WO2023004704A1 - Method and apparatus for monitoring closed space environment, and computer-readable storage medium - Google Patents

Method and apparatus for monitoring closed space environment, and computer-readable storage medium Download PDF

Info

Publication number
WO2023004704A1
WO2023004704A1 PCT/CN2021/109345 CN2021109345W WO2023004704A1 WO 2023004704 A1 WO2023004704 A1 WO 2023004704A1 CN 2021109345 W CN2021109345 W CN 2021109345W WO 2023004704 A1 WO2023004704 A1 WO 2023004704A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
current
historical
influencing factor
space environment
Prior art date
Application number
PCT/CN2021/109345
Other languages
French (fr)
Chinese (zh)
Inventor
孙天瑞
白新
周晓舟
曹佃松
Original Assignee
西门子（中国）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 西门子（中国）有限公司 filed Critical 西门子（中国）有限公司
Priority to CN202180098829.8A priority Critical patent/CN117413273A/en
Priority to PCT/CN2021/109345 priority patent/WO2023004704A1/en
Publication of WO2023004704A1 publication Critical patent/WO2023004704A1/en

Links

Images

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F11/00—Control or safety arrangements
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]

Definitions

  • the present disclosure relates to the field of environmental monitoring, and more particularly, to a method, apparatus, computing device, computer-readable storage medium, and computer program product for monitoring an enclosed space environment.
  • Physical sensors are often used to monitor critical environmental parameters in enclosed spaces. Since the monitoring of environmental conditions using physical sensors is limited to certain locations, due to cost and physical constraints, only local environmental conditions can be presented but not the overall environmental conditions. As a result, monitoring systems may not be able to address locations within a space where environmental requirements are not met (e.g., hot/cold spots). Virtual sensors based on sensor inputs and system models have been applied to monitor the status of equipment (eg, heating ventilation and air conditioning (HVAC) equipment, motors, etc.). The application of virtual sensors in environmental monitoring is still limited due to the extensive and expensive computational fluid dynamics (CFD) analysis required to build a system model and the challenges of virtual calibration.
  • CFD computational fluid dynamics
  • the CFD simulation calculation time for establishing a simulation-based model is long (for example, CFD modeling analysis may take up to several hours), which cannot meet the needs of real-time environmental monitoring/control, and model calibration requiring a large amount of computing resources is of great importance for establishing high Accurate simulation-driven models are necessary.
  • a first embodiment of the present disclosure proposes a method for monitoring an enclosed space environment, the method comprising: obtaining current sensing data and current influencing factor data associated with the enclosed space environment; obtaining a proxy for the enclosed space environment model, a proxy model is generated based on historical sensing data and historical influencing factor data associated with the enclosed space environment and a CFD model; based on the current sensing data and current influencing factor data, the proxy model is used to determine the enclosed space environment The current overall environment state value for .
  • model calibration using a large number of computing resources can be avoided, and modeling and analysis time can be saved, and the overall environmental state of the closed space environment can be presented quickly and with high precision, so as to Meet real-time monitoring needs.
  • a second embodiment of the present disclosure proposes an apparatus for monitoring an enclosed space environment, the apparatus comprising: an acquisition unit configured to acquire current sensing data and current influencing factor data associated with an enclosed space environment; a model unit , configured to obtain a proxy model for the closed space environment, the proxy model is generated based on historical sensing data and historical influencing factor data associated with the closed space environment and a CFD model; the determining unit is configured to be based on the current sensing The measured data and the current influencing factor data are used to determine the current overall environmental state value of the enclosed space environment by using a proxy model.
  • a third embodiment of the present disclosure proposes a computing device, which includes: a processor; and a memory for storing computer-executable instructions that, when executed, cause the processor to perform the first embodiment. Methods.
  • a fourth embodiment of the present disclosure proposes a computer-readable storage medium having computer-executable instructions stored thereon, and the computer-executable instructions are used to execute the method of the first embodiment.
  • a fifth embodiment of the present disclosure proposes a computer program product that is tangibly stored on a computer-readable storage medium and includes computer-executable instructions that, when executed, cause at least one processing The device executes the method of the first embodiment.
  • FIG. 1 illustrates an exemplary enclosed space environment in which embodiments of the present disclosure may be applied
  • Figure 2 illustrates a flowchart of an exemplary method for monitoring an enclosed space environment according to an embodiment of the present disclosure
  • FIG. 3 shows a flowchart of an exemplary method for generating a proxy model according to an embodiment of the present disclosure
  • FIG. 4 shows a flowchart of another exemplary method for generating a proxy model according to an embodiment of the present disclosure
  • FIG. 5 illustrates an exemplary apparatus for monitoring an enclosed space environment according to an embodiment of the present disclosure
  • FIG. 6 illustrates an exemplary computing device for monitoring an enclosed space environment according to an embodiment of the disclosure.
  • FIG. 1 illustrates an exemplary enclosed space environment 100 in which embodiments of the present disclosure may be applied.
  • the enclosed space environment 100 may be, for example, a data center room, a greenhouse, a cold storage, a clean room, an operating room, and the like.
  • Several objects may be distributed within the enclosed space environment 100 , for example, devices 101 and 102 , and sensor 103 .
  • the equipment 101 may be, for example, racks, and the rows of the racks are stacked such as servers, switches, etc.
  • the equipment 102 may be, for example, air conditioner units (AHUs)
  • sensors 103 It may be, for example, sensors for temperature, humidity, flow rate, etc., which are set on or near the device 101 or 102 or other facilities or other positions that can sense environmental conditions.
  • the environmental state of the enclosed space environment 100 may be sensed by the sensors 103 to generate or record sensed data associated with the enclosed space environment 100, such as real-time sensed data may be referred to as current sensed data, and past sensed data may be referred to as As historical sensing data, the sensing data may be stored for access, for example, in a database or other storage device.
  • Influencing factor data associated with the enclosed space environment 100 may include internal factors (e.g., rack loads in a data center room), external factors (e.g., weather information for a greenhouse), control parameters (e.g., HVAC set points for an office building) and other factors that have a significant impact on environmental KPIs.
  • real-time influencing factor data may be referred to as current influencing factor data
  • past influencing factor data may be referred to as historical sensing data
  • influencing factor data may be stored in a database or other storage devices for access, for example.
  • FIG. 2 illustrates a flowchart of an exemplary method 200 for monitoring an enclosed space environment, according to disclosed embodiments.
  • the exemplary method 200 may be applied to an enclosed space environment 100 as shown in FIG. 1 .
  • step 201 current sensing data and current influencing factor data associated with the enclosed space environment 100 are acquired.
  • sensing data and internal factor data may be collected from sensors or devices within the environment 100
  • external factor data eg, weather information, etc.
  • step 202 a surrogate model for the enclosed space environment 100 is obtained, the surrogate model is generated based on historical sensing data and historical influencing factor data associated with the enclosed space environment 100 and a CFD model.
  • the CFD model is already calibrated with a reference database, so you can avoid time-consuming CFD modeling and use the calibrated CFD model to quickly generate proxy models.
  • the generated proxy model can be stored in a model database or other storage device such that the generated proxy model can be accessed from the model database or other storage device.
  • step 203 based on the current sensing data and current influencing factor data, a proxy model is used to determine the current overall environmental state value (eg, temperature field, airflow field, etc.) of the enclosed space environment 100 .
  • a proxy model is used to determine the current overall environmental state value (eg, temperature field, airflow field, etc.) of the enclosed space environment 100 .
  • the enclosed space environment 100 may be divided into a plurality of grid points, and step 203 may include: based on current sensing data and current influencing factor data, using a proxy model to determine each of the plurality of grid points Environmental parameters of grid points; based on the environmental parameters of each grid point, determine the current overall environmental state value of the enclosed space environment.
  • the environmental state values of the continuous spatial positions of the environment 100 can be predicted based on the environmental parameters determined by the discrete grid points. For example, based on the environmental parameters of multiple grid points near any spatial position of the environment 100, the environmental state value of the spatial position may be obtained by interpolation.
  • step 203 may further include: performing data cleaning and preprocessing on the current sensing data and current influencing factor data; based on the cleaned and preprocessed current sensing data and current influencing factor data, using a proxy model to An environmental parameter is determined for each grid point of the plurality of grid points.
  • data accuracy, completeness, and consistency can be improved through data cleaning and preprocessing, so as to avoid affecting the accuracy of the generated environmental state values.
  • data cleaning and preprocessing can be performed on historical sensing data and historical influencing factor data to improve data accuracy, completeness, and consistency, and avoid affecting the accuracy of the generated proxy model.
  • the proxy model is generated by performing a CFD simulation using a CFD model to construct a simulation of the environmental state of the enclosed space environment based on historical sensing data and historical influencing factor data and a plurality of grid points Result data set; based on the historical sensing data and historical influencing factor data and the simulation result data set, the agent model is trained and generated.
  • historical sensing data and historical influencing factor data are input into the CFD model to perform CFD simulation to generate a simulation result data set (on multiple grid points) about the environmental state of the closed space environment; and the historical sensing data and The historical influencing factor data is used as input, the simulation result data set (on multiple grid points) is used as output, and a part of the data set composed of input and corresponding output is used as a training set and a part is used as a test set; based on the training set, through Use algorithms (eg, kriging, KPLS, multi-precision kriging, etc.) to train the generation proxy model, and based on the test set, to verify the proxy model.
  • Use algorithms eg, kriging, KPLS, multi-precision kriging, etc.
  • FIG. 3 a flowchart of one exemplary method 300 for generating a proxy model is shown, according to an embodiment of the present disclosure.
  • step 301 historical sensing data and historical influencing factor data are divided into multiple subgroups according to a set of division strategies.
  • step 302 for each of the plurality of subgroups: using a CFD model to construct a corresponding subset of simulation result data.
  • a proxy model is generated based on a plurality of sub-agent models corresponding to a plurality of sub-groups.
  • the parameters of the subgroups can be taken as input and a subset of simulation result data can be used as output, and the data set consisting of the input and the corresponding output can be divided into a training set and a test set.
  • the sub-agent model is trained by using an algorithm (for example, kriging, KPLS, multi-precision kriging, etc.), and based on the test set, the sub-agent model is verified.
  • FIG. 4 there is shown a flowchart of another exemplary method 400 for generating a proxy model according to an embodiment of the present disclosure.
  • step 401 the division strategy set includes a plurality of strategies, and each strategy in the plurality of strategies specifies different divisions from historical sensing data and historical influencing factor data into a plurality of subgroups, and method 300 may be used for each strategy to Proxy models are generated from multiple subgroups corresponding to each strategy.
  • step 402 based on historical sensing data and historical influencing factor data, corresponding overall environmental state values of the enclosed space environment 100 are generated through multiple agent models corresponding to multiple policies.
  • step 403 the generated overall environmental state value for the corresponding enclosed space environment 100 is compared with historical environmental state values associated with the enclosed space environment 100 .
  • the historical environmental state value of the environment 100 may be obtained by arranging relevant sensing or detection devices in the environment 100 for sensing or detection within a period of time.
  • step 404 one of the plurality of proxy models corresponding to the plurality of policies is determined as the proxy model for the enclosed space environment 100 according to the comparison result. For example, a proxy model corresponding to an overall environmental state value with the best accuracy or minimum error (e.g., mean square error, etc.) with historical environmental state values may be set as the proxy model for the enclosed space environment 100 to achieve the best predictive effect.
  • the best accuracy or minimum error e.g., mean square error, etc.
  • the proxy model with the best prediction effect can be selected from the proxy models generated based on different partitioning strategies, so as to further improve the accuracy of the proxy model.
  • the data on the influencing factors of the room temperature state include weather information, the load of each rack in the 4 rows, and the control parameters of the 3 air conditioner equipment (AHU), and the sensing data includes The temperature in the middle of each row and the air volume of each AHU.
  • An exemplary set of partition policies and corresponding exemplary subgroup partitions are shown in Table 1 below.
  • the partition policy set may include policy 1, policy 2, policy 3 and policy 4.
  • These strategies 1-4 can, for example, divide inputs into categories (eg, internal factors, external factors, and control parameters), divide inputs into detailed categories (eg, divide heat sources into categories by location), combine sensor information with model inputs Combine inputs into one category (for example combine heat sources and temperature sensors within a location into one category).
  • the sensing data and influencing factor data can be divided into adjustable control parameters and non-adjustable model inputs, and the control parameters and model inputs can be divided into multiple subgroups.
  • strategy 1 divides control parameters and model inputs into 3 subgroups 1.1-1.3
  • strategy 2 divides control parameters and model inputs into 5 subgroups 2.1-2.5
  • strategy 3 divides control parameters and model inputs into 8 subgroups 3.1 ⁇ 3.8
  • Strategy 4 divides control parameters and model inputs into 8 subgroups 4.1 ⁇ 4.8.
  • strategy 1-4 only a part of the control parameters and model inputs can be divided into multiple subgroups, so that some control parameters and model inputs are not included in any subgroup.
  • Table 1 Partition strategy set and subgroup partition
  • a corresponding sub-agent model E i can be obtained, representing the relative influence of the parameters in the subgroup i on the environment state of a certain grid point.
  • the influence of all parameters on the environmental state of a certain grid point can be integrated through proxy models generated for each subgroup. For example, if linear superposition is applied during integration, the integrated surrogate model E all can be expressed as follows:
  • weighting may be applied to the sub-agent models to generate an integrated proxy model E all for the environment 100 to obtain an integrated proxy model E all with the best accuracy.
  • the method 300 can be used to obtain multiple proxy models E all (j) corresponding to multiple partition strategies, and determine the proxy model with the best predictive effect from these multiple proxy models, for example, by using the proxy
  • the output of the model is compared with the actual obtained historical environmental state values.
  • the plurality of grid points comprises a set of regions for which a proxy model for the enclosed space environment 100 is determined.
  • all grid points can be regarded as a region, so that all grid points use the same proxy model with the best prediction effect, such as the proxy model generated according to the method 300 or 400 .
  • a surrogate model with the best predictive performance can be applied at each grid point.
  • the grid points can be divided into multiple regions (which can be divided according to physical location or other methods), and the proxy model with the best predictive effect can be applied to each region.
  • the method 200 may also optionally include steps 204 - 207 .
  • the method 200 may include: acquiring a plurality of proxy models corresponding to a plurality of strategies; based on current sensing data and current influencing factor data, using each of the plurality of proxy models to determine the a corresponding set of current overall environment state values; based on the plurality of sets of current overall environment state values corresponding to the plurality of agent models, anomalies of objects within the enclosed space environment are detected by cross-validation, the objects being consistent with the current sensed data Associated with one of the current influencer data. For example, multiple proxy models may be accessed from model data or other storage.
  • each strategy specifies the different divisions of historical sensing data and historical influencing factor data (or control parameters and model inputs) into multiple subgroups, that is to say, control parameters and model inputs can be made under different strategies are classified into different subgroups, or under certain policies are not included in any subgroup. Therefore, this step 207 can cross-validate to compare the influence of the presence or absence of control parameters and model inputs on the overall environment state value, thereby detecting whether the objects (such as sensors or equipment, etc.) associated with control parameters and model inputs are abnormal .
  • the method 200 may include: setting the current overall environmental state value or predicted overall environmental state value (for example, a future expected overall environmental state value) as an initial state; based on the initial state, using an optimization algorithm (for example, Differential evolution (DE), particle swarm optimization (PSO), simulated annealing, ant colony algorithm, etc.) perform the following steps:
  • DE Differential evolution
  • PSO particle swarm optimization
  • simulated annealing for example, a future expected overall environmental state value
  • ant colony algorithm etc.
  • the adjustment step is to adjust the adjustable parameters in the current sensing data and the current influencing factor data
  • the judging step is to determine whether the updated environmental state value of the enclosed space environment reaches the optimization target, if not, go to the adjustment step, and if the optimization target has been reached, provide the optimal value of the adjustable parameter.
  • the optimization goal is to obtain optimized energy efficiency while keeping key environmental KPIs within the design range
  • the adjustable parameters are control parameters such as AHU (e.g., temperature set point, wind speed set point, etc.), external influencing factors
  • the data may, for example, include a stepwise electricity price distribution interval and the like.
  • the existing CFD modeling analysis cannot realize this kind of optimal control.
  • parameters can be adjusted, input the adjusted whole parameters to the proxy model to obtain model output, and compare whether the optimization goal is achieved, and use an optimization algorithm to quickly and iteratively realize optimal control.
  • the optimized value of the adjusted parameter can be provided to the management unit, that is, an optimized control strategy can be provided to the management unit to adjust the control parameter.
  • the method 200 may include: detecting an environmental anomaly of the enclosed space environment based on the current overall environmental state value; and generating an alarm in response to detecting the environmental anomaly. For example, conditions such as localized overheating or overcooling, abnormal wind speed, etc. can be detected, and based on the detected abnormal conditions, alarms are generated to alert eg engineers or operators.
  • the current overall environmental state value can also be processed into a data format that can be displayed in a visual form, so that the current overall environmental state value can be presented graphically on a user interface or a display device.
  • the method 200 may include: updating historical sensing data and historical influencing factor data with current sensing data and current influencing factor data; and updating based on the updated historical sensing data and historical influencing factor data and the CFD model The proxy model.
  • the historical data can be updated over time, and the proxy model can be regenerated using the updated historical data, so as to improve the accuracy of the proxy model.
  • the aforementioned method 200 it has the following advantages: by reading the input proxy model from multiple input sources, it is possible to perform high-precision real-time three-dimensional overall environmental monitoring of the enclosed space, therefore, it is possible to detect and report the environment requirements in the entire enclosed space.
  • the deviation of the proxy model can be used to provide dynamic control strategy recommendations for reliable overall environmental state assurance and/or energy consumption optimization; the parameter division strategy is used to significantly reduce the computing resources for constructing proxy models, speed up system deployment, and reduce simulation time.
  • Cost; sensor/equipment anomalies can be detected and diagnosed through cross-validation of calculation results of multiple integrated proxy models, which can increase the stability and reliability of environmental control; can be integrated into external management systems (e.g., asset management systems, etc.) , for additional functions such as capacity planning for data centers and demand/sequence design for chilled water systems.
  • external management systems e.g., asset management systems, etc.
  • FIG. 5 illustrates an exemplary apparatus 500 for monitoring an enclosed space environment according to an embodiment of the disclosure.
  • an apparatus 500 includes an acquisition unit 501 , a model unit 502 and a determination unit 503 .
  • the obtaining unit 501 is configured to perform the process described above in relation to step 201 of the method 200
  • the model unit 502 is configured to perform the process described above in relation to step 202 in the method 200
  • the determining unit 503 is configured to perform the process described above in relation to The process described in step 203 in the method 200, wherein the proxy model is generated according to the process described in relation to the method 200 above.
  • the apparatus 500 may also optionally include a detection unit 504 , an optimization unit 505 , an exception processing unit 506 and an update unit 507 .
  • the detection unit 504 is configured to execute the process as described above about step 204 in the method 200
  • the optimization unit 505 is configured to execute the process described above about step 205 in the method 200
  • the exception processing unit 506 is configured to execute the process described above about step 205 in the method 200.
  • the updating unit 507 is configured to perform the process as described above in relation to step 207 in method 200 .
  • the apparatus 500 may also optionally include a model generation unit configured to perform the process of generating a proxy model as described above in relation to the method 200 .
  • FIG. 6 illustrates a block diagram of an exemplary computing device 600 for monitoring an enclosed space environment, according to an embodiment of the disclosure.
  • the computing device 600 includes a processor 601 and a memory 602 coupled with the processor 601 .
  • the memory 602 is used to store computer-executable instructions, and when the computer-executable instructions are executed, the processor 601 executes the methods in the above embodiments (for example, any one or more steps of the foregoing method 200).
  • a computer-readable storage medium carries computer-readable program instructions for implementing various embodiments of the present disclosure.
  • a computer readable storage medium may be a tangible device that can retain and store instructions for use by an instruction execution device.
  • a computer readable storage medium may be, for example, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.
  • Computer-readable storage media include: portable computer diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or flash memory), static random access memory (SRAM), compact disc read only memory (CD-ROM), digital versatile disc (DVD), memory stick, floppy disk, mechanically encoded device, such as a printer with instructions stored thereon A hole card or a raised structure in a groove, and any suitable combination of the above.
  • RAM random access memory
  • ROM read-only memory
  • EPROM erasable programmable read-only memory
  • flash memory static random access memory
  • SRAM static random access memory
  • CD-ROM compact disc read only memory
  • DVD digital versatile disc
  • memory stick floppy disk
  • mechanically encoded device such as a printer with instructions stored thereon
  • a hole card or a raised structure in a groove and any suitable combination of the above.
  • computer-readable storage media are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., pulses of light through fiber optic cables), or transmitted electrical signals.
  • the present disclosure provides a computer-readable storage medium having computer-executable instructions stored thereon for performing various implementations of the present disclosure. method in the example.
  • the present disclosure provides a computer program product tangibly stored on a computer-readable storage medium and comprising computer-executable instructions that, when executed, cause At least one processor executes the methods in various embodiments of the present disclosure.
  • the various example embodiments of the present disclosure may be implemented in hardware or special purpose circuits, software, firmware, logic, or any combination thereof. Certain aspects may be implemented in hardware, while other aspects may be implemented in firmware or software, which may be executed by a controller, microprocessor or other computing device.
  • aspects of the embodiments of the present disclosure are illustrated or described as block diagrams, flowcharts, or using some other graphical representation, it is to be understood that the blocks, devices, systems, techniques, or methods described herein may serve as non-limiting Examples are implemented in hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controllers or other computing devices, or some combination thereof.
  • the computer-readable program instructions or computer program products used to execute various embodiments of the present disclosure can also be stored in the cloud, and when called, the user can access the program stored on the cloud for execution through the mobile Internet, fixed network or other networks.
  • the computer-readable program instructions of an embodiment of the present disclosure implement the technical solutions disclosed in accordance with various embodiments of the present disclosure.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Geometry (AREA)
  • Health & Medical Sciences (AREA)
  • Mathematical Physics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Computer Hardware Design (AREA)
  • Mathematical Analysis (AREA)
  • Fluid Mechanics (AREA)
  • Computing Systems (AREA)
  • Algebra (AREA)
  • Mathematical Optimization (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Automation & Control Theory (AREA)
  • Chemical & Material Sciences (AREA)
  • Combustion & Propulsion (AREA)
  • Mechanical Engineering (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

A method (200) for monitoring a closed space environment. The method comprises: step 201, acquiring the current sensing data and the current influence factor data, which are associated with a closed space environment (100); step 202, acquiring a proxy model for the closed space environment (100), wherein the proxy model is generated on the basis of historical sensing data and historical influence factor data, which are associated with the closed space environment (100), and a CFD model; and step 203, determining, on the basis of the current sensing data and the current influence factor data, the current overall environmental state value of the closed space environment (100) by using the proxy model. An apparatus (500) for monitoring a closed space environment. The apparatus comprises: an acquisition unit (501) for executing step 201, a model unit (502) for executing step 202, and a determination unit (503) for executing step 203. Further provided are a computing device, a computer-readable storage medium and a computer program product.

Description

用于监控封闭空间环境的方法、装置和计算机存储介质Method, apparatus and computer storage medium for monitoring an enclosed space environment 技术领域technical field
本公开涉及环境监控领域，更具体地说，涉及用于监控封闭空间环境的方法、装置、计算设备、计算机可读存储介质和计算机程序产品。The present disclosure relates to the field of environmental monitoring, and more particularly, to a method, apparatus, computing device, computer-readable storage medium, and computer program product for monitoring an enclosed space environment.
背景技术Background technique
环境保障和关键参数(如温度、湿度、清洁度、气流)在内部/外部影响下的动态控制对于诸如数据中心室、温室、冷库、洁净室和操作室等的封闭空间环境的环境敏感行业至关重要。可靠的环境保障和精确控制需要对目标基础设施/设施和涉及多个***以及内部和外部环境的集成解决方案进行整体监控。还存在其他的挑战，例如在将关键环境KPI(关键性能指标)保持在设计范围内的同时获得优化的能源效率。Environmental assurance and dynamic control of key parameters (such as temperature, humidity, cleanliness, airflow) under internal/external influences are essential for environmentally sensitive industries such as data center rooms, greenhouses, cold stores, clean rooms and operating rooms in enclosed space environments important. Reliable environmental assurance and precise control requires holistic monitoring of the target infrastructure/facilities and integrated solutions involving multiple systems and internal and external environments. There are other challenges, such as obtaining optimized energy efficiency while keeping key environmental KPIs (key performance indicators) within design limits.
物理传感器通常用于监测封闭空间中的关键环境参数。由于使用物理传感器对环境状态的监测仅限于某些位置，出于成本和物理约束的限制，只能呈现局部的环境状态而无法呈现整体的环境状态。因此，监视***可能无法解决空间内未满足环境要求的位置(例如，热点/冷点)。基于传感器输入和***模型的虚拟传感器已应用于监测设备(例如，暖通空调(HVAC)设备、电机等)的状态。由于建立***模型需要大量昂贵的计算流体力学(CFD)分析以及虚拟校准的挑战，虚拟传感器在环境监测中的应用仍然受到限制。Physical sensors are often used to monitor critical environmental parameters in enclosed spaces. Since the monitoring of environmental conditions using physical sensors is limited to certain locations, due to cost and physical constraints, only local environmental conditions can be presented but not the overall environmental conditions. As a result, monitoring systems may not be able to address locations within a space where environmental requirements are not met (e.g., hot/cold spots). Virtual sensors based on sensor inputs and system models have been applied to monitor the status of equipment (eg, heating ventilation and air conditioning (HVAC) equipment, motors, etc.). The application of virtual sensors in environmental monitoring is still limited due to the extensive and expensive computational fluid dynamics (CFD) analysis required to build a system model and the challenges of virtual calibration.
虽然基于物理/虚拟传感器的开环/闭环控制***(例如，PID)广泛用于环境管理，但控制参数、内部/外部影响和环境KPI之间的相关性对于可靠的环境保证是必要的。可以使用数据驱动模型和/或模拟驱动模型来建立相关性。数据驱动模型的准确性受到训练数据的数量和质量的限制，即数据收集周期和传感器的准确性。另一方面，建立基于仿真的模型的CFD仿真计算时间长(例如，CFD建模分析可能需要多达数小时)，无法满足实时环境监测/控制的需求，需要大量计算资源的模型校准对于建立高精度仿真驱动模型是必要的。Although physical/virtual sensor-based open-loop/closed-loop control systems (e.g., PID) are widely used in environmental management, correlations among control parameters, internal/external influences, and environmental KPIs are necessary for reliable environmental assurance. Correlations can be established using data-driven models and/or simulation-driven models. The accuracy of data-driven models is limited by the quantity and quality of training data, i.e. data collection cycle and sensor accuracy. On the other hand, the CFD simulation calculation time for establishing a simulation-based model is long (for example, CFD modeling analysis may take up to several hours), which cannot meet the needs of real-time environmental monitoring/control, and model calibration requiring a large amount of computing resources is of great importance for establishing high Accurate simulation-driven models are necessary.
因此，亟需一种改进的用于监控封闭空间环境的解决方案。Therefore, there is a need for an improved solution for monitoring an enclosed space environment.
发明内容Contents of the invention
采用物理传感器来监测封闭空间环境通常只能呈现局部的环境状态，但无法呈现整体的环境状态，而通过CFD建模分析耗费了大量的时间进行仿真计算，难以满足实时环境监测/控制的需求。可靠的环境保障和精确控制需要对目标基础设施/设施和涉及多个***以及内部和外部环境的集成解决方案进行实时的整体监控。The use of physical sensors to monitor the closed space environment can usually only present the local environmental state, but cannot present the overall environmental state. CFD modeling analysis consumes a lot of time for simulation calculations, and it is difficult to meet the needs of real-time environmental monitoring/control. Reliable environmental assurance and precise control requires real-time holistic monitoring of target infrastructure/facilities and integrated solutions involving multiple systems and internal and external environments.
本公开的第一实施例提出了一种用于监控封闭空间环境的方法，该方法包括：获取与封闭空间环境相关联的当前感测数据和当前影响因素数据；获取用于封闭空间环境的代理模型，代理模型是基于与封闭空间环境相关联的历史感测数据和历史影响因素数据以及CFD模型来生成的；基于当前感测数据以及当前影响因素数据，使用所述代理模型来确定封闭空间环境的当前整体环境状态值。A first embodiment of the present disclosure proposes a method for monitoring an enclosed space environment, the method comprising: obtaining current sensing data and current influencing factor data associated with the enclosed space environment; obtaining a proxy for the enclosed space environment model, a proxy model is generated based on historical sensing data and historical influencing factor data associated with the enclosed space environment and a CFD model; based on the current sensing data and current influencing factor data, the proxy model is used to determine the enclosed space environment The current overall environment state value for .
在该实施例中，通过使用代理模型取代CFD建模，可以避免使用大量计算资源的模型校准，并节省了建模分析时间，可快速地且高精度地呈现封闭空间环境的整体环境状态，以满足实时监控需求。In this embodiment, by using a proxy model instead of CFD modeling, model calibration using a large number of computing resources can be avoided, and modeling and analysis time can be saved, and the overall environmental state of the closed space environment can be presented quickly and with high precision, so as to Meet real-time monitoring needs.
本公开的第二实施例提出了一种用于监控封闭空间环境的装置，该装置包括：获取单元，被配置为获取与封闭空间环境相关联的当前感测数据和当前影响因素数据；模型单元，被配置为获取用于封闭空间环境的代理模型，代理模型是基于与封闭空间环境相关联的历史感测数据和历史影响因素数据以及CFD模型来生成的；确定单元，被配置为基于当前感测数据以及当前影响因素数据，使用代理模型来确定所述封闭空间环境的当前整体环境状态值。A second embodiment of the present disclosure proposes an apparatus for monitoring an enclosed space environment, the apparatus comprising: an acquisition unit configured to acquire current sensing data and current influencing factor data associated with an enclosed space environment; a model unit , configured to obtain a proxy model for the closed space environment, the proxy model is generated based on historical sensing data and historical influencing factor data associated with the closed space environment and a CFD model; the determining unit is configured to be based on the current sensing The measured data and the current influencing factor data are used to determine the current overall environmental state value of the enclosed space environment by using a proxy model.
本公开的第三实施例提出了一种计算设备，该计算设备包括：处理器；以及存储器，其用于存储计算机可执行指令，当计算机可执行指令被执行时使得处理器执行第一实施例的方法。A third embodiment of the present disclosure proposes a computing device, which includes: a processor; and a memory for storing computer-executable instructions that, when executed, cause the processor to perform the first embodiment. Methods.
本公开的第四实施例提出了一种计算机可读存储介质，该计算机可读存储介质具有存储在其上的计算机可执行指令，计算机可执行指令用于执行第一实施例的方法。A fourth embodiment of the present disclosure proposes a computer-readable storage medium having computer-executable instructions stored thereon, and the computer-executable instructions are used to execute the method of the first embodiment.
本公开的第五实施例提出了一种计算机程序产品，该计算机程序产品被有形地存储在计算机可读存储介质上，并且包括计算机可执行指令，计算机 可执行指令在被执行时使至少一个处理器执行第一实施例的方法。A fifth embodiment of the present disclosure proposes a computer program product that is tangibly stored on a computer-readable storage medium and includes computer-executable instructions that, when executed, cause at least one processing The device executes the method of the first embodiment.
附图说明Description of drawings
结合附图并参考以下详细说明，本公开的各实施例的特征、优点及其他方面将变得更加明显，在此以示例性而非限制性的方式示出了本公开的若干实施例，在附图中：The features, advantages and other aspects of the various embodiments of the present disclosure will become more apparent with reference to the following detailed description when taken in conjunction with the accompanying drawings, which show several embodiments of the present disclosure by way of illustration and not limitation. In the attached picture:
图1示出了其中可应用本公开的实施例的示例性封闭空间环境；FIG. 1 illustrates an exemplary enclosed space environment in which embodiments of the present disclosure may be applied;
图2示出根据本公开的实施例的用于监控封闭空间环境的示例性方法的流程图；Figure 2 illustrates a flowchart of an exemplary method for monitoring an enclosed space environment according to an embodiment of the present disclosure;
图3示出了根据本公开的实施例的用于生成代理模型的一个示例性方法的流程图；FIG. 3 shows a flowchart of an exemplary method for generating a proxy model according to an embodiment of the present disclosure;
图4示出了根据本公开的实施例的用于生成代理模型的另一个示例性方法的流程图；FIG. 4 shows a flowchart of another exemplary method for generating a proxy model according to an embodiment of the present disclosure;
图5示出了根据本公开的实施例的用于监控封闭空间环境的示例性装置；FIG. 5 illustrates an exemplary apparatus for monitoring an enclosed space environment according to an embodiment of the present disclosure;
图6示出了根据本公开的实施例的用于监控封闭空间环境的示例性计算设备。FIG. 6 illustrates an exemplary computing device for monitoring an enclosed space environment according to an embodiment of the disclosure.
具体实施方式Detailed ways
以下参考附图详细描述本公开的各个示例性实施例。虽然以下所描述的示例性方法、装置包括在其它组件当中的硬件上执行的软件和/或固件，但是应当注意，这些示例仅仅是说明性的，而不应看作是限制性的。例如，考虑在硬件中独占地、在软件中独占地、或在硬件和软件的任何组合中可以实施任何或所有硬件、软件和固件组件。因此，虽然以下已经描述了示例性的方法和装置，但是本领域的技术人员应容易理解，所提供的示例并不用于限制用于实现这些方法和装置的方式。Various exemplary embodiments of the present disclosure are described in detail below with reference to the accompanying drawings. While the example methods, apparatus described below include software and/or firmware executing on hardware, among other components, it should be noted that these examples are illustrative only and should not be viewed as limiting. For example, it is contemplated that any or all hardware, software, and firmware components may be implemented exclusively in hardware, exclusively in software, or in any combination of hardware and software. Therefore, although exemplary methods and apparatuses have been described below, those skilled in the art will readily understand that the examples provided are not intended to limit the manner in which these methods and apparatuses are implemented.
此外，附图中的流程图和框图示出了根据本公开的各个实施例的方法和***的可能实现的体系架构、功能和操作。应当注意，方框中所标注的功能也可以按照不同于附图中所标注的顺序发生。例如，两个接连地表示的方框实际上可以基本并行地执行，或者它们有时也可以按照相反的顺序执行，这取决于所涉及的功能。同样应当注意的是，流程图和/或框图中的每个方框、以及流程图和/或框图中的方框的组合，可以使用执行规定的功能或操作的专 用的基于硬件的***来实现，或者可以使用专用硬件与计算机指令的组合来实现。Furthermore, the flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of methods and systems according to various embodiments of the present disclosure. It should be noted that the functions noted in the block may also occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved. It should also be noted that each block in the flowchart and/or block diagrams, and combinations of blocks in the flowchart and/or block diagrams, can be implemented using a dedicated hardware-based system that performs the specified functions or operations , or can be implemented using a combination of dedicated hardware and computer instructions.
本文所使用的术语“包括”、“包含”及类似术语是开放性的术语，即“包括/包含但不限于”，表示还可以包括其他内容。术语“基于”是“至少部分地基于”。术语“一个实施例”表示“至少一个实施例”；术语“另一实施例”表示“至少一个另外的实施例”等等。The terms "including", "comprising" and similar terms used herein are open-ended terms, that is, "including/including but not limited to", which means that other contents may also be included. The term "based on" is "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one further embodiment" and so on.
图1示出了其中可以应用本公开的实施例的示例性封闭空间环境100。封闭空间环境100可以是例如数据中心室、温室、冷库、洁净室和操作室等。封闭空间环境100内可以分布有若干对象，举例来说，设备101和102、以及传感器103。在封闭空间环境100是数据中心室的示例中，设备101可以例如是机架，并且机架的各排上堆放了诸如服务器、交换机等，设备102可以是例如空调器设备(AHU)，传感器103可以例如是温度、湿度、流速等传感器，设置在设备101或设备102或其他设施上或附近或其他可以感测环境状态的位置。可以由传感器103感测封闭空间环境100的环境状态以生成或记录与封闭空间环境100相关联的感测数据，例如实时的感测数据可以称为当前感测数据，过去的感测数据可以称为历史感测数据，感测数据可以例如保存在数据库或其他存储装置中以供存取。与封闭空间环境100相关联的影响因素数据可以包括内部因素(例如，数据中心室的机架负载)、外部因素(例如，温室的天气信息)、控制参数(例如，办公楼的HVAC设置点)以及其他对环境KPI有重大影响的因素。类似地，例如实时的影响因素数据可以称为当前影响因素数据，过去的影响因素数据可以称为历史感测数据，影响因素数据可以例如保存在数据库或其他存储装置中以供存取。FIG. 1 illustrates an exemplary enclosed space environment 100 in which embodiments of the present disclosure may be applied. The enclosed space environment 100 may be, for example, a data center room, a greenhouse, a cold storage, a clean room, an operating room, and the like. Several objects may be distributed within the enclosed space environment 100 , for example, devices 101 and 102 , and sensor 103 . In an example where the enclosed space environment 100 is a data center room, the equipment 101 may be, for example, racks, and the rows of the racks are stacked such as servers, switches, etc., the equipment 102 may be, for example, air conditioner units (AHUs), sensors 103 It may be, for example, sensors for temperature, humidity, flow rate, etc., which are set on or near the device 101 or 102 or other facilities or other positions that can sense environmental conditions. The environmental state of the enclosed space environment 100 may be sensed by the sensors 103 to generate or record sensed data associated with the enclosed space environment 100, such as real-time sensed data may be referred to as current sensed data, and past sensed data may be referred to as As historical sensing data, the sensing data may be stored for access, for example, in a database or other storage device. Influencing factor data associated with the enclosed space environment 100 may include internal factors (e.g., rack loads in a data center room), external factors (e.g., weather information for a greenhouse), control parameters (e.g., HVAC set points for an office building) and other factors that have a significant impact on environmental KPIs. Similarly, real-time influencing factor data may be referred to as current influencing factor data, past influencing factor data may be referred to as historical sensing data, and influencing factor data may be stored in a database or other storage devices for access, for example.
如前所述，要呈现封闭空间环境100的整体环境状态，需要在环境100内布置大量的物理传感器103，然而出于成本和物理约束的限制，这是不切实际的。虽然基于传感器输入和***模型的虚拟传感器已应用于监测设备，但是建立***模型需要大量昂贵的CFD分析以及虚拟校准的挑战，使得虚拟传感器在环境监测中的应用仍然受到限制。此外，建立基于仿真的模型的CFD仿真计算时间长(例如，CFD建模分析可能需要多达数小时)，无法满足实时环境监测/控制的需求。As previously mentioned, to represent the overall environmental state of the enclosed space environment 100, a large number of physical sensors 103 would need to be placed within the environment 100, however this is impractical due to cost and physical constraints. Although virtual sensors based on sensor input and system models have been applied to monitoring equipment, the establishment of system models requires a lot of expensive CFD analysis and the challenge of virtual calibration, so that the application of virtual sensors in environmental monitoring is still limited. In addition, the calculation time of CFD simulation to establish a simulation-based model is long (for example, CFD modeling analysis may take up to several hours), which cannot meet the needs of real-time environmental monitoring/control.
图2示出了根据公开的实施例的用于监控封闭空间环境的示例性方法200的流程图。该示例性方法200可以应用于如图1所示的封闭空间环境100。FIG. 2 illustrates a flowchart of an exemplary method 200 for monitoring an enclosed space environment, according to disclosed embodiments. The exemplary method 200 may be applied to an enclosed space environment 100 as shown in FIG. 1 .
参考图2，方法200从步骤201开始。在步骤201中，获取与封闭空间环境100相关联的当前感测数据和当前影响因素数据。例如，可以从环境100内传感器或设备采集感测数据和内部因素数据，以及从环境100外部的气象服务中心接收外部因素数据(例如，天气信息等)。Referring to FIG. 2 , method 200 starts at step 201 . In step 201, current sensing data and current influencing factor data associated with the enclosed space environment 100 are acquired. For example, sensing data and internal factor data may be collected from sensors or devices within the environment 100 , and external factor data (eg, weather information, etc.) may be received from a weather service center outside the environment 100 .
接着，方法200行进到步骤202。在步骤202中，获取用于封闭空间环境100的代理模型(surrogate model)，代理模型是基于与封闭空间环境100相关联的历史感测数据和历史影响因素数据以及CFD模型来生成的。例如，CFD模型是已通过参考数据库进行校准的，因此可以避免耗费大量的时间进行CFD建模，并利用经校准的CFD模型来快速生成代理模型。例如，所生成的代理模型可以存储在模型数据库或其他存储装置中，使得可以从模型数据库或其他存储装置访问所生成的代理模型。Next, the method 200 proceeds to step 202 . In step 202, a surrogate model for the enclosed space environment 100 is obtained, the surrogate model is generated based on historical sensing data and historical influencing factor data associated with the enclosed space environment 100 and a CFD model. For example, the CFD model is already calibrated with a reference database, so you can avoid time-consuming CFD modeling and use the calibrated CFD model to quickly generate proxy models. For example, the generated proxy model can be stored in a model database or other storage device such that the generated proxy model can be accessed from the model database or other storage device.
接着，方法200行进到步骤203。在步骤203中，基于当前感测数据以及当前影响因素数据，使用代理模型来确定封闭空间环境100的当前整体环境状态值(例如，温度场、气流场等)。Next, the method 200 proceeds to step 203 . In step 203 , based on the current sensing data and current influencing factor data, a proxy model is used to determine the current overall environmental state value (eg, temperature field, airflow field, etc.) of the enclosed space environment 100 .
在一些实施例中，封闭空间环境100可以被划分为多个网格点，并且步骤203可以包括：基于当前感测数据以及当前影响因素数据，使用代理模型来确定多个网格点中的每个网格点的环境参数；基于每个网格点的环境参数，确定所述封闭空间环境的当前整体环境状态值。在该步骤中，可以通过将空间环境100分成离散的网格点，基于离散网格点确定的环境参数来预测环境100的连续空间位置的环境状态值。例如，基于环境100的任何空间位置附近的多个网格点的环境参数，可以通过插值法来获得该空间位置的环境状态值。In some embodiments, the enclosed space environment 100 may be divided into a plurality of grid points, and step 203 may include: based on current sensing data and current influencing factor data, using a proxy model to determine each of the plurality of grid points Environmental parameters of grid points; based on the environmental parameters of each grid point, determine the current overall environmental state value of the enclosed space environment. In this step, by dividing the space environment 100 into discrete grid points, the environmental state values of the continuous spatial positions of the environment 100 can be predicted based on the environmental parameters determined by the discrete grid points. For example, based on the environmental parameters of multiple grid points near any spatial position of the environment 100, the environmental state value of the spatial position may be obtained by interpolation.
在一些实施例中，步骤203可以进一步包括：对当前感测数据以及当前影响因素数据进行数据清洗和预处理；基于清洗和预处理后的当前感测数据以及当前影响因素数据，使用代理模型来确定多个网格点中的每个网格点的环境参数。在该步骤中，通过数据清洗和预处理可以提升数据准确性、完整性、一致性，以避免影响所生成的环境状态值的准确性。类似地，在生成代理模型之前，可以对历史感测数据以及历史影响因素数据进行数据清洗和预处理，以提升数据准确性、完整性、一致性，避免影响所生成的代理模型的准确性。In some embodiments, step 203 may further include: performing data cleaning and preprocessing on the current sensing data and current influencing factor data; based on the cleaned and preprocessed current sensing data and current influencing factor data, using a proxy model to An environmental parameter is determined for each grid point of the plurality of grid points. In this step, data accuracy, completeness, and consistency can be improved through data cleaning and preprocessing, so as to avoid affecting the accuracy of the generated environmental state values. Similarly, before generating the proxy model, data cleaning and preprocessing can be performed on historical sensing data and historical influencing factor data to improve data accuracy, completeness, and consistency, and avoid affecting the accuracy of the generated proxy model.
在一些实施例中，代理模型是通过以下方式来生成的：基于历史感测数 据和历史影响因素数据以及多个网格点，使用CFD模型进行CFD仿真以构建关于封闭空间环境的环境状态的仿真结果数据集；基于历史感测数据和历史影响因素数据以及仿真结果数据集，训练生成代理模型。例如，将历史感测数据和历史影响因素数据输入到CFD模型进行CFD仿真以生成关于封闭空间环境的环境状态的(多个网格点上的)仿真结果数据集；并且将历史感测数据和历史影响因素数据作为输入，(多个网格点上的)仿真结果数据集作为输出，并将由输入和相应输出组成的数据集合中的一部分作为训练集以及一部分作为测试集；基于训练集，通过使用算法(例如，kriging、KPLS、多精度kriging等)来训练生成代理模型，并基于测试集，来对代理模型进行验证。In some embodiments, the proxy model is generated by performing a CFD simulation using a CFD model to construct a simulation of the environmental state of the enclosed space environment based on historical sensing data and historical influencing factor data and a plurality of grid points Result data set; based on the historical sensing data and historical influencing factor data and the simulation result data set, the agent model is trained and generated. For example, historical sensing data and historical influencing factor data are input into the CFD model to perform CFD simulation to generate a simulation result data set (on multiple grid points) about the environmental state of the closed space environment; and the historical sensing data and The historical influencing factor data is used as input, the simulation result data set (on multiple grid points) is used as output, and a part of the data set composed of input and corresponding output is used as a training set and a part is used as a test set; based on the training set, through Use algorithms (eg, kriging, KPLS, multi-precision kriging, etc.) to train the generation proxy model, and based on the test set, to verify the proxy model.
转到图3，示出根据本公开的实施例的用于生成代理模型的一个示例性方法300的流程图。Turning to FIG. 3 , a flowchart of one exemplary method 300 for generating a proxy model is shown, according to an embodiment of the present disclosure.
参考图3，方法300从步骤301开始。在步骤301中，根据划分策略集合，将历史感测数据和历史影响因素数据划分到多个子组。Referring to FIG. 3 , method 300 starts at step 301 . In step 301, historical sensing data and historical influencing factor data are divided into multiple subgroups according to a set of division strategies.
接着，方法300行进到步骤302。在步骤302中，针对多个子组中的每个子组：使用CFD模型来构建相应的仿真结果数据子集。Next, the method 300 proceeds to step 302 . In step 302, for each of the plurality of subgroups: using a CFD model to construct a corresponding subset of simulation result data.
接着，方法300行进到步骤303。在步骤303中，基于与多个子组相对应的多个子代理模型，生成代理模型。类似地，可以将子组的参数作为输入，仿真结果数据子集作为输出，将由输入和相应输出组成的数据集合划分为训练集和测试集。基于训练集，通过使用算法(例如，kriging、KPLS、多精度kriging等)来训练生成子代理模型，并基于测试集，来对子代理模型进行验证。Next, the method 300 proceeds to step 303 . In step 303, a proxy model is generated based on a plurality of sub-agent models corresponding to a plurality of sub-groups. Similarly, the parameters of the subgroups can be taken as input and a subset of simulation result data can be used as output, and the data set consisting of the input and the corresponding output can be divided into a training set and a test set. Based on the training set, the sub-agent model is trained by using an algorithm (for example, kriging, KPLS, multi-precision kriging, etc.), and based on the test set, the sub-agent model is verified.
在图3的方法300中，通过将历史感测数据和历史影响因素数据划分成多个子组，可以有效地降低CFD仿真生成的数据量，例如，数据量与参数个数成指数比例，通过将8个参数分组为2个参数、3个参数、3个参数的三个子组，可以将仿真数据量从2 8＝256减少到2 2+2 3+2 3＝20，这不仅大大降低了数据存储，而且可以快速生成子代理模型，并基于子代理模型生成代理模型。由于历史感测数据和历史影响因素数据的数量可以是几十、几百甚至更多，将这些历史感测数据和历史影响因素数据分成子组可以显著减少构建代理模型所需的参考数据，并可以通过整合由每个子组计算的预测影响来计算历史感测数据和历史影响因素数据对环境状态的总体影响。 In the method 300 of FIG. 3 , by dividing the historical sensing data and historical influencing factor data into multiple subgroups, the amount of data generated by CFD simulation can be effectively reduced, for example, the amount of data is exponentially proportional to the number of parameters, by dividing 8 parameters are grouped into three subgroups of 2 parameters, 3 parameters, and 3 parameters, which can reduce the amount of simulation data from 2 8 =256 to 2 2 +2 3 +2 3 =20, which not only greatly reduces the data storage, and can quickly generate a sub-agent model, and generate a proxy model based on the sub-agent model. Since the amount of historical sensing data and historical influencing factor data can be dozens, hundreds or even more, dividing these historical sensing data and historical influencing factor data into subgroups can significantly reduce the reference data required to construct the proxy model, and The overall impact of the historical sensing data and historical influencing factor data on the state of the environment can be calculated by integrating the predicted impact calculated by each subgroup.
转到图4，示出根据本公开的实施例的用于生成代理模型的另一个示例性方法400的流程图。Turning to FIG. 4 , there is shown a flowchart of another exemplary method 400 for generating a proxy model according to an embodiment of the present disclosure.
参考图4，方法400从步骤401开始。在步骤401中，划分策略集合包括多个策略，多个策略中的每个策略指定从历史感测数据和历史影响因素数据到多个子组的不同划分，可以针对每个策略，采用方法300来根据每个策略相对应的多个子组来生成代理模型。Referring to FIG. 4 , method 400 starts at step 401 . In step 401, the division strategy set includes a plurality of strategies, and each strategy in the plurality of strategies specifies different divisions from historical sensing data and historical influencing factor data into a plurality of subgroups, and method 300 may be used for each strategy to Proxy models are generated from multiple subgroups corresponding to each strategy.
接着，方法400行进到步骤402。在步骤402中，基于历史感测数据和历史影响因素数据，通过与多个策略相对应的多个代理模型生成相应的封闭空间环境100的整体环境状态值。Next, the method 400 proceeds to step 402 . In step 402 , based on historical sensing data and historical influencing factor data, corresponding overall environmental state values of the enclosed space environment 100 are generated through multiple agent models corresponding to multiple policies.
接着，方法400行进到步骤403。在步骤403中，将所生成的相应的封闭空间环境100的整体环境状态值和与封闭空间环境100相关联的历史环境状态值进行比较。例如，环境100的历史环境状态值可以是通过在一段时间内在环境100内布置相关的传感或检测设备进行感测或检测来获得的。Next, the method 400 proceeds to step 403 . In step 403 , the generated overall environmental state value for the corresponding enclosed space environment 100 is compared with historical environmental state values associated with the enclosed space environment 100 . For example, the historical environmental state value of the environment 100 may be obtained by arranging relevant sensing or detection devices in the environment 100 for sensing or detection within a period of time.
接着，方法400行进到步骤404。在步骤404中，根据比较结果，将与多个策略相对应的多个代理模型中的一个确定为用于封闭空间环境100的代理模型。例如，可以将与历史环境状态值具有最佳精度或最小误差(例如，均方差等)的整体环境状态值相对应的代理模型设置为用于封闭空间环境100的代理模型，以实现最佳的预测效果。Next, the method 400 proceeds to step 404 . In step 404, one of the plurality of proxy models corresponding to the plurality of policies is determined as the proxy model for the enclosed space environment 100 according to the comparison result. For example, a proxy model corresponding to an overall environmental state value with the best accuracy or minimum error (e.g., mean square error, etc.) with historical environmental state values may be set as the proxy model for the enclosed space environment 100 to achieve the best predictive effect.
在图4的方法400中，可以从基于不同划分策略生成的代理模型中选择具有最佳预测效果的代理模型，来进一步提高代理模型的准确性。In the method 400 of FIG. 4 , the proxy model with the best prediction effect can be selected from the proxy models generated based on different partitioning strategies, so as to further improve the accuracy of the proxy model.
下面给出一个示例来具体说明上述图3和图4的用于生成代理模型的方法300和400。An example is given below to specifically illustrate the above-mentioned methods 300 and 400 for generating the proxy model in FIG. 3 and FIG. 4 .
以封闭空间环境100为数据中心室为示例，例如，室温状态的影响因素数据包括天气信息、4排的每个机架的负载、3个空调器设备(AHU)的控制参数，感测数据包括每排中间的温度和每个AHU的风量。在以下的表1中示出了示例性划分策略集合和相应的示例性子组划分。Taking the closed space environment 100 as an example of a data center room, for example, the data on the influencing factors of the room temperature state include weather information, the load of each rack in the 4 rows, and the control parameters of the 3 air conditioner equipment (AHU), and the sensing data includes The temperature in the middle of each row and the air volume of each AHU. An exemplary set of partition policies and corresponding exemplary subgroup partitions are shown in Table 1 below.
划分策略集合可以包括策略1、策略2、策略3和策略4。这些策略1～4可以例如将输入按类别划分(例如，内部因素、外部因素和控制参数)，将输入按详细类别划分(例如，将热源按位置分为几类)，将传感器信息与模型输入结合起来输入到一个类别中(例如将位置内的热源和温度传感器合并为一个类别)。换句话说，可以将感测数据和影响因素数据分为可调整的控 制参数和不可调整的模型输入，并将控制参数和模型输入划分成多个子组。例如，策略1将控制参数和模型输入划分成3个子组1.1～1.3，策略2将控制参数和模型输入划分成5个子组2.1～2.5，策略3将控制参数和模型输入划分成8个子组3.1～3.8，策略4将控制参数和模型输入划分成8个子组4.1～4.8。在这些策略1～4中，可以仅将控制参数和模型输入的一部分划分到多个子组中，使得某些控制参数和模型输入并不包含在任何子组内。The partition policy set may include policy 1, policy 2, policy 3 and policy 4. These strategies 1-4 can, for example, divide inputs into categories (eg, internal factors, external factors, and control parameters), divide inputs into detailed categories (eg, divide heat sources into categories by location), combine sensor information with model inputs Combine inputs into one category (for example combine heat sources and temperature sensors within a location into one category). In other words, the sensing data and influencing factor data can be divided into adjustable control parameters and non-adjustable model inputs, and the control parameters and model inputs can be divided into multiple subgroups. For example, strategy 1 divides control parameters and model inputs into 3 subgroups 1.1-1.3, strategy 2 divides control parameters and model inputs into 5 subgroups 2.1-2.5, and strategy 3 divides control parameters and model inputs into 8 subgroups 3.1 ~3.8, Strategy 4 divides control parameters and model inputs into 8 subgroups 4.1~4.8. In these strategies 1-4, only a part of the control parameters and model inputs can be divided into multiple subgroups, so that some control parameters and model inputs are not included in any subgroup.
Figure PCTCN2021109345-appb-000001
Figure PCTCN2021109345-appb-000001
表1：划分策略集合和子组划分Table 1: Partition strategy set and subgroup partition
根据方法300和方法400，可以针对具有M个子组中的每个子组i，获得相应的子代理模型E i，表示子组i中的参数对某个网格点的环境状态的相对影响。所有参数对某个网格点的环境状态的影响可以通过为每个子组生成的代理模型进行整合。例如，如果在整合过程中应用线性叠加，则整合的代理模型E all可以表示成如下： According to the method 300 and the method 400, for each subgroup i in M subgroups, a corresponding sub-agent model E i can be obtained, representing the relative influence of the parameters in the subgroup i on the environment state of a certain grid point. The influence of all parameters on the environmental state of a certain grid point can be integrated through proxy models generated for each subgroup. For example, if linear superposition is applied during integration, the integrated surrogate model E all can be expressed as follows:
Figure PCTCN2021109345-appb-000002
Figure PCTCN2021109345-appb-000002
或者，可以对子代理模型采用加权来生成用于环境100的整合的代理模型E all，以获得具有最佳精度的整合代理模型E all。 Alternatively, weighting may be applied to the sub-agent models to generate an integrated proxy model E all for the environment 100 to obtain an integrated proxy model E all with the best accuracy.
根据方法400，可以采用方法300得到与多个划分策略相对应的多个代理模型E all (j)，并从这多个代理模型中确定具有最佳预测效果的代理模型，例如，通过将代理模型的输出与实际获得的历史环境状态值进行比较。 According to the method 400, the method 300 can be used to obtain multiple proxy models E all (j) corresponding to multiple partition strategies, and determine the proxy model with the best predictive effect from these multiple proxy models, for example, by using the proxy The output of the model is compared with the actual obtained historical environmental state values.
在一些实施例中，多个网格点包括区域集合，用于封闭空间环境100的 代理模型是针对区域集合中的区域来确定的。在一些示例中，可以在所有网格点作为一个区域，使得所有网格点上使用相同的具有最佳预测效果的代理模型，如根据方法300或400所生成的代理模型。或者，在一些示例中，可以在每个网格点上应用具有最佳预测效果的代理模型。或者，在一些示例中，可以将网格点划分为多个区域(可以按照物理位置，或其他方式进行划分)，并在每个区域上确定应用具有最佳预测效果的代理模型。In some embodiments, the plurality of grid points comprises a set of regions for which a proxy model for the enclosed space environment 100 is determined. In some examples, all grid points can be regarded as a region, so that all grid points use the same proxy model with the best prediction effect, such as the proxy model generated according to the method 300 or 400 . Or, in some examples, a surrogate model with the best predictive performance can be applied at each grid point. Or, in some examples, the grid points can be divided into multiple regions (which can be divided according to physical location or other methods), and the proxy model with the best predictive effect can be applied to each region.
回到图2，方法200还可以可选地包括步骤204～207。Returning to FIG. 2 , the method 200 may also optionally include steps 204 - 207 .
在步骤204，方法200可以包括：获取与多个策略相对应的多个代理模型；基于当前感测数据以及当前影响因素数据，使用多个代理模型中的每个代理模型来确定封闭空间环境的相应的一组当前整体环境状态值；基于与多个代理模型相对应的多组当前整体环境状态值，通过交叉验证来检测所述封闭空间环境内的对象的异常，该对象与当前感测数据和当前影响因素数据中的一个相关联。例如，可以从模型数据或其他存储装置访问多个代理模型。如前所述，每个策略指定了历史感测数据和历史影响因素数据(或者说控制参数和模型输入)到多个子组的不同划分，也就是说可以使得控制参数和模型输入在不同策略下被划分到不同的子组内，或者在某些策略下没有包含在任何子组内。因此，该步骤207可以交叉验证来比较控制参数和模型输入的存在或不存在对整体环境状态值的影响，从而检测与控制参数和模型输入相关联的对象(例如，传感器或设备等)是否异常。At step 204, the method 200 may include: acquiring a plurality of proxy models corresponding to a plurality of strategies; based on current sensing data and current influencing factor data, using each of the plurality of proxy models to determine the a corresponding set of current overall environment state values; based on the plurality of sets of current overall environment state values corresponding to the plurality of agent models, anomalies of objects within the enclosed space environment are detected by cross-validation, the objects being consistent with the current sensed data Associated with one of the current influencer data. For example, multiple proxy models may be accessed from model data or other storage. As mentioned earlier, each strategy specifies the different divisions of historical sensing data and historical influencing factor data (or control parameters and model inputs) into multiple subgroups, that is to say, control parameters and model inputs can be made under different strategies are classified into different subgroups, or under certain policies are not included in any subgroup. Therefore, this step 207 can cross-validate to compare the influence of the presence or absence of control parameters and model inputs on the overall environment state value, thereby detecting whether the objects (such as sensors or equipment, etc.) associated with control parameters and model inputs are abnormal .
在步骤205，方法200可以包括：将当前整体环境状态值或预测的整体环境状态值(例如，未来的期望的整体环境状态值)设定为初始状态；基于初始状态，利用优化算法(例如，差分演化(DE)、粒子群(PSO)、模拟退火、蚁群算法等)执行以下步骤：At step 205, the method 200 may include: setting the current overall environmental state value or predicted overall environmental state value (for example, a future expected overall environmental state value) as an initial state; based on the initial state, using an optimization algorithm (for example, Differential evolution (DE), particle swarm optimization (PSO), simulated annealing, ant colony algorithm, etc.) perform the following steps:
调整步骤，对当前感测数据和当前影响因素数据中可调整的参数进行调整；The adjustment step is to adjust the adjustable parameters in the current sensing data and the current influencing factor data;
更新步骤，基于经调整的感测数据和影响因素数据，使用代理模型来确定封闭空间环境的更新的环境状态值；an updating step of using the proxy model to determine an updated environmental state value for the enclosed space environment based on the adjusted sensing data and influencing factor data;
判断步骤，确定封闭空间环境的更新的环境状态值是否达到优化目标，如果未达到优化目标，则转至调整步骤，如果已达到优化目标，则提供可调整的参数的优化值。The judging step is to determine whether the updated environmental state value of the enclosed space environment reaches the optimization target, if not, go to the adjustment step, and if the optimization target has been reached, provide the optimal value of the adjustable parameter.
例如，优化目标是在将关键环境KPI保持在设计范围内的同时获得优 化的能源效率，可调整的参数是诸如AHU的控制参数(例如，温度设置点、风速设置点等)等，外部影响因素数据可以例如包括阶梯电价分布区间等。然而，现有的CFD建模分析是无法实现此类优化控制的。相比之下，在该步骤205中，可以通过调整参数，将调后的整参数输入到代理模型来获得模型输出，并比较是否达到优化目标，使用优化算法来快速迭代地实现优化控制。当优化目标实现时，可以将所调整的参数的优化值提供给管理单元，也就是说，可以向管理单元提供优化控制策略以调整控制参数。For example, the optimization goal is to obtain optimized energy efficiency while keeping key environmental KPIs within the design range, and the adjustable parameters are control parameters such as AHU (e.g., temperature set point, wind speed set point, etc.), external influencing factors The data may, for example, include a stepwise electricity price distribution interval and the like. However, the existing CFD modeling analysis cannot realize this kind of optimal control. In contrast, in step 205, parameters can be adjusted, input the adjusted whole parameters to the proxy model to obtain model output, and compare whether the optimization goal is achieved, and use an optimization algorithm to quickly and iteratively realize optimal control. When the optimization target is achieved, the optimized value of the adjusted parameter can be provided to the management unit, that is, an optimized control strategy can be provided to the management unit to adjust the control parameter.
在步骤206，方法200可以包括：基于当前整体环境状态值，检测封闭空间环境的环境异常；响应于检测到环境异常，生成报警。例如，可以检测到局部过热或过冷，风速异常等状况，并基于检测到的异常状况，生成报警以提醒例如工程师或操作人员。在一些示例中，还可以将当前整体环境状态值处理成能以可视化形式展现的数据格式，使得当前整体环境状态值能够图形化形式呈现在用户界面或显示设备上。At step 206, the method 200 may include: detecting an environmental anomaly of the enclosed space environment based on the current overall environmental state value; and generating an alarm in response to detecting the environmental anomaly. For example, conditions such as localized overheating or overcooling, abnormal wind speed, etc. can be detected, and based on the detected abnormal conditions, alarms are generated to alert eg engineers or operators. In some examples, the current overall environmental state value can also be processed into a data format that can be displayed in a visual form, so that the current overall environmental state value can be presented graphically on a user interface or a display device.
在步骤207，方法200可以包括：利用当前感测数据和当前影响因素数据来更新历史感测数据和历史影响因素数据；并且基于经更新的历史感测数据和历史影响因素数据以及CFD模型来更新所述代理模型。在该步骤207中，可以随着时间对历史数据进行更新，并利用更新后的历史数据来重新生成代理模型，以提高代理模型的准确性。At step 207, the method 200 may include: updating historical sensing data and historical influencing factor data with current sensing data and current influencing factor data; and updating based on the updated historical sensing data and historical influencing factor data and the CFD model The proxy model. In this step 207, the historical data can be updated over time, and the proxy model can be regenerated using the updated historical data, so as to improve the accuracy of the proxy model.
根据前述的方法200，具有如下优点：可以通过从多个输入源读取输入的代理模型，对封闭空间进行高精度实时三维的整体环境监测，因此，可以检测和报告整个封闭空间中与环境要求的偏差；通过应用代理模型可以为可靠的整体环境状态保证和/或能耗优化提供动态控制策略推荐；采用参数划分策略，显著减少了构建代理模型的计算资源，加快了***部署，降低了仿真成本；可以通过多个整合的代理模型计算结果的交叉验证来检测和诊断传感器/设备异常，这可以增加环境控制的稳定性和可靠性；可以集成到外部管理***(例如，资产管理***等)中，用于实现其他功能，如数据中心的容量规划和冷冻水***的需求/顺序设计。According to the aforementioned method 200, it has the following advantages: by reading the input proxy model from multiple input sources, it is possible to perform high-precision real-time three-dimensional overall environmental monitoring of the enclosed space, therefore, it is possible to detect and report the environment requirements in the entire enclosed space. The deviation of the proxy model can be used to provide dynamic control strategy recommendations for reliable overall environmental state assurance and/or energy consumption optimization; the parameter division strategy is used to significantly reduce the computing resources for constructing proxy models, speed up system deployment, and reduce simulation time. Cost; sensor/equipment anomalies can be detected and diagnosed through cross-validation of calculation results of multiple integrated proxy models, which can increase the stability and reliability of environmental control; can be integrated into external management systems (e.g., asset management systems, etc.) , for additional functions such as capacity planning for data centers and demand/sequence design for chilled water systems.
图5示出了根据本公开实施例的用于监控封闭空间环境的示例性装置500。FIG. 5 illustrates an exemplary apparatus 500 for monitoring an enclosed space environment according to an embodiment of the disclosure.
参考图5，装置500包括获取单元501、模型单元502和确定单元503。获取单元501被配置为执行如上文关于方法200中的步骤201描述的过程， 模型单元502被配置为执行如上文关于方法200中的步骤202描述的过程，确定单元503被配置为执行如上文关于方法200中的步骤203描述的过程，其中，代理模型是根据如上文关于方法200中所描述的过程来生成的。Referring to FIG. 5 , an apparatus 500 includes an acquisition unit 501 , a model unit 502 and a determination unit 503 . The obtaining unit 501 is configured to perform the process described above in relation to step 201 of the method 200, the model unit 502 is configured to perform the process described above in relation to step 202 in the method 200, and the determining unit 503 is configured to perform the process described above in relation to The process described in step 203 in the method 200, wherein the proxy model is generated according to the process described in relation to the method 200 above.
装置500还可以可选地包括检测单元504、优化单元505、异常处理单元506和更新单元507。检测单元504被配置为执行如上文关于方法200中的步骤204描述的过程，优化单元505被配置为执行如上文关于方法200中的步骤205描述的过程，异常处理单元506被配置为执行如上文关于方法200中的步骤206描述的过程，更新单元507被配置为执行如上文关于方法200中的步骤207描述的过程。装置500还可以可选地包括模型生成单元，模型生成单元被配置为执行如上文关于方法200中生成代理模型的过程。The apparatus 500 may also optionally include a detection unit 504 , an optimization unit 505 , an exception processing unit 506 and an update unit 507 . The detection unit 504 is configured to execute the process as described above about step 204 in the method 200, the optimization unit 505 is configured to execute the process described above about step 205 in the method 200, and the exception processing unit 506 is configured to execute the process described above about step 205 in the method 200. Regarding the process described in step 206 in method 200 , the updating unit 507 is configured to perform the process as described above in relation to step 207 in method 200 . The apparatus 500 may also optionally include a model generation unit configured to perform the process of generating a proxy model as described above in relation to the method 200 .
图6出了根据本公开的实施例的用于监控封闭空间环境的示例性计算设备600的框图。计算设备600包括处理器601和与处理器601耦合的存储器602。存储器602用于存储计算机可执行指令，当计算机可执行指令被执行时使得处理器601执行以上实施例中的方法(例如，前述的方法200的任何一个或多个步骤)。FIG. 6 illustrates a block diagram of an exemplary computing device 600 for monitoring an enclosed space environment, according to an embodiment of the disclosure. The computing device 600 includes a processor 601 and a memory 602 coupled with the processor 601 . The memory 602 is used to store computer-executable instructions, and when the computer-executable instructions are executed, the processor 601 executes the methods in the above embodiments (for example, any one or more steps of the foregoing method 200).
此外，替代地，上述方法能够通过计算机可读存储介质来实现。计算机可读存储介质上载有用于执行本公开的各个实施例的计算机可读程序指令。计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是但不限于电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括：便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身，诸如无线电波或者其他自由传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如，通过光纤电缆的光脉冲)、或者通过电线传输的电信号。In addition, alternatively, the above method can be implemented by a computer-readable storage medium. A computer-readable storage medium carries computer-readable program instructions for implementing various embodiments of the present disclosure. A computer readable storage medium may be a tangible device that can retain and store instructions for use by an instruction execution device. A computer readable storage medium may be, for example, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or flash memory), static random access memory (SRAM), compact disc read only memory (CD-ROM), digital versatile disc (DVD), memory stick, floppy disk, mechanically encoded device, such as a printer with instructions stored thereon A hole card or a raised structure in a groove, and any suitable combination of the above. As used herein, computer-readable storage media are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., pulses of light through fiber optic cables), or transmitted electrical signals.
因此，在另一个实施例中，本公开提出了一种计算机可读存储介质，该计算机可读存储介质具有存储在其上的计算机可执行指令，计算机可执行指 令用于执行本公开的各个实施例中的方法。Accordingly, in another embodiment, the present disclosure provides a computer-readable storage medium having computer-executable instructions stored thereon for performing various implementations of the present disclosure. method in the example.
在另一个实施例中，本公开提出了一种计算机程序产品，该计算机程序产品被有形地存储在计算机可读存储介质上，并且包括计算机可执行指令，该计算机可执行指令在被执行时使至少一个处理器执行本公开的各个实施例中的方法。In another embodiment, the present disclosure provides a computer program product tangibly stored on a computer-readable storage medium and comprising computer-executable instructions that, when executed, cause At least one processor executes the methods in various embodiments of the present disclosure.
一般而言，本公开的各个示例实施例可以在硬件或专用电路、软件、固件、逻辑，或其任何组合中实施。某些方面可以在硬件中实施，而其他方面可以在可以由控制器、微处理器或其他计算设备执行的固件或软件中实施。当本公开的实施例的各方面被图示或描述为框图、流程图或使用某些其他图形表示时，将理解此处描述的方框、装置、***、技术或方法可以作为非限制性的示例在硬件、软件、固件、专用电路或逻辑、通用硬件或控制器或其他计算设备，或其某些组合中实施。In general, the various example embodiments of the present disclosure may be implemented in hardware or special purpose circuits, software, firmware, logic, or any combination thereof. Certain aspects may be implemented in hardware, while other aspects may be implemented in firmware or software, which may be executed by a controller, microprocessor or other computing device. When aspects of the embodiments of the present disclosure are illustrated or described as block diagrams, flowcharts, or using some other graphical representation, it is to be understood that the blocks, devices, systems, techniques, or methods described herein may serve as non-limiting Examples are implemented in hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controllers or other computing devices, or some combination thereof.
用于执行本公开的各个实施例的计算机可读程序指令或者计算机程序产品也能够存储在云端，在需要调用时，用户能够通过移动互联网、固网或者其他网络访问存储在云端上的用于执行本公开的一个实施例的计算机可读程序指令，从而实施依据本公开的各个实施例所公开的技术方案。The computer-readable program instructions or computer program products used to execute various embodiments of the present disclosure can also be stored in the cloud, and when called, the user can access the program stored on the cloud for execution through the mobile Internet, fixed network or other networks. The computer-readable program instructions of an embodiment of the present disclosure implement the technical solutions disclosed in accordance with various embodiments of the present disclosure.
虽然已经参考若干具体实施例描述了本公开的实施例，但是应当理解，本公开的实施例并不限于所公开的具体实施例。本公开的实施例旨在涵盖在所附权利要求的精神和范围内所包括的各种修改和等同布置。权利要求的范围符合最宽泛的解释，从而包含所有这样的修改及等同结构和功能。While embodiments of the present disclosure have been described with reference to several specific embodiments, it is to be understood that the embodiments of the present disclosure are not limited to the specific embodiments disclosed. Embodiments of the present disclosure are intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. The scope of the claims is accorded the broadest interpretation to encompass all such modifications and equivalent structures and functions.

Claims (25)

  1. 用于监控封闭空间环境的方法，其特征在于，所述方法包括：A method for monitoring an enclosed space environment, characterized in that the method comprises:
    获取与封闭空间环境相关联的当前感测数据和当前影响因素数据；obtaining current sensing data and current influencing factor data associated with the enclosed space environment;
    获取用于所述封闭空间环境的代理模型，所述代理模型是基于与所述封闭空间环境相关联的历史感测数据和历史影响因素数据以及CFD模型来生成的；obtaining a proxy model for the enclosure environment, the proxy model generated based on historical sensing data and historical influencing factor data associated with the enclosure environment and a CFD model;
    基于当前感测数据以及当前影响因素数据，使用所述代理模型来确定所述封闭空间环境的当前整体环境状态值。Based on current sensing data and current influencing factor data, the proxy model is used to determine a current overall environmental state value of the enclosed space environment.
  2. 根据权利要求1所述的方法，其特征在于，The method according to claim 1, characterized in that,
    所述封闭空间环境被划分为多个网格点，其中，确定所述封闭空间环境的当前整体环境状态值包括：The enclosed space environment is divided into a plurality of grid points, wherein determining the current overall environmental state value of the enclosed space environment includes:
    基于当前感测数据以及当前影响因素数据，使用所述代理模型来确定所述多个网格点中的每个网格点的环境参数；using the proxy model to determine an environmental parameter for each grid point of the plurality of grid points based on current sensing data and current influencing factor data;
    基于所述每个网格点的环境参数，确定所述封闭空间环境的当前整体环境状态值。Based on the environmental parameters of each grid point, the current overall environmental state value of the enclosed space environment is determined.
  3. 根据权利要求2所述的方法，其特征在于，基于当前感测数据以及当前影响因素数据，使用所述代理模型来确定所述多个网格点中的每个网格点的环境参数包括：The method according to claim 2, wherein, based on current sensing data and current influencing factor data, using the proxy model to determine the environmental parameters of each grid point in the plurality of grid points comprises:
    对当前感测数据以及当前影响因素数据进行数据清洗和预处理；Perform data cleaning and preprocessing on current sensing data and current influencing factor data;
    基于清洗和预处理后的当前感测数据以及当前影响因素数据，使用所述代理模型来确定所述多个网格点中的每个网格点的环境参数。Based on the cleaned and preprocessed current sensing data and current influencing factor data, the proxy model is used to determine an environmental parameter for each grid point of the plurality of grid points.
  4. 根据权利要求2或3所述的方法，其特征在于，所述代理模型是通过以下方式来生成的：The method according to claim 2 or 3, wherein the proxy model is generated in the following manner:
    基于历史感测数据和历史影响因素数据以及所述多个网格点，使用所述CFD模型进行CFD仿真以构建关于所述封闭空间环境的环境状态的仿真结果数据集；Based on historical sensing data and historical influencing factor data and the plurality of grid points, using the CFD model to perform CFD simulation to construct a simulation result data set about the environmental state of the enclosed space environment;
    基于历史感测数据和历史影响因素数据以及所述仿真结果数据集，训练 生成所述代理模型。Based on the historical sensing data and historical influencing factor data and the simulation result data set, training generates the agent model.
  5. 根据权利要求4所述的方法，其特征在于，所述代理模型是通过以下方式来生成的：The method according to claim 4, wherein the proxy model is generated in the following manner:
    根据划分策略集合，将历史感测数据和历史影响因素数据划分到多个子组；Divide historical sensing data and historical influencing factor data into multiple subgroups according to the set of division strategies;
    针对所述多个子组中的每个子组：For each subgroup of the plurality of subgroups:
    使用所述CFD模型来构建相应的仿真结果数据子集；using the CFD model to construct a corresponding simulation result data subset;
    基于所述每个子组以及相应的仿真结果数据子集，训练生成对应的子代理模型；Based on each subgroup and the corresponding subset of simulation result data, train and generate a corresponding sub-agent model;
    基于与所述多个子组相对应的多个子代理模型，生成所述代理模型。The proxy model is generated based on a plurality of sub-agent models corresponding to the plurality of subgroups.
  6. 根据权利要求5所述的方法，其特征在于，所述代理模型是通过以下方式来生成的：The method according to claim 5, wherein the proxy model is generated in the following manner:
    所述划分策略集合包括多个策略，所述多个策略中的每个策略指定从历史感测数据和历史影响因素数据到多个子组的不同划分，并且所述每个策略对应于根据相应的多个子组生成的代理模型；The set of partitioning strategies includes a plurality of strategies, each of which specifies a different partitioning of historical sensing data and historical influencing factor data into a plurality of subgroups, and each strategy corresponds to Surrogate models generated by multiple subgroups;
    基于历史感测数据和历史影响因素数据，通过与所述多个策略相对应的多个代理模型生成相应的所述封闭空间环境的整体环境状态值；generating corresponding overall environmental state values of the enclosed space environment through a plurality of agent models corresponding to the plurality of strategies based on historical sensing data and historical influencing factor data;
    将所生成的相应的所述封闭空间环境的整体环境状态值和与所述封闭空间环境相关联的历史环境状态值进行比较；comparing the generated corresponding overall environmental state value of the enclosure environment with historical environmental state values associated with the enclosure environment;
    根据比较结果，将与所述多个策略相对应的多个代理模型中的一个确定为用于所述封闭空间环境的代理模型。According to the comparison result, one of the plurality of proxy models corresponding to the plurality of policies is determined as the proxy model for the enclosed space environment.
  7. 根据权利要求6所述的方法，其特征在于，所述多个网格点包括区域集合，用于所述封闭空间环境的代理模型是针对所述区域集合中的区域来确定的。The method of claim 6, wherein the plurality of grid points comprises a set of regions, and a proxy model for the enclosed space environment is determined for regions in the set of regions.
  8. 根据权利要求6所述的方法，其特征在于，所述方法还包括：The method according to claim 6, further comprising:
    获取与所述多个策略相对应的多个代理模型；obtaining a plurality of proxy models corresponding to the plurality of strategies;
    基于当前感测数据以及当前影响因素数据，使用所述多个代理模型中的 每个代理模型来确定所述封闭空间环境的相应的一组当前整体环境状态值；using each of the plurality of proxy models to determine a corresponding set of current overall environmental state values for the enclosure environment based on current sensed data and current influencing factor data;
    基于与所述多个代理模型相对应的多组当前整体环境状态值，通过交叉验证来检测所述封闭空间环境内的对象的异常，所述对象与当前感测数据和当前影响因素数据中的一个相关联。Based on multiple sets of current overall environment state values corresponding to the plurality of agent models, anomalies of objects in the enclosed space environment that are consistent with current sensing data and current influencing factor data are detected by cross-validation an associated.
  9. 根据权利要求1所述的方法，其特征在于，所述方法还包括：The method according to claim 1, further comprising:
    将当前整体环境状态值或预测的整体环境状态值设定为初始状态；Set the current overall environmental state value or the predicted overall environmental state value as the initial state;
    基于所述初始状态，利用优化算法执行以下步骤：Based on the initial state, the optimization algorithm is used to perform the following steps:
    调整步骤，对所述当前感测数据和当前影响因素数据中可调整的参数进行调整；An adjusting step, adjusting the adjustable parameters in the current sensing data and the current influencing factor data;
    更新步骤，基于经调整的感测数据和影响因素数据，使用所述代理模型来确定所述封闭空间环境的更新的环境状态值；an updating step of using the proxy model to determine an updated environmental state value for the enclosure environment based on the adjusted sensed data and influencing factor data;
    判断步骤，确定所述封闭空间环境的更新的环境状态值是否达到优化目标，如果未达到优化目标，则转至调整步骤，如果已达到优化目标，则提供所述可调整的参数的优化值。The judging step is to determine whether the updated environmental state value of the closed space environment has reached the optimization target, if not, go to the adjustment step, and if the optimization target has been reached, provide the optimal value of the adjustable parameter.
  10. 根据权利要求1所述的方法，其特征在于，所述方法还包括：The method according to claim 1, further comprising:
    基于所述当前整体环境状态值，检测所述封闭空间环境的环境异常；Detecting an environmental anomaly of the enclosed space environment based on the current overall environmental state value;
    响应于检测到所述环境异常，生成报警。An alert is generated in response to detecting the environmental anomaly.
  11. 根据权利要求1所述的方法，其特征在于，所述方法还包括：The method according to claim 1, further comprising:
    利用当前感测数据和当前影响因素数据来更新历史感测数据和历史影响因素数据；并且updating historical sensing data and historical influencing factor data with current sensing data and current influencing factor data; and
    基于经更新的历史感测数据和历史影响因素数据以及CFD模型来更新所述代理模型。The proxy model is updated based on updated historical sensory data and historical influencer data and the CFD model.
  12. 用于监控封闭空间环境的装置，其特征在于，所述装置包括：A device for monitoring an enclosed space environment, characterized in that the device comprises:
    获取单元，所述获取单元被配置为获取与封闭空间环境相关联的当前感测数据和当前影响因素数据；an acquisition unit configured to acquire current sensing data and current influencing factor data associated with the enclosed space environment;
    模型单元，所述模型单元被配置为获取用于所述封闭空间环境的代理模型，所述代理模型是基于与所述封闭空间环境相关联的历史感测数据和历史 影响因素数据以及CFD模型来生成的；a model unit configured to obtain a proxy model for the enclosed space environment, the proxy model based on historical sensing data and historical influencing factor data associated with the enclosed space environment and a CFD model Generated;
    确定单元，所述确定单元被配置为基于当前感测数据以及当前影响因素数据，使用所述代理模型来确定所述封闭空间环境的当前整体环境状态值。A determining unit configured to use the proxy model to determine a current overall environmental state value of the enclosed space environment based on current sensing data and current influencing factor data.
  13. 根据权利要求12所述的装置，其特征在于，The device according to claim 12, characterized in that,
    所述封闭空间环境被划分为多个网格点，其中，所述确定单元被进一步配置为：The enclosed space environment is divided into a plurality of grid points, wherein the determining unit is further configured to:
    基于当前感测数据以及当前影响因素数据，使用所述代理模型来确定所述多个网格点中的每个网格点的环境参数；using the proxy model to determine an environmental parameter for each grid point of the plurality of grid points based on current sensing data and current influencing factor data;
    基于所述每个网格点的环境参数，确定所述封闭空间环境的当前整体环境状态值。Based on the environmental parameters of each grid point, the current overall environmental state value of the enclosed space environment is determined.
  14. 根据权利要求13所述的装置，其特征在于，所述确定单元被进一步配置为：The device according to claim 13, wherein the determining unit is further configured to:
    对当前感测数据以及当前影响因素数据进行数据清洗和预处理；Perform data cleaning and preprocessing on current sensing data and current influencing factor data;
    基于清洗和预处理后的当前感测数据以及当前影响因素数据，使用所述代理模型来确定所述多个网格点中的每个网格点的环境参数。Based on the cleaned and preprocessed current sensing data and current influencing factor data, the proxy model is used to determine an environmental parameter for each grid point of the plurality of grid points.
  15. 根据权利要求13或14所述的装置，其特征在于，所述代理模型是通过以下方式来生成的：The device according to claim 13 or 14, wherein the proxy model is generated in the following manner:
    基于历史感测数据和历史影响因素数据以及所述多个网格点，使用所述CFD模型进行CFD仿真以构建关于所述封闭空间环境的环境状态的仿真结果数据集；Based on historical sensing data and historical influencing factor data and the plurality of grid points, using the CFD model to perform CFD simulation to construct a simulation result data set about the environmental state of the enclosed space environment;
    基于历史感测数据和历史影响因素数据以及所述仿真结果数据集，训练生成所述代理模型。Based on historical sensing data, historical influencing factor data and the simulation result data set, training generates the agent model.
  16. 根据权利要求15所述的装置，其特征在于，所述代理模型是通过以下方式来生成的：The device according to claim 15, wherein the proxy model is generated in the following manner:
    根据划分策略集合，将历史感测数据和历史影响因素数据划分到多个子组；Divide historical sensing data and historical influencing factor data into multiple subgroups according to the set of division strategies;
    针对所述多个子组中的每个子组：For each subgroup of the plurality of subgroups:
    使用所述CFD模型来构建相应的仿真结果数据子集；using the CFD model to construct a corresponding simulation result data subset;
    基于所述每个子组以及相应的仿真结果数据子集，训练生成对应的子代理模型；Based on each subgroup and the corresponding subset of simulation result data, train and generate a corresponding sub-agent model;
    基于与所述多个子组相对应的多个子代理模型，生成所述代理模型。The proxy model is generated based on a plurality of sub-agent models corresponding to the plurality of subgroups.
  17. 根据权利要求16所述的装置，其特征在于，所述代理模型是通过以下方式来生成的：The device according to claim 16, wherein the proxy model is generated by:
    所述划分策略集合包括多个策略，所述多个策略中的每个策略指定从历史感测数据和历史影响因素数据到多个子组的不同划分，并且所述每个策略对应于根据相应的多个子组生成的代理模型；The set of partitioning strategies includes a plurality of strategies, each of which specifies a different partitioning of historical sensing data and historical influencing factor data into a plurality of subgroups, and each strategy corresponds to Surrogate models generated by multiple subgroups;
    基于历史感测数据和历史影响因素数据，通过与所述多个策略相对应的多个代理模型生成相应的所述封闭空间环境的整体环境状态值；generating corresponding overall environmental state values of the enclosed space environment through a plurality of agent models corresponding to the plurality of strategies based on historical sensing data and historical influencing factor data;
    将所生成的相应的所述封闭空间环境的整体环境状态值和与所述封闭空间环境相关联的历史环境状态值进行比较；comparing the generated corresponding overall environmental state value of the enclosure environment with historical environmental state values associated with the enclosure environment;
    根据比较结果，将与所述多个策略相对应的多个代理模型中的一个确定为用于所述封闭空间环境的代理模型。According to the comparison result, one of the plurality of proxy models corresponding to the plurality of policies is determined as the proxy model for the enclosed space environment.
  18. 根据权利要求17所述的装置，其特征在于，所述多个网格点包括区域集合，用于所述封闭空间环境的代理模型是针对所述区域集合中的区域来确定的。The apparatus of claim 17, wherein the plurality of grid points comprises a set of regions, a proxy model for the enclosed space environment is determined for regions in the set of regions.
  19. 根据权利要求17所述的装置，其特征在于，所述装置还包括：The device according to claim 17, further comprising:
    检测单元，所述检测单元被配置为：a detection unit, the detection unit is configured to:
    获取与所述多个策略相对应的多个代理模型；obtaining a plurality of proxy models corresponding to the plurality of strategies;
    基于当前感测数据以及当前影响因素数据，使用所述多个代理模型中的每个代理模型来确定所述封闭空间环境的相应的一组当前整体环境状态值；using each of the plurality of proxy models to determine a corresponding set of current overall environmental state values for the enclosure environment based on current sensed data and current influencing factor data;
    基于与所述多个代理模型相对应的多组当前整体环境状态值，通过交叉验证来检测所述封闭空间环境内的对象的异常，所述对象与当前感测数据和当前影响因素数据中的一个相关联。Based on multiple sets of current overall environment state values corresponding to the plurality of agent models, anomalies of objects in the enclosed space environment that are consistent with current sensing data and current influencing factor data are detected by cross-validation an associated.
  20. 根据权利要求12所述的装置，其特征在于，所述装置还包括：The device according to claim 12, wherein the device further comprises:
    优化单元，所述优化单元被配置为：an optimization unit, the optimization unit is configured to:
    将当前整体环境状态值或预测的整体环境状态值设定为初始状态；Set the current overall environmental state value or the predicted overall environmental state value as the initial state;
    基于所述初始状态，利用优化算法执行以下步骤：Based on the initial state, the optimization algorithm is used to perform the following steps:
    调整步骤，对所述当前感测数据和当前影响因素数据中可调整的参数进行调整；An adjusting step, adjusting the adjustable parameters in the current sensing data and the current influencing factor data;
    更新步骤，基于经调整的感测数据和影响因素数据，使用所述代理模型来确定所述封闭空间环境的更新的环境状态值；an updating step of using the proxy model to determine an updated environmental state value for the enclosure environment based on the adjusted sensed data and influencing factor data;
    判断步骤，确定所述封闭空间环境的更新的环境状态值是否达到优化目标，如果未达到优化目标，则转至调整步骤，如果已达到优化目标，则提供所述可调整的参数的优化值。The judging step is to determine whether the updated environmental state value of the closed space environment has reached the optimization target, if not, go to the adjustment step, and if the optimization target has been reached, provide the optimal value of the adjustable parameter.
  21. 根据权利要求12所述的装置，其特征在于，所述装置还包括：The device according to claim 12, wherein the device further comprises:
    异常处理单元，所述异常处理单元被配置为：An exception handling unit, the exception handling unit is configured to:
    基于所述当前整体环境状态值，检测所述封闭空间环境的环境异常；Detecting an environmental anomaly of the enclosed space environment based on the current overall environmental state value;
    响应于检测到所述环境异常，生成报警。An alert is generated in response to detecting the environmental anomaly.
  22. 根据权利要求12所述的装置，其特征在于，所述装置还包括：The device according to claim 12, wherein the device further comprises:
    更新单元，所述更新单元被配置为：an update unit, the update unit is configured to:
    利用当前感测数据和当前影响因素数据来更新历史感测数据和历史影响因素数据；并且updating historical sensing data and historical influencing factor data with current sensing data and current influencing factor data; and
    基于经更新的历史感测数据和历史影响因素数据以及CFD模型来更新所述代理模型。The proxy model is updated based on updated historical sensory data and historical influencer data and the CFD model.
  23. 计算设备，其特征在于，所述计算设备包括：A computing device, wherein the computing device includes:
    处理器；以及processor; and
    存储器，其用于存储计算机可执行指令，当所述计算机可执行指令被执行时使得所述处理器执行根据权利要求1-11中任一项所述的方法。A memory for storing computer-executable instructions which, when executed, cause the processor to perform the method according to any one of claims 1-11.
  24. 计算机可读存储介质，所述计算机可读存储介质具有存储在其上的计算机可执行指令，所述计算机可执行指令用于执行根据权利要求1-11中任 一项所述的方法。A computer-readable storage medium having computer-executable instructions stored thereon for performing the method according to any one of claims 1-11.
  25. 计算机程序产品，所述计算机程序产品被有形地存储在计算机可读存储介质上，并且包括计算机可执行指令，所述计算机可执行指令在被执行时使至少一个处理器执行根据权利要求1-11中任一项所述的方法。A computer program product tangibly stored on a computer-readable storage medium and comprising computer-executable instructions which, when executed, cause at least one processor to perform the any one of the methods described.
PCT/CN2021/109345 2021-07-29 2021-07-29 Method and apparatus for monitoring closed space environment, and computer-readable storage medium WO2023004704A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202180098829.8A CN117413273A (en) 2021-07-29 2021-07-29 Method, apparatus and computer storage medium for monitoring an enclosure environment
PCT/CN2021/109345 WO2023004704A1 (en) 2021-07-29 2021-07-29 Method and apparatus for monitoring closed space environment, and computer-readable storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/109345 WO2023004704A1 (en) 2021-07-29 2021-07-29 Method and apparatus for monitoring closed space environment, and computer-readable storage medium

Publications (1)

Publication Number Publication Date
WO2023004704A1 true WO2023004704A1 (en) 2023-02-02

Family

ID=85086008

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/109345 WO2023004704A1 (en) 2021-07-29 2021-07-29 Method and apparatus for monitoring closed space environment, and computer-readable storage medium

Country Status (2)

Country Link
CN (1) CN117413273A (en)
WO (1) WO2023004704A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116502319A (en) * 2023-06-28 2023-07-28 长江三峡集团实业发展(北京)有限公司 Three-dimensional temperature field reconstruction method and device for concrete dam and electronic equipment

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007072991A (en) * 2005-09-09 2007-03-22 Kajima Corp Environment evaluation system and environment evaluation method
CN103995548A (en) * 2014-05-13 2014-08-20 江苏大学 Indoor thermal environment control method based on model reduction and multi-model prediction control
US20150134123A1 (en) * 2013-11-14 2015-05-14 Ces Group, Llc Predictive monitoring and control of an environment using cfd
CN104633856A (en) * 2015-01-27 2015-05-20 天津大学 Method for controlling artificial environment by combining CFD numerical simulation and BP neural network
CN107545100A (en) * 2017-07-31 2018-01-05 江苏大学 A kind of high-resolution Reduced Modeling Methods of venlo types greenhouse
CN109840307A (en) * 2018-12-29 2019-06-04 重庆大学 A kind of indoor temperature and humidity field quick predict system and method
KR20190141935A (en) * 2018-06-15 2019-12-26 (주)미래환경플랜건축사사무소 system and method predicting effective ventilation amount to improve indoor air quality in apartment buildings
CN113158593A (en) * 2021-04-13 2021-07-23 中南建筑设计院股份有限公司 Intelligent design method of hospital ward ventilation system based on artificial intelligence

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007072991A (en) * 2005-09-09 2007-03-22 Kajima Corp Environment evaluation system and environment evaluation method
US20150134123A1 (en) * 2013-11-14 2015-05-14 Ces Group, Llc Predictive monitoring and control of an environment using cfd
CN103995548A (en) * 2014-05-13 2014-08-20 江苏大学 Indoor thermal environment control method based on model reduction and multi-model prediction control
CN104633856A (en) * 2015-01-27 2015-05-20 天津大学 Method for controlling artificial environment by combining CFD numerical simulation and BP neural network
CN107545100A (en) * 2017-07-31 2018-01-05 江苏大学 A kind of high-resolution Reduced Modeling Methods of venlo types greenhouse
KR20190141935A (en) * 2018-06-15 2019-12-26 (주)미래환경플랜건축사사무소 system and method predicting effective ventilation amount to improve indoor air quality in apartment buildings
CN109840307A (en) * 2018-12-29 2019-06-04 重庆大学 A kind of indoor temperature and humidity field quick predict system and method
CN113158593A (en) * 2021-04-13 2021-07-23 中南建筑设计院股份有限公司 Intelligent design method of hospital ward ventilation system based on artificial intelligence

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116502319A (en) * 2023-06-28 2023-07-28 长江三峡集团实业发展(北京)有限公司 Three-dimensional temperature field reconstruction method and device for concrete dam and electronic equipment
CN116502319B (en) * 2023-06-28 2023-10-20 长江三峡集团实业发展(北京)有限公司 Three-dimensional temperature field reconstruction method and device for concrete dam and electronic equipment

Also Published As

Publication number Publication date
CN117413273A (en) 2024-01-16

Similar Documents

Publication Publication Date Title
CN103766015B (en) Measuring system and the method for prediction is assisted for data center's air temperature and current value
US20150134123A1 (en) Predictive monitoring and control of an environment using cfd
JP6339238B2 (en) Virtual data center environmental monitoring system
US9857779B2 (en) Multi-dimensional optimization for controlling environmental maintenance modules
JP5593387B2 (en) Methods, products and apparatus for knowledge base modeling for data centers
US8825451B2 (en) System and methods for rack cooling analysis
Yu et al. Virtual in-situ calibration method in building systems
US9996092B2 (en) Determining a time for corrective action in a data center
JP2014182694A (en) Device, method, and program for sensor failure detection
US9857235B2 (en) Real-time modeling of heat distributions
CN110036418B (en) Performance parameterization of care equipment and systems
AU2012236965A1 (en) Systems and methods for predicting fluid dynamics in a data center
US20140316720A1 (en) Data processing system with real-time data center air flow simulator
TW202030617A (en) Data center infrastructure optimization method based on causal learning
JP6181079B2 (en) Computational fluid dynamics system and method of use
JP6438124B2 (en) Operation management system and operation management method
WO2023004704A1 (en) Method and apparatus for monitoring closed space environment, and computer-readable storage medium
CN112989642B (en) Non-transitory computer readable medium, system, and method for generating a display
CN117412555B (en) Air conditioner temperature control system of data center computer lab
US20160188764A1 (en) Raised floor plenum tool
US11561536B2 (en) Building health assessment and commissioning tool with dynamic report generation
JP2013195170A (en) Air flow state prediction device, air flow state prediction method, diffusion state prediction device and diffusion state prediction method
CN115776813B (en) Efficient machine room control method and device
CN113170592A (en) Thermal control optimization based on monitoring/control mechanisms
KR20230075861A (en) Iintelligent complex environmental sensor system and its operating method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21951311

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 202180098829.8

Country of ref document: CN

NENP Non-entry into the national phase

Ref country code: DE