Lean-flooding pre-sensing internet-of-things multi-objective optimization heating and ventilation system

Abstract

The invention discloses a poor-flooding pre-sensing internet-of-things multi-target optimization heating and ventilation system which is obtained by executing the following steps: step 1, data are collected through front sensors arranged on heating equipment, ventilation equipment, air conditioning equipment, corollary equipment and auxiliary equipment, and an initial collection sample library is formed; 2, preprocessing the initial acquisition sample library, performing difference compensation and dimension reduction normalization processing on different modes of rough data after enhancement and generalization through principal component analysis (PCA) and a wavelet modulus maximum if a small amount of data appears due to insufficiency in the data acquired by the front sensor, and extracting key feature vectors, generating an expansion sample library; 3, training the mathematical model; and 4, testing the mathematical model.

Description

Technical Field

[0001] The present invention relates to the field of HVAC technology, and in particular to a poor and universal pre-sensing Internet of Things multi-objective optimization HVAC system. Background Art

[0002] Integrated building energy management mainly involves the management of water, electricity and gas, and an important part of it is the energy-saving management of the HVAC system.

[0003] When social capital participates in integrated energy management, behavioral management is an important source of energy efficiency.

[0004] However, the lack of unconventional data such as energy consumption surges, equipment overload scenarios or equipment anomalies, extreme environmental data, etc. makes it difficult to cover all scenario conditions and optimize behavior management strategies. It is also easy to misjudge the operating status in fault diagnosis and anomaly detection, and behavior management may also lead to a poor user experience.

[0005] Therefore, how to manage and monitor anomalies in real time and optimize energy consumption, and balance multi-objective control strategies to achieve the ultimate goal of behavior management has become a technical problem that technical personnel in this field urgently need to solve. Summary of the invention

[0006] In view of the above-mentioned defects of the prior art, the present invention provides a poor and universal pre-sensing Internet of Things multi-objective optimization HVAC system, which aims to manage and monitor abnormalities in real time and optimize energy consumption, and balance multi-objective control strategies to achieve the ultimate goal of behavior management.

[0007] To achieve the above object, the present invention discloses a poor and universal pre-sensing IoT multi-objective optimization HVAC system, which is obtained by executing the following steps:

[0008] Step 1: Collect data through front sensors installed on heating equipment, ventilation equipment, air conditioning equipment, supporting equipment and auxiliary equipment, and form an initial collection sample library;

[0009] Step 2: pre-process the initial sample library to remove the small amount of data due to the lack of data collected by the front sensor.

[0010] Then, principal component analysis (PCA) and wavelet modulus maximum are used to normalize and compensate the rough data of different modes after enhancement and generalization, and key feature vectors are extracted to generate an expanded sample library.

[0011] Step 3: Training mathematical model;

[0012] Step 4: Mathematical model testing.

[0013] Preferably, the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment all complete the masterless information exchange of data between the corresponding communication modules through the corresponding communication modules and the corresponding pre-sensors in the Internet of Things network of the mesh data stream in the duplex channel.

[0014] Or realize the data management of the cloud server through the edge computing gateway of the aggregation layer;

[0015] The Internet of Things network of the mesh data stream relies on the edge computing gateway of the aggregation layer to complete the communication of the control management system of the cloud server through the cellular / WI-FI network in the MQTT / HTTP protocol.

[0016] The data collected by the pre-sensors includes parameters such as temperature and humidity, energy consumption, and equipment health status, specifically temperature and humidity, energy consumption, and the health status of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment.

[0017] Preferably, in step 1, the data collected by the pre-sensors includes the following four types of data:

[0018] The first type of data is the cooling data, heating data, dehumidification data, ventilation data, sleep data and energy-saving data of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment in the normal energy consumption mode.

[0019] The second type of data is the temperature difference data, airtightness data, long-term operation data, and frequent start-stop data of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment in the abnormal energy consumption mode.

[0020] The third type of data is the strong mode data, initial start-up data, electric auxiliary heating data, and extreme environment operation data of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment in the peak energy consumption mode.

[0021] The fourth type of data is the compressor data, refrigerant data, filter screen data, control system data, circuit data, and fan data of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment in the equipment failure energy consumption mode.

[0022] In this embodiment, the first type of data is sufficient data;

[0023] The second type of data, the third type of data and the fourth type of data are all the small amount of data;

[0024] The method for forming the initial acquisition sample library is as follows: The data collected by the pre-sensor is classified into corresponding target information parameters and task attribute parameters according to the refrigeration mode, heating mode, dehumidification mode, ventilation mode, sleep mode, and energy-saving mode under the normal energy consumption mode,

[0025] the large temperature difference mode, poor room sealing mode, long-term operation mode, and frequent start-stop mode under the abnormal energy consumption mode,

[0026] the strong mode, initial start mode, electric auxiliary heating mode, and extreme environment operation mode under the peak energy consumption mode,

[0027] the compressor failure mode, refrigerant shortage or leakage mode, filter screen blockage mode, control system failure mode, circuit problem mode, fan failure mode, and long-term non-maintenance under the equipment failure energy consumption mode,

[0028] ;

[0029] Among them,

[0030] in the refrigeration mode of the heating equipment, ventilation equipment, air conditioning equipment, supporting equipment, and auxiliary equipment under the normal energy consumption mode, the corresponding target information parameters include the indoor temperature dropping to the set value, the compressor power ratio, and the power consumption, and the corresponding task attribute parameters include the temperature set value, the compressor start-stop logic, and the refrigerant flow control;

[0031] in the heating mode of the heating equipment, ventilation equipment, air conditioning equipment, supporting equipment, and auxiliary equipment under the normal energy consumption mode, the corresponding target information parameters include the indoor temperature rising to the set value, the heat pump efficiency, and the electric auxiliary heating enable rate, and the corresponding task attribute parameters include the temperature set value, the heat pump / electric auxiliary switching threshold, and the outside temperature acquisition parameter;

[0032] in the dehumidification mode of the heating equipment, ventilation equipment, air conditioning equipment, supporting equipment, and auxiliary equipment under the normal energy consumption mode, the corresponding target information parameters include the indoor humidity dropping to the target value, the compressor on time, and the fan speed, and the corresponding task attribute parameters include the humidity set value, the compressor operation time, and the fan speed adjustment;

[0033] in the ventilation mode of the heating equipment, ventilation equipment, air conditioning equipment, supporting equipment, and auxiliary equipment under the normal energy consumption mode, the corresponding target information parameters include the air circulation rate, the fan power ratio, and the environmental ventilation frequency, and the corresponding task attribute parameters include the air volume control target, the filter screen resistance detection, and the air circulation time;

[0034] In the sleep mode of the normal energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the temperature change curve, the night power optimization ratio and the user comfort score, and the corresponding task attribute parameters include the temperature control gradual change curve, the night target power and the compressor dormancy frequency;

[0035] In the energy-saving mode of the normal energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the system energy-saving rate, the target temperature control deviation range and the compressor start-stop frequency, and the corresponding task attribute parameters include the energy-saving optimization algorithm, the power limit parameter and the load balancing coefficient;

[0036] In the excessive temperature difference mode of the abnormal energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the indoor-outdoor temperature difference, the compressor load ratio and the unnecessary energy consumption growth ratio, and the corresponding task attribute parameters include the outdoor temperature acquisition frequency, the compressor high-frequency operation threshold and the alarm parameter;

[0037] In the poor room airtightness mode of the abnormal energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the energy loss rate, the cold / hot air leakage point analysis and the temperature control delay time, and the corresponding task attribute parameters include the cold / hot air loss point analysis task and the environmental heat insulation material evaluation parameter;

[0038] In the long-time operation mode of the abnormal energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the cumulative operation time, the continuous operation time of the compressor and the energy consumption per unit time, and the corresponding task attribute parameters include the timing function trigger task and the equipment continuous high-load monitoring threshold;

[0039] In the frequent start-stop mode of the abnormal energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the start-stop frequency, the peak compressor start-up power and the cyclic energy consumption growth ratio, and the corresponding task attribute parameters include the user behavior record and the minimum switch interval time parameter;

[0040] In the strong mode of the peak energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the system full-load operation time, the target temperature achievement time and the power peak value, and the corresponding task attribute parameters include the full-load start time limit, the target temperature control accuracy and the power upper limit parameter;

[0041] In the initial start mode of the peak energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the compressor current at startup, the target temperature control response time and the refrigerant pressure value, and the corresponding task attribute parameters include the start current limit, the refrigerant pressure monitoring task and the compressor overload protection;

[0042] In the electric auxiliary heating mode of the peak energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the electric auxiliary heater activation duration, the ambient temperature and the target temperature maintenance time, and the corresponding task attribute parameters include the electric auxiliary power adjustment parameter, the minimum activation temperature threshold and the electric heating efficiency evaluation;

[0043] In the extreme environment operation mode of the peak energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the outside temperature deviation degree, the system stability and the compressor load coefficient, and the corresponding task attribute parameters include the outside environment perception parameter, the compressor heat dissipation optimization and the operation stability index;

[0044] In the compressor failure mode of the equipment failure energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the compressor power deviation, the abnormal refrigerant flow rate and the start-stop efficiency decrease ratio, and the corresponding task attribute parameters include the compressor operating frequency, the abnormal vibration alarm and the efficiency deviation threshold;

[0045] In the refrigerant shortage or leakage mode of the equipment failure energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the remaining refrigerant amount, the refrigeration efficiency decrease ratio and the environmental impact coefficient, and the corresponding task attribute parameters include the refrigerant pressure sensor data, the automatic replenishment task and the leakage detection algorithm;

[0046] In the filter clogging mode of the equipment failure energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the air volume obstruction ratio, the filter cleanliness and the system heat dissipation delay rate, and the corresponding task attribute parameters include the air flow resistance analysis, the regular cleaning reminder and the air volume measurement task;

[0047] In the control system failure mode of the equipment failure energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the temperature control deviation range, the system response delay time and the number of abnormal alarms, and the corresponding task attribute parameters include the temperature control logic test, the sensor calibration task and the data abnormality record;

[0048] In the circuit problem mode of the equipment failure energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the current fluctuation amplitude, the overload rate and the aging degree of key components, and the corresponding task attribute parameters include the circuit current stability, the overload monitoring algorithm and the component aging life assessment;

[0049] In the fan failure mode of the equipment failure energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the wind speed deviation, the circulation efficiency decline ratio and the abnormal noise index, and the corresponding task attribute parameters include the fan speed deviation detection, the abnormal noise data collection and the component wear assessment;

[0050] In the long-term unmaintained mode of the equipment failure energy consumption mode of the heating equipment, the ventilation equipment, the air conditioning equipment, the supporting equipment and the auxiliary equipment, the corresponding target information parameters include the dust accumulation coverage rate, the heat exchange efficiency and the operating power consumption increase ratio, and the corresponding task attribute parameters include the maintenance task plan, the component aging coefficient and the trigger of the regular cleaning task.

[0051] Preferably, the heating equipment includes a heat source device for providing heat, a heat dissipation device for transferring heat to the room, and a heating auxiliary device for supporting hot water circulation and control;

[0052] The heat source device includes a boiler, a heat pump and / or an electric heater;

[0053] The heat dissipation device includes a radiator, underfloor heating and / or a fan coil unit;

[0054] The heating auxiliary device includes a water pump and / or a heating control system;

[0055] The ventilation equipment includes natural ventilation that utilizes natural wind to circulate air through windows and ventilation openings, mechanical ventilation for forced ventilation and air exchange, and filtration equipment for purifying air;

[0056] The mechanical ventilation includes exhaust fans, fresh air systems, and / or smoke exhaust equipment;

[0057] The filtration equipment includes air filters and / or electrostatic precipitators;

[0058] The air conditioning equipment includes cold source equipment that provides a cooling function, terminal equipment for adjusting indoor air, and cooling equipment for cooling the cold source equipment;

[0059] The cold source equipment includes chillers and / or absorption refrigerators;

[0060] The terminal equipment includes fan coil units and / or air handling units;

[0061] The cooling equipment includes cooling towers and / or condensers;

[0062] The supporting equipment includes a pipe network system for cold and hot water and air transportation pipelines, and a supporting control system for automatic control;

[0063] The supporting control system includes thermostats and / or intelligent building management systems;

[0064] The auxiliary equipment includes humidifying equipment for increasing air humidity, dehumidifying equipment for reducing air humidity, water treatment equipment for improving water quality, and safety equipment for ensuring safety;

[0065] The humidifying equipment includes humidifiers and / or spray humidifiers;

[0066] The dehumidifying equipment includes condensing dehumidifiers and / or rotary dehumidifiers;

[0067] The water treatment equipment includes water softening equipment and / or descaling equipment;

[0068] The safety equipment includes fire dampers, air volume regulating valves, and / or expansion tanks.

[0069] Preferably, in step 3, based on the generative adversarial network WGAN, energy consumption sparse data in abnormal energy consumption patterns, peak energy consumption patterns, and equipment failure energy consumption patterns are used to extract feature vectors through the intermediate layer of the discriminator to achieve data enhancement, while avoiding overfitting of small sample events;

[0070] The classification model SVM quickly classifies and detects the states of the heating equipment, ventilation equipment, air conditioning equipment, supporting equipment, and auxiliary equipment in normal energy consumption patterns, abnormal energy consumption patterns, peak energy consumption patterns, and equipment failure energy consumption patterns;

[0071] Among them, the classification model SVM is trained and tested on the target information parameters and task attribute parameters in the extended sample library according to the enhanced generalization derivation rules of the WGAN;

[0072] Select the classification model SVM that meets the accuracy requirements for real-time identification of target signal parameters;

[0073] Search for a specific rule sample library that conforms to specific rules in the extended sample library according to the indexing rules and output the corresponding task attribute parameters that meet the conditions;

[0074] Based on the WGAN-SVM model, the initial population of real-time target information parameters and task attribute parameters generates the combination of device switch status and temperature. Through the utility evaluation of energy consumption, operating cost, anomaly detection, and comfort, combined with the crossover and mutation of GA, the control strategy is optimized by the solution with the highest fitness.

[0075] More preferably, the multi-objective optimization NSGA-II function is designed as follows: the design of the preprocessing anomaly detection model is as follows:

[0076] Minimize:

[0077] Among them, |W k (f)| is the modulus maximum value of the collected signal f at the wavelet scale k;

[0078] K represents the number of wavelet transform scales;

[0079] The enhanced generalization model includes:

[0080] Generator:

[0081] Minimize:

[0082] It refers to the evaluation of generating the sample X~ of the environmental data predicted by the generator;

[0083] P gen represents the environmental data predicted by the generator, including indoor temperature and humidity;

[0084] Discriminator:

[0085] Minimize:

[0086] It refers to the evaluation of the real sample X of the real indoor environmental data collected by the air conditioning system;

[0087] P real represents the real indoor environmental data collected by the air conditioning system, including temperature and humidity;

[0088] The classification model is designed as:

[0089] Minimize:

[0090] where ε is the weight vector of the hyperplane of the classification model;

[0091] ξ n is the slack variable;

[0092] C is the penalty coefficient;

[0093] n represents the air-conditioning operation data sample at a certain time point, including room temperature, humidity, energy consumption, current;

[0094] N represents the total number of samples in the entire training data set;

[0095] The energy consumption model is designed as:

[0096] Minimize:

[0097] where P (t) is the power consumption of the HVAC system at time t; T is the time period;

[0098] The comfort model is designed as:

[0099] Minimize:

[0100] where B is the total number of samples involved in comfort calculation;

[0101] Q i is the actual temperature of the i-th space; Q set is the set temperature;

[0102] The cost model is designed as:

[0103]

[0104] where C cost (t) is the electricity cost at time t;

[0105] The multi-objective optimization model is designed as:

[0106] Minimize: J total = θJ1 + μJ5 + πJ6 + ρJ7;

[0107] where θ, μ, π, and ρ are the weight values for balancing energy consumption, operating cost, anomaly detection, and comfort evaluation respectively.

[0108] Preferably, the interaction of the mathematical model is as follows: the IoT network of the mesh data stream collects the environment through the front-end sensors and the corresponding communication modules, and the device operation data of the corresponding heating equipment, the corresponding ventilation equipment, the corresponding air-conditioning equipment, the corresponding supporting equipment, and / or the corresponding auxiliary equipment, and then combines the real-time data and the abnormal missing IoT data, and performs data completion and prediction through WGAN;

[0109] Among them, the abnormal data extracts the fault feature vector through principal component analysis and wavelet modulus maximum processing, trains and tests the classification model with the comprehensive feature vector, and performs pattern recognition and behavior prediction through the optimal model;

[0110] The abnormal data includes compressor signals and / or power curves;

[0111] The sample library of target information parameters and task attribute parameters under the combined rule conditions, when the real-time target information parameters are input, obtains the task attribute parameters corresponding to the rules through the SVM classification model index, and then combines the sample library of target information parameters and task attribute parameters under the specific rule conditions, and through the crossover and mutation of GA, multi-objectively optimizes the energy consumption, operation cost, abnormal detection, and comfort objective functions, and the behavior management sample with the optimization strategy iteratively updates the initial acquisition sample library.

[0112] Preferably, in step 4, one or more target information parameters are input to the trained mathematical model, and the mathematical model generates a specific sample library and obtains the corresponding task attribute parameters.

[0113] Preferably, the initial acquisition sample library generates the combination of device switch states and temperatures based on the initial population of real-time target information parameters and task attribute parameters of the mathematical model, evaluates the practicality of energy consumption, operation cost, abnormal detection, and comfort, and combines the crossover and mutation of GA, and optimizes the control strategy with the highest fitness solution.

[0114] The beneficial effects of the present invention:

[0115] In the ownerless blockchain-based IoT HVAC system of the present invention, relying on IoT nodes and edge gateways to collect parameters such as temperature, humidity, energy consumption, and device status, the robustness of the generative adversarial network WGAN is extended for extreme situation management, and the multi-modal SVM binary classifier is trained and tested on the simulated global data. The task attribute parameters are identified from the target information parameters by the optimization model, and then according to the broad task attribute parameters, the comfort multi-objective optimization of the genetic algorithm provides accurate real-time behavior management and feeds back and iterates the expert sample library through the post-evaluation database.

[0116] The present invention is a typical model for integrated energy behavior management based on WGAN, SVM, and intelligent swarm algorithms, suitable for scenarios in the IoT field such as IoT lighting systems and IoT devices with communication modules.

[0117] The following will further illustrate the concept, specific structure, and technical effects of the present invention in conjunction with the accompanying drawings to fully understand the purpose, features, and effects of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0118] Figure 1 The network topology schematic diagram showing an embodiment of the present invention.

[0119] Figure 2 The schematic diagram of the multi-objective optimization principle showing an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0120] Embodiment

[0121] As Figure 1 and Figure 2 shown, the poor and general pre-perception IoT multi-objective optimization HVAC system is obtained by performing the following steps:

[0122] Step 1: Collect data through the front-end sensors 1 installed in the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45, and form an initial acquisition sample library;

[0123] Step 2: Preprocess the initial acquisition sample library. For the situation of a small amount of data due to deficiencies in the data collected by the front-end sensors 1, such as the lack of data in the extreme environmental operation modes of the peak energy consumption modes of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45, which refers to the lack of data caused by abnormal equipment operation, extreme weather, etc.

[0124] Then, perform compensation, dimensionality reduction, and normalization processing on the rough data of different modes after enhancing generalization through principal component analysis (PCA) and wavelet modulus maxima, and extract key feature vectors to generate an extended sample library; for example, the internal relationship between temperature, humidity, and human flow.

[0125] Step 3: Train the mathematical model;

[0126] Step 4: Test the mathematical model.

[0127] In some embodiments, the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 all complete the data exchange without master information between the corresponding communication modules through the corresponding communication modules and the corresponding front-end sensors 1 in the IoT network 2 of the mesh data stream in the duplex channel.

[0128] Or, data management of the cloud server 3 is implemented through the edge computing gateway 5 at the aggregation layer;

[0129] The IoT network 2 of the mesh data stream relies on the edge computing gateway 5 at the aggregation layer to complete the communication of the control management system of the cloud server 3 through the cellular / WI-FI network using the MQTT / HTTP protocol;

[0130] The data collected by the front-end sensor 1 includes parameters such as temperature and humidity, energy consumption, and equipment health status, specifically the temperature and humidity, energy consumption, and the health status of the heating equipment 41, ventilation equipment 42, air-conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45.

[0131] The present invention relies on the IoT network of the mesh data stream, performs extended derivation for small-sample data, and then combines SVM classifier identification to realize the perception and identification of HVAC data, and realizes the optimal energy-saving behavior management through GA multi-objective optimization.

[0132] In some embodiments, in step 1, the data collected by the front-end sensor 1 includes the following four types of data:

[0133] The first type of data is the cooling data, heating data, dehumidification data, ventilation data, sleep data, and energy-saving data of the heating equipment 41, ventilation equipment 42, air-conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 under normal energy consumption modes;

[0134] The second type of data is the temperature difference data, airtightness data, long-term operation data, and frequent start-stop data of the heating equipment 41, ventilation equipment 42, air-conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 under abnormal energy consumption modes;

[0135] The third type of data is the strong mode data, initial start-up data, electric auxiliary heating data, and extreme environment operation data of the heating equipment 41, ventilation equipment 42, air-conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 under peak energy consumption modes;

[0136] The fourth type of data is the compressor data, refrigerant data, filter screen data, control system data, circuit data, and fan data of the heating equipment 41, ventilation equipment 42, air-conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 under equipment failure energy consumption modes;

[0137] In this embodiment, the first type of data is sufficient data;

[0138] The second type of data, the third type of data, and the fourth type of data are all small amounts of data;

[0139] The method for forming the initial acquisition sample library is: for the data collected by the front-end sensor 1, according to the cooling mode, heating mode, dehumidification mode, ventilation mode, sleep mode, and energy-saving mode under normal energy consumption modes,

[0140] The over-temperature difference mode, poor room sealing mode, long-term operation mode, and frequent on / off mode under the abnormal energy consumption mode

[0141] The strong mode, initial startup mode, electric auxiliary heating mode, and extreme environment operation mode under the peak energy consumption mode

[0142] The compressor failure mode, refrigerant shortage or leakage mode, filter screen blockage mode, control system failure mode, circuit problem mode, fan failure mode, and long-term lack of maintenance under the equipment failure energy consumption mode

[0143] It is divided into corresponding target information parameters and task attribute parameters;

[0144] Among them,

[0145] In the cooling mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 under the normal energy consumption mode, the corresponding target information parameters include the indoor temperature dropping to the set value, the compressor power ratio, and the power consumption, and the corresponding task attribute parameters include the temperature set value, the compressor start / stop logic, and the refrigerant flow control;

[0146] In the heating mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 under the normal energy consumption mode, the corresponding target information parameters include the indoor temperature rising to the set value, the heat pump efficiency, and the electric auxiliary heating enable rate, and the corresponding task attribute parameters include the temperature set value, the heat pump / electric auxiliary switching threshold, and the outside temperature acquisition parameter;

[0147] In the dehumidification mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 under the normal energy consumption mode, the corresponding target information parameters include the indoor humidity dropping to the target value, the compressor on time, and the fan speed, and the corresponding task attribute parameters include the humidity set value, the compressor operation time, and the fan speed regulation;

[0148] In the ventilation mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 under the normal energy consumption mode, the corresponding target information parameters include the air circulation rate, the fan power ratio, and the environmental ventilation frequency, and the corresponding task attribute parameters include the air volume control target, the filter screen resistance detection, and the air circulation time;

[0149] In the sleep mode of the normal energy consumption mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45, the corresponding target information parameters include the temperature change curve, night power optimization ratio and user comfort score, and the corresponding task attribute parameters include the temperature control gradual change curve, night target power and compressor dormancy frequency;

[0150] In the energy-saving mode of the normal energy consumption mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45, the corresponding target information parameters include the system energy-saving rate, target temperature control deviation range and compressor start-stop frequency, and the corresponding task attribute parameters include the energy-saving optimization algorithm, power limit parameter and load balancing coefficient;

[0151] In the excessive temperature difference mode of the abnormal energy consumption mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45, the corresponding target information parameters include the indoor-outdoor temperature difference, compressor load ratio and unnecessary energy consumption growth ratio, and the corresponding task attribute parameters include the outdoor temperature acquisition frequency, compressor high-frequency operation threshold and alarm parameter;

[0152] In the poor room sealing mode of the abnormal energy consumption mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45, the corresponding target information parameters include the energy loss rate, cold / hot air leakage point analysis and temperature control delay time, and the corresponding task attribute parameters include the cold / hot air loss point analysis task and environmental heat insulation material evaluation parameter;

[0153] In the long-time operation mode of the abnormal energy consumption mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45, the corresponding target information parameters include the cumulative operation time, continuous operation time of the compressor and energy consumption per unit time, and the corresponding task attribute parameters include the timing function trigger task and equipment continuous high-load monitoring threshold;

[0154] In the frequent start-stop mode of the abnormal energy consumption mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45, the corresponding target information parameters include the start-stop frequency, peak compressor start-up power and cyclic energy consumption growth ratio, and the corresponding task attribute parameters include the user behavior record and minimum switch interval time parameter;

[0155] In the strong mode of the peak energy consumption mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45, the corresponding target information parameters include the system full-load operation time, target temperature achievement time and power peak, and the corresponding task attribute parameters include the full-load start time limit, target temperature control accuracy and power upper limit parameter;

[0156] In the initial startup mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 in the peak energy consumption mode, the corresponding target information parameters include the compressor current at startup, the target temperature control response time, and the refrigerant pressure value, and the corresponding task attribute parameters include the startup current limit, the refrigerant pressure monitoring task, and the compressor overload protection;

[0157] In the electric auxiliary heating mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 in the peak energy consumption mode, the corresponding target information parameters include the duration of electric auxiliary heater activation, the ambient temperature, and the target temperature maintenance time, and the corresponding task attribute parameters include the electric auxiliary power adjustment parameter, the minimum activation temperature threshold, and the electric heating efficiency evaluation;

[0158] In the extreme environment operation mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 in the peak energy consumption mode, the corresponding target information parameters include the deviation of the outside temperature, the system stability, and the compressor load factor, and the corresponding task attribute parameters include the outside environment perception parameter, the compressor heat dissipation optimization, and the operation stability index;

[0159] In the compressor failure mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 in the equipment failure energy consumption mode, the corresponding target information parameters include the compressor power deviation, the abnormal refrigerant flow rate, and the decrease ratio of the start-stop efficiency, and the corresponding task attribute parameters include the compressor operating frequency, the abnormal vibration alarm, and the efficiency deviation threshold;

[0160] In the refrigerant shortage or leakage mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 in the equipment failure energy consumption mode, the corresponding target information parameters include the remaining amount of refrigerant, the decrease ratio of the refrigeration efficiency, and the environmental impact coefficient, and the corresponding task attribute parameters include the refrigerant pressure sensor data, the automatic replenishment task, and the leakage detection algorithm;

[0161] In the filter clogging mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44, and auxiliary equipment 45 in the equipment failure energy consumption mode, the corresponding target information parameters include the air volume obstruction ratio, the filter cleanliness, and the system heat dissipation delay rate, and the corresponding task attribute parameters include the air flow resistance analysis, the regular cleaning reminder, and the air volume measurement task;

[0162] In the control system failure mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45 in the equipment failure energy consumption mode, the corresponding target information parameters include the temperature control deviation range, system response delay time and abnormal alarm count, and the corresponding task attribute parameters include temperature control logic test, sensor calibration task and data anomaly record;

[0163] In the circuit problem mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45 in the equipment failure energy consumption mode, the corresponding target information parameters include current fluctuation amplitude, overload rate and critical component aging degree, and the corresponding task attribute parameters include circuit current stability, overload monitoring algorithm and component aging life assessment;

[0164] In the fan failure mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45 in the equipment failure energy consumption mode, the corresponding target information parameters include wind speed deviation, circulation efficiency decline ratio and abnormal noise index, and the corresponding task attribute parameters include fan speed deviation detection, abnormal noise data collection and component wear assessment;

[0165] In the long-term unmaintained mode of the heating equipment 41, ventilation equipment 42, air conditioning equipment 43, supporting equipment 44 and auxiliary equipment 45 in the equipment failure energy consumption mode, the corresponding target information parameters include dust accumulation coverage rate, heat transfer efficiency and operating power consumption increase ratio, and the corresponding task attribute parameters include maintenance task plan, component aging coefficient and regular cleaning task trigger. As shown in the following table:

[0166]

[0167]

[0168] In some embodiments, the heating equipment 41 includes a heat source device that provides heat, a heat dissipation device that transfers heat to the room, and heating auxiliary equipment that supports hot water circulation and control;

[0169] The heat source device includes a boiler, a heat pump and / or an electric heater;

[0170] The heat dissipation device includes a radiator, underfloor heating and / or fan coil units;

[0171] The heating auxiliary equipment includes a water pump and / or a heating control system;

[0172] The ventilation equipment 42 includes natural ventilation that utilizes natural wind to circulate air through windows and ventilation openings, mechanical ventilation for forced ventilation and air exchange, and filtration equipment for purifying air;

[0173] The mechanical ventilation includes exhaust fans, fresh air systems and / or smoke exhaust equipment;

[0174] The filtering device includes an air filter and / or an electrostatic precipitator;

[0175] The air conditioning device 43 includes a cold source device providing a cooling function, a terminal device for regulating indoor air, and a cooling device for cooling the cold source device;

[0176] The cold source device includes a chiller and / or an absorption chiller;

[0177] The terminal device includes a fan coil unit and / or an air handling unit;

[0178] The cooling device includes a cooling tower and / or a condenser;

[0179] The supporting device 44 includes a pipe network system of a pipe network for cold and hot water and air transportation, and a supporting control system for automatic control;

[0180] The supporting control system includes a thermostat and / or an intelligent building management system;

[0181] The auxiliary device 45 includes a humidifying device for increasing air humidity, a dehumidifying device for reducing air humidity, a water treatment device for improving water quality, and a safety device for ensuring safety;

[0182] The humidifying device includes a humidifier and / or a spray humidifier;

[0183] The dehumidifying device includes a condensing dehumidifier and / or a rotary dehumidifier;

[0184] The water treatment device includes a water softening device and / or a descaling device;

[0185] The safety device includes a fire damper, an air volume regulating valve and / or an expansion tank.

[0186] In some embodiments, in step 3, based on the generative adversarial network WGAN, the energy consumption sparse data in the abnormal energy consumption mode, the peak energy consumption mode and the equipment failure energy consumption mode is used to extract feature vectors through the middle layer of the discriminator to achieve data enhancement, while avoiding overfitting of small sample events;

[0187] Data enhancement is performed by the above means.

[0188] The classification model SVM quickly classifies and detects the states of the heating device 41, the ventilation device 42, the air conditioning device 43, the supporting device 44 and the auxiliary device 45 in the normal energy consumption mode, the abnormal energy consumption mode, the peak energy consumption mode and the equipment failure energy consumption mode;

[0189] Normal and abnormal are classified and trained together by the above means.

[0190] Among them, the classification model SVM is trained and tested on the target information parameters and task attribute parameters in the extended sample library according to the enhanced generalization derivation rules of WGAN;

[0191] Select the classification model SVM that meets the accuracy requirements for real-time identification of target signal parameters;

[0192] Complete model training through the above means.

[0193] Search for a specific rule sample library that meets specific rules in the extended sample library according to the indexing rules and output the corresponding task attribute parameters that meet the conditions;

[0194] Complete model testing through the above means.

[0195] Based on the WGAN-SVM model, generate the combination of device switch status and temperature for the initial population of real-time target information parameters and task attribute parameters. Through the utility evaluation of energy consumption, operating cost, anomaly detection, and comfort, combined with the crossover and mutation of GA, optimize the control strategy with the highest fitness solution.

[0196] Optimize comfort for real-time multi-objectives through genetic algorithms.

[0197] In some embodiments, the multi-objective optimization NSGA-II function is designed as follows: Design the preprocessing anomaly detection model, specifically as follows:

[0198] Minimize:

[0199] Among them, |W k (f)| is the modulus maximum value of the collected signal f at the wavelet scale k;

[0200] K represents the number of wavelet transform scales;

[0201] The enhanced generalization model includes:

[0202] Generator:

[0203] Minimize:

[0204] Refers to the evaluation of generating samples for the environmental data predicted by the generator of;

[0205] P gen represents the environmental data predicted by the generator, including indoor temperature and humidity;

[0206] Discriminator:

[0207] Minimize:

[0208] It refers to the evaluation of the true sample X of the real indoor environment data collected by the air conditioning system;

[0209] P real represents the real indoor environment data collected by the air conditioning system, including temperature and humidity;

[0210] The classification model is designed as:

[0211] Minimize:

[0212] where ε is the weight vector of the classification model hyperplane;

[0213] ξ n is the slack variable;

[0214] C is the penalty coefficient;

[0215] n represents the air conditioning operation data sample at a certain time point, including room temperature, humidity, energy consumption, current;

[0216] N represents the total number of samples in the entire training data set;

[0217] The energy consumption model is designed as:

[0218] Minimize:

[0219] where P (t) is the power consumption of the HVAC system at time t; T is the time period;

[0220] The comfort model is designed as:

[0221] Minimize:

[0222] where B is the total number of samples involved in comfort calculation;

[0223] Q i is the actual temperature of the ith space; Q set is the set temperature;

[0224] The cost model is designed as:

[0225]

[0226] where C cost (t) is the electricity cost at time t;

[0227] The multi-objective optimization model is designed as:

[0228] Minimize: J total = θJ1 + μJ5 + πJ6 + ρJ7;

[0229] Among them, θ, μ, π, and ρ are the weight values for balancing energy consumption, operating cost, anomaly detection, and comfort evaluation, respectively.

[0230] In some embodiments, the interaction of the mathematical model is as follows: The IoT network 2 of the mesh data stream collects the environment through the front-end sensor 1 and the corresponding communication module, and the device operation data of the corresponding heating device 41, the corresponding ventilation device 42, the corresponding air-conditioning device 43, the corresponding supporting device 44, and / or the corresponding auxiliary device 45. Then, combined with the real-time data and the abnormal missing IoT data, it is complemented and predicted by the WGAN data.

[0231] Among them, the abnormal data extracts the fault feature vector through principal component analysis and wavelet modulus maxima processing, trains and tests the classification model with the comprehensive feature vector, and performs pattern recognition and behavior prediction through the optimal model.

[0232] The abnormal data includes compressor signals and / or power curves.

[0233] For the sample library of the target information parameters and task attribute parameters under the combined rule conditions, when the real-time target information parameters are input, the task attribute parameters corresponding to the rules are obtained through the SVM classification model index. Then, combined with the sample library of the target information parameters and task attribute parameters under specific rule conditions, through the crossover and mutation of the GA, the multi-objective optimization of the energy consumption, operating cost, anomaly detection, and comfort objective functions, and the behavior management sample with the optimization strategy are used to iteratively update the initial acquisition sample library.

[0234] In some embodiments, in step 4, more than one target information parameter is input into the trained mathematical model, and the mathematical model generates a specific sample library and obtains the corresponding task attribute parameters.

[0235] In some embodiments, the initial acquisition sample library generates the combination of device switch states and temperatures based on the initial population of the real-time target information parameters and task attribute parameters of the mathematical model. After the utility evaluation of energy consumption, operating cost, anomaly detection, and comfort, combined with the crossover and mutation of the GA, the control strategy is optimized by the solution with the highest fitness.

[0236] The preferred specific embodiments of the present invention have been described in detail above. It should be understood that those of ordinary skill in the art can make many modifications and variations according to the concept of the present invention without creative labor. Therefore, all technical solutions that can be obtained by those skilled in the art in the technical field of the present invention based on the concept of the present invention through logical analysis, reasoning, or limited experiments on the basis of the prior art should be within the protection scope determined by the claims.

Claims

1. Poor and extensive pre-perceived IoT multi-objective optimized HVAC system, characterized in that, Obtained by performing the following steps: Step 1: Collect data through the pre - sensors (1) set in the heating equipment (41), ventilation equipment (42), air - conditioning equipment (43), supporting equipment (44) and auxiliary equipment (45), and form an initial acquisition sample library; Step 2: Pre - process the initial acquisition sample library to handle the situation of a small amount of data due to insufficient data collected by the pre - sensor (1). Then, perform complementary - difference dimensionality reduction and normalization processing on the rough data of different patterns after enhanced generalization through principal component analysis (PCA) and wavelet modulus maxima, and extract key feature vectors to generate an extended sample library; Step 3: Train a mathematical model; Step 4: Test the mathematical model.

2. The lean pre-perception IoT multi-objective optimized HVAC system according to claim 1, characterized in that The heating equipment (41), the ventilation equipment (42), the air - conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45) all complete the masterless information exchange of data between the corresponding communication modules through the corresponding communication modules and the corresponding pre - sensors (1) in the Internet of Things network (2) of the mesh data stream in the duplex channel, or realize the data management of the cloud server (3) through the aggregation - layer edge computing gateway (5); The Internet of Things network (2) of the mesh data stream relies on the aggregation - layer edge computing gateway (5) to complete the communication of the control management system of the cloud server (3) through the cellular / WI - FI network using the MQTT / HTTP protocol; The data collected by the pre - sensor (1) includes parameters such as temperature and humidity, energy consumption, and equipment health status, specifically temperature and humidity, energy consumption, and the health status of the heating equipment (41), the ventilation equipment (42), the air - conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45).

3. The poverty-prevention pre-perception IoT multi-objective optimized HVAC system according to claim 1, characterized in that In step 1, the data collected by the pre - sensor (1) includes the following four types of data: The first type of data is the cooling data, heating data, dehumidification data, ventilation data, sleep data and energy - saving data in the normal energy - consumption mode of the heating equipment (41), the ventilation equipment (42), the air - conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45); The second type of data is the temperature difference data, airtightness data, long - time operation data, and frequent start - up and shutdown data in the abnormal energy - consumption mode of the heating equipment (41), the ventilation equipment (42), the air - conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45); The third type of data is the strong - mode data, initial start - up data, electric auxiliary heating data, and extreme - environment operation data in the peak energy - consumption mode of the heating equipment (41), the ventilation equipment (42), the air - conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45); The fourth type of data is the compressor data, refrigerant data, filter - screen data, control - system data, circuit data, and fan data in the equipment - failure energy - consumption mode of the heating equipment (41), the ventilation equipment (42), the air - conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45). In this embodiment, the first type of data is sufficient data; The second type of data, the third type of data, and the fourth type of data are all the small amount of data; The method for forming the initial acquisition sample library is as follows: The data collected by the pre-sensor (1) is classified according to the refrigeration mode, heating mode, dehumidification mode, ventilation mode, sleep mode, and energy-saving mode under the normal energy consumption mode, The over-temperature difference mode, poor room sealing mode, long-term operation mode, and frequent start-stop mode under the abnormal energy consumption mode, The strong mode, initial start mode, electric auxiliary heating mode, and extreme environment operation mode under the peak energy consumption mode, The compressor failure mode, refrigerant shortage or leakage mode, filter screen blockage mode, control system failure mode, circuit problem mode, fan failure mode, and long-term non-maintenance under the equipment failure energy consumption mode, into corresponding target information parameters and task attribute parameters; Among them, In the refrigeration mode under the normal energy consumption mode of the heating device (41), the ventilation device (42), the air-conditioning device (43), the supporting device (44), and the auxiliary device (45), the corresponding target information parameters include the indoor temperature dropping to the set value, the compressor power ratio, and the power consumption, and the corresponding task attribute parameters include the temperature set value, the compressor start-stop logic, and the refrigerant flow control; In the heating mode under the normal energy consumption mode of the heating device (41), the ventilation device (42), the air-conditioning device (43), the supporting device (44), and the auxiliary device (45), the corresponding target information parameters include the indoor temperature rising to the set value, the heat pump efficiency, and the electric auxiliary heating enable rate, and the corresponding task attribute parameters include the temperature set value, the heat pump / electric auxiliary switching threshold, and the outside temperature acquisition parameter; In the dehumidification mode under the normal energy consumption mode of the heating device (41), the ventilation device (42), the air-conditioning device (43), the supporting device (44), and the auxiliary device (45), the corresponding target information parameters include the indoor humidity dropping to the target value, the compressor on time, and the fan speed, and the corresponding task attribute parameters include the humidity set value, the compressor operation time, and the fan speed adjustment; In the ventilation mode under the normal energy consumption mode of the heating device (41), the ventilation device (42), the air-conditioning device (43), the supporting device (44), and the auxiliary device (45), the corresponding target information parameters include the air circulation rate, the fan power ratio, and the environmental air change frequency, and the corresponding task attribute parameters include the air volume control target, the filter screen resistance detection, and the air circulation time; In the sleep mode of the normal energy consumption mode, for the heating device (41), the ventilation device (42), the air conditioning device (43), the supporting device (44) and the auxiliary device (45), the corresponding target information parameters include the temperature change curve, the night power optimization ratio and the user comfort score, and the corresponding task attribute parameters include the temperature control gradual change curve, the night target power and the compressor dormancy frequency; In the energy saving mode of the normal energy consumption mode, for the heating device (41), the ventilation device (42), the air conditioning device (43), the supporting device (44) and the auxiliary device (45), the corresponding target information parameters include the system energy saving rate, the target temperature control deviation range and the compressor start-stop frequency, and the corresponding task attribute parameters include the energy saving optimization algorithm, the power limit parameter and the load balancing coefficient; In the large temperature difference mode of the abnormal energy consumption mode, for the heating device (41), the ventilation device (42), the air conditioning device (43), the supporting device (44) and the auxiliary device (45), the corresponding target information parameters include the indoor-outdoor temperature difference, the compressor load ratio and the unnecessary energy consumption growth ratio, and the corresponding task attribute parameters include the outdoor temperature acquisition frequency, the compressor high-frequency operation threshold and the alarm parameter; In the poor room airtightness mode of the abnormal energy consumption mode, for the heating device (41), the ventilation device (42), the air conditioning device (43), the supporting device (44) and the auxiliary device (45), the corresponding target information parameters include the energy loss rate, the cold / hot air leakage point analysis and the temperature control delay time, and the corresponding task attribute parameters include the cold / hot air loss point analysis task and the environmental heat insulation material evaluation parameter; In the long-time operation mode of the abnormal energy consumption mode, for the heating device (41), the ventilation device (42), the air conditioning device (43), the supporting device (44) and the auxiliary device (45), the corresponding target information parameters include the accumulated operation time, the continuous operation time of the compressor and the energy consumption per unit time, and the corresponding task attribute parameters include the timing function trigger task and the device continuous high-load monitoring threshold; In the frequent start-stop mode of the abnormal energy consumption mode, for the heating device (41), the ventilation device (42), the air conditioning device (43), the supporting device (44) and the auxiliary device (45), the corresponding target information parameters include the start-stop frequency, the peak compressor start power and the cyclic energy consumption growth ratio, and the corresponding task attribute parameters include the user behavior record and the minimum switch interval time parameter; In the strong mode of the peak energy consumption mode, for the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45), the corresponding target information parameters include the system full load operation time, the target temperature achievement time and the power peak value, and the corresponding task attribute parameters include the full load start time limit, the target temperature control accuracy and the power upper limit parameter; In the initial start mode of the peak energy consumption mode, for the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45), the corresponding target information parameters include the compressor current at startup, the target temperature control response time and the refrigerant pressure value, and the corresponding task attribute parameters include the startup current limit, the refrigerant pressure monitoring task and the compressor overload protection; In the electric auxiliary heating mode of the peak energy consumption mode, for the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45), the corresponding target information parameters include the electric auxiliary heater enabling duration, the ambient temperature and the target temperature maintenance time, and the corresponding task attribute parameters include the electric auxiliary power adjustment parameter, the enabling minimum temperature threshold and the electric heating efficiency evaluation; In the extreme environment operation mode of the peak energy consumption mode, for the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45), the corresponding target information parameters include the outside temperature deviation degree, the system stability and the compressor load factor, and the corresponding task attribute parameters include the outside environment perception parameter, the compressor heat dissipation optimization and the operation stability index; In the compressor failure mode of the equipment failure energy consumption mode, for the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45), the corresponding target information parameters include the compressor power deviation, the refrigerant flow abnormality and the start-stop efficiency decrease ratio, and the corresponding task attribute parameters include the compressor operating frequency, the abnormal vibration alarm and the efficiency deviation threshold; In the refrigerant shortage or leakage mode of the equipment failure energy consumption mode, for the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45), the corresponding target information parameters include the remaining refrigerant amount, the refrigeration efficiency decrease ratio and the environmental impact coefficient, and the corresponding task attribute parameters include the refrigerant pressure sensor data, the automatic replenishment task and the leakage detection algorithm; In the filter clogging mode of the equipment failure energy consumption mode of the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44), and the auxiliary equipment (45), the corresponding target information parameters include the air volume obstruction ratio, the filter cleanliness, and the system heat dissipation delay rate, and the corresponding task attribute parameters include the air flow resistance analysis, the regular cleaning reminder, and the air volume measurement task; In the control system failure mode of the equipment failure energy consumption mode of the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44), and the auxiliary equipment (45), the corresponding target information parameters include the temperature control deviation range, the system response delay time, and the number of abnormal alarms, and the corresponding task attribute parameters include the temperature control logic test, the sensor calibration task, and the data abnormality record; In the circuit problem mode of the equipment failure energy consumption mode of the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44), and the auxiliary equipment (45), the corresponding target information parameters include the current fluctuation amplitude, the overload rate, and the aging degree of key components, and the corresponding task attribute parameters include the circuit current stability, the overload monitoring algorithm, and the component aging life assessment; In the fan failure mode of the equipment failure energy consumption mode of the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44), and the auxiliary equipment (45), the corresponding target information parameters include the wind speed deviation, the circulation efficiency decline ratio, and the abnormal noise index, and the corresponding task attribute parameters include the fan speed deviation detection, the abnormal noise data collection, and the component wear assessment; In the long-term unmaintained mode of the equipment failure energy consumption mode of the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44), and the auxiliary equipment (45), the corresponding target information parameters include the dust accumulation coverage rate, the heat exchange efficiency, and the operating power consumption increase ratio, and the corresponding task attribute parameters include the maintenance task plan, the component aging coefficient, and the regular cleaning task trigger.

4. The lean pre-perception IoT multi-objective optimized HVAC system according to claim 1, characterized in that, The heating equipment (41) includes a heat source device that provides heat, a heat dissipation device that transfers heat to the indoor space, and a heating auxiliary device that supports hot water circulation and control; The heat source device includes a boiler, a heat pump, and / or an electric heater; The heat dissipation device includes a radiator, underfloor heating, and / or fan coil units; The heating auxiliary device includes a water pump and / or a heating control system; The ventilation equipment (42) includes natural ventilation that uses natural wind to circulate air through windows and vents, mechanical ventilation for forced ventilation and air change, and a filtering device for purifying air; The mechanical ventilation includes an exhaust fan, a fresh air system, and / or a smoke exhaust device; The filtering device includes an air filter and / or an electrostatic dust removal device; The air conditioning equipment (43) includes a cold source device that provides a cooling function, a terminal device that adjusts indoor air, and a cooling device that cools the cold source device; The cold source device includes a chiller and / or an absorption chiller; The terminal device includes a fan coil unit and / or an air handling unit; The cooling device includes a cooling tower and / or a condenser; The supporting equipment (44) includes a pipe network system of a cold and hot water and air transportation pipeline network, and a supporting control system with automatic control; The supporting control system includes a thermostat and / or an intelligent building management system; The auxiliary equipment (45) includes a humidifying device that increases air humidity, a dehumidifying device that reduces air humidity, a water treatment device that improves water quality, and a safety device that ensures safety; The humidifying device includes a humidifier and / or a spray humidifier; The dehumidifying device includes a condensing dehumidifier and / or a rotary dehumidifier; The water treatment device includes a water softening device and / or a descaling device; The safety device includes a fire damper, an air volume regulating valve and / or an expansion tank.

5. The poverty-prevention pre-perception IoT multi-objective optimized HVAC system according to claim 1, characterized in that In step 3, based on the generative adversarial network WGAN, the energy consumption sparse data in the abnormal energy consumption mode, peak energy consumption mode and equipment failure energy consumption mode is used to extract feature vectors through the middle layer of the discriminator to achieve data enhancement, while avoiding overfitting of small sample events; The classification model SVM quickly classifies and detects the states of the heating equipment (41), the ventilation equipment (42), the air conditioning equipment (43), the supporting equipment (44) and the auxiliary equipment (45) in the normal energy consumption mode, abnormal energy consumption mode, peak energy consumption mode and equipment failure energy consumption mode; Among them, the classification model SVM trains and tests the target information parameters and task attribute parameters in the extended sample library according to the enhanced WGAN generalization derivation rules; Select the classification model SVM that meets the accuracy requirements for real-time target signal parameter identification; Search for a specific rule sample library that meets specific rules in the extended sample library according to the indexing rules and output the corresponding task attribute parameters that meet the conditions; Based on the real-time target information parameters and task attribute parameters of the WGAN-SVM model, an initial population of equipment switch states and temperatures is generated. Through the practicality evaluation of energy consumption, operating cost, anomaly detection and comfort, combined with the crossover and mutation of GA, the control strategy is optimized by the solution with the highest fitness.

6. The sparse pre-perception IoT multi-objective optimized HVAC system according to claim 5, characterized in that, The multi-objective optimization NSGA-II function is designed as follows: the preprocessing anomaly detection model is designed as follows: Minimize: where, |W k (f)| is the modulus maximum of the acquired signal f at the wavelet scale k; K represents the scale number of wavelet transform; The enhanced generalization model includes: Generator: Minimize: It refers to the evaluation of generating a sample X~ for the environmental data predicted by the generator; P gen represents the environmental data predicted by the generator, including indoor temperature and humidity; Discriminator: Minimize: It refers to the evaluation of the true sample X of the true indoor environment data collected by the air conditioning system; P real Represents the real indoor environmental data collected by the air conditioning system, including temperature and humidity; The classification model is designed as: Minimize: Among them, ε is the weight vector of the classification model hyperplane; ξ n is a slack variable; C is the penalty coefficient; n represents the air conditioning operation data sample at a certain time point, including room temperature, humidity, energy consumption, current, etc.; N represents the total number of samples in the entire training data set; The energy consumption model is designed as: Minimize: Among them, P (t) is the power consumption of the HVAC system at time t; T is the time period; The comfort model is designed as: Minimize: Among them, B is the total number of samples involved in comfort calculation; Q i is the actual temperature of the i-th space; Q set is the set temperature; The cost model is designed as: Among them, C cost (t) is the electricity consumption at time t; The multi-objective optimization model is designed as: Minimize: J total = θJ1 + μJ5 + πJ6 + ρJ7; Among them, θ, μ, π and ρ are the weight values for balancing energy consumption, operating cost, anomaly detection and comfort evaluation respectively.

7. The sparse pre-perception IoT multi-objective optimized HVAC system according to claim 1, wherein The interaction of the mathematical model lies in that the Internet of Things network (2) of the mesh data stream collects the environment through the preposed sensors (1) and the corresponding communication modules, and the device operation data of the corresponding heating equipment (41), the corresponding ventilation equipment (42), the corresponding air conditioning equipment (43), the corresponding supporting equipment (44) and / or the corresponding auxiliary equipment (45), and then combines real-time data and abnormal missing IoT data, and performs data completion and prediction through WGAN; Among them, the abnormal data extracts the fault feature vector through principal component analysis and wavelet modulus maximum processing, trains and tests the classification model with the comprehensive feature vector, and performs pattern recognition and behavior prediction through the optimal model; The abnormal data includes compressor signals and / or power curves; The sample library of target information parameters and task attribute parameters under the combined rule conditions, when the real-time target information parameters are input, obtains the task attribute parameters of the corresponding rule through the SVM classification model index, and then combines the sample library of target information parameters and task attribute parameters under the specific rule conditions, and through the crossover and mutation of GA, multi-objectively optimizes the energy consumption, operation cost, abnormal detection and comfort objective functions, and the behavior management sample with the optimization strategy, and iteratively updates the initial acquisition sample library.

8. The poverty-prevention pre-perception IoT multi-objective optimized HVAC system according to claim 1, characterized in that In step 4, input one or more target information parameters to the trained mathematical model, and the mathematical model generates a specific sample library and obtains the corresponding task attribute parameters.

9. The poverty-prevalent pre-perception IoT multi-objective optimized HVAC system according to claim 1, characterized in that, The initial acquisition sample library generates a combination of device switch states and temperatures based on the initial population of real-time target information parameters and task attribute parameters of the mathematical model, evaluates the practicality of energy consumption, operation cost, abnormal detection and comfort, and combines the crossover and mutation of GA, and optimizes the control strategy with the scheme with the highest fitness.

Images