5. The central air conditioning installation method based on intelligent perception mechanism according to claim 1 is characterized in that: The dynamic adjustment step of the installation plan also includes inputting the initial candidate installation plan set and the heat source spatial distribution data into the optimization engine, setting the regional temperature priority weight coefficient, the temperature difference threshold between key areas, and the air conditioning thermal efficiency evaluation index, generating a three-dimensional temperature field cloud map through computational fluid dynamics simulation based on the initial candidate installation plan set, calculating the air conditioning thermal efficiency evaluation index based on the spatial relationship between the air supply path and the heat source core area; calculating the weighted average temperature based on the regional temperature priority weight coefficient, and then calculating the temperature balance index by statistically analyzing the weighted sum of the squares of the temperature differences at the monitoring points, and calculating the optimized installation plan set through a non-dominated sorting genetic algorithm.

6. A central air conditioning installation method based on intelligent perception mechanism according to claim 5, characterized in that: The temperature balance index calculation includes parsing the regional weight coefficient definition table to obtain the priority weight values ​​of the office area and the open area; extracting the average temperature of each partition in the simulated temperature field and calculating the global weighted average temperature according to the weight; calculating the instantaneous difference between the data of each temperature and humidity monitoring point and the weighted average temperature; multiplying the difference of the office area monitoring point by the weight coefficient and taking the square operation, and directly taking the square operation for the open area monitoring point; finally, calculating the arithmetic mean of all calculation results and outputting it as the temperature balance quantitative index.

7. A central air conditioning installation method based on intelligent perception mechanism according to claim 1, characterized in that: The solution verification and calibration step includes collecting actual temperature and humidity distribution, equipment power, and air outlet wind speed data through a sensor network after the air conditioning system is installed and operated according to the optimized installation solution set; Calculate the measured thermal efficiency index, temperature balance index and energy consumption index; differentiate the measured value and the simulated value item by item. If the absolute value of the difference of any index exceeds the preset tolerance threshold, start the deviation tracing analysis: detect the deviation of the thermal capacity parameter of the building envelope structure, the heat source intensity mapping error or the offset of the equipment air volume-power consumption curve; correct the corresponding parameters in the digital twin through the back propagation algorithm, generate a model correction parameter table and output an updated installation plan; the air-conditioning system includes an air-conditioning indoor unit, an air vent and a sensor network for real-time data detection, and the sensor network includes temperature and humidity sensors, light intensity sensors, infrared thermal imagers, air volume sensors and equipment power sensors.

8. A central air conditioning installation method based on intelligent perception mechanism according to claim 7, characterized in that: The generation of the model correction parameter table includes: when the temperature balance index deviation exceeds the limit, the heat source radiation range parameter is reversely adjusted according to the weight ratio of the office area and the open area; when the thermal efficiency index deviation exceeds the limit, the turbulence coefficient of the airflow organization model is corrected according to the spatial topological relationship between the air supply angle and the heat source area; when the energy consumption index deviation exceeds the limit, the slope of the energy efficiency curve is updated based on the ratio of the actual power of the equipment to the simulated air volume; all correction parameters are normalized by the error distribution model of the historical database to form a calibration parameter comparison table that matches the current building scenario.

9. A central air conditioning installation system based on an intelligent sensing mechanism, applicable to a central air conditioning installation method based on an intelligent sensing mechanism as claimed in any one of claims 1 to 8, characterized in that: include: The installation solution set acquisition module obtains the geometric structure and thermal parameters of the installation area based on the three-dimensional model of the installation area to construct a digital twin. It then calls the air conditioning installation data of the same scenario as the digital twin in the historical operation database to generate an initial candidate installation solution set including the location of the indoor unit, the direction and number of the air outlet; a heat source prediction model building module, which establishes a dynamic heat source distribution prediction model based on the three-dimensional model of the installation area and historical heat source data, and outputs the heat source prediction spatial position, heat source prediction intensity and prediction coverage data of the current installation area as heat source spatial distribution data; The installation plan dynamic adjustment module simulates the deployment of environmental perception sensors to collect real-time temperature, humidity and heat source information based on the initial candidate installation plan set and the heat source spatial distribution data, and uses a multi-objective optimization algorithm to simulate and iterate the candidate plans and output an optimized installation plan set; The solution verification and calibration module actually installs the air-conditioning system according to the optimized installation solution set, obtains operating data through the actual installation sensor network, compares the deviation values ​​of the measured operating data with the simulation results, corrects the digital twin parameters and generates an updated installation solution.

10. A central air conditioning installation system based on intelligent perception mechanism according to claim 9, characterized in that: The environmental perception sensors include temperature and humidity sensors, infrared thermal imagers, light intensity sensors, wind volume sensors and equipment power sensors.

Images