CN114484731A - Method and device for diagnosing faults of central air conditioner based on stacking fusion algorithm - Google Patents

Method and device for diagnosing faults of central air conditioner based on stacking fusion algorithm Download PDF

Info

Publication number
CN114484731A
CN114484731A CN202111614025.7A CN202111614025A CN114484731A CN 114484731 A CN114484731 A CN 114484731A CN 202111614025 A CN202111614025 A CN 202111614025A CN 114484731 A CN114484731 A CN 114484731A
Authority
CN
China
Prior art keywords
central air
model
data set
conditioning system
training
Prior art date
Legal status (The legal status 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 status listed.)
Granted
Application number
CN202111614025.7A
Other languages
Chinese (zh)
Other versions
CN114484731B (en
Inventor
赵琼
穆佩红
裘天阅
谢金芳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Yingji Power Technology Co ltd
Original Assignee
Zhejiang Yingji Power Technology Co ltd
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 Zhejiang Yingji Power Technology Co ltd filed Critical Zhejiang Yingji Power Technology Co ltd
Priority to CN202111614025.7A priority Critical patent/CN114484731B/en
Publication of CN114484731A publication Critical patent/CN114484731A/en
Application granted granted Critical
Publication of CN114484731B publication Critical patent/CN114484731B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • F24F11/30—Control or safety arrangements for purposes related to the operation of the system, e.g. for safety or monitoring
    • F24F11/32—Responding to malfunctions or emergencies
    • F24F11/38—Failure diagnosis
    • 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
    • F24F11/50—Control or safety arrangements characterised by user interfaces or communication
    • F24F11/54—Control or safety arrangements characterised by user interfaces or communication using one central controller connected to several sub-controllers
    • 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
    • F24F11/62—Control or safety arrangements characterised by the type of control or by internal processing, e.g. using fuzzy logic, adaptive control or estimation of values
    • F24F11/63—Electronic processing
    • F24F11/64—Electronic processing using pre-stored data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G06F18/253—Fusion techniques of extracted features
    • 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
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F2140/00—Control inputs relating to system states
    • F24F2140/10—Pressure
    • F24F2140/12—Heat-exchange fluid pressure
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F2140/00—Control inputs relating to system states
    • F24F2140/20—Heat-exchange fluid temperature
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/06—Power analysis or power optimisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/08—Thermal analysis or thermal optimisation
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
    • Y02B30/00—Energy efficient heating, ventilation or air conditioning [HVAC]
    • Y02B30/70—Efficient control or regulation technologies, e.g. for control of refrigerant flow, motor or heating

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Chemical & Material Sciences (AREA)
  • Mechanical Engineering (AREA)
  • Combustion & Propulsion (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Software Systems (AREA)
  • Probability & Statistics with Applications (AREA)
  • Signal Processing (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Biomedical Technology (AREA)
  • Operations Research (AREA)
  • Health & Medical Sciences (AREA)
  • Algebra (AREA)
  • Databases & Information Systems (AREA)
  • Fuzzy Systems (AREA)
  • Human Computer Interaction (AREA)
  • Medical Informatics (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Air Conditioning Control Device (AREA)

Abstract

The invention discloses a method for diagnosing faults of a central air conditioner based on a stacking fusion algorithm, which comprises the following steps: establishing a digital twin model of a central air-conditioning system; acquiring state data of a central air-conditioning system in normal operation and different faults, and performing data preprocessing and feature extraction to obtain a sample data set; dividing a sample data set into a training data set and a testing data set, and simultaneously building a double-layer stacking model; training each base learner by adopting a k-fold cross validation method, and obtaining a prediction result of each base learner as a secondary training data set; when training each base learner, selecting a plurality of groups of different machine learning algorithms for combination to generate a plurality of groups of secondary training data sets under a combination mode; inputting a plurality of groups of secondary training data sets into a secondary learner for training to obtain a plurality of central air conditioner fault diagnosis models; and evaluating the prediction performance of the plurality of central air conditioner fault diagnosis models, and selecting the model with the best performance as the optimal central air conditioner fault diagnosis model for fault diagnosis.

Description

Central air conditioner fault diagnosis method and device based on stacking fusion algorithm
Technical Field
The invention belongs to the technical field of central air conditioners, and particularly relates to a method and a device for diagnosing faults of a central air conditioner based on a stacking fusion algorithm.
Background
In the development process of urbanization, with the enlargement of urban scale, the number and scale of large public buildings are remarkably increased, wherein two major trends of rapidly developed building electrification and building intellectualization which is gradually rising are widely concerned by all communities. Due to the refrigeration requirement of large public buildings, the scale of the central air-conditioning system and the automatic control system thereof is increasingly large, and the variety and the number of the equipment are increasingly diversified, so that the complexity of the system is increasingly high. In the running process of the system, various faults inevitably occur, and if the faults cannot be eliminated in time, the running parameters of the system are inevitably deviated from the required set values, so that discomfort is brought to indoor workers, the working efficiency and the working quality are affected, the energy consumption of the system is increased, and the service life of equipment is shortened. Moreover, once a fault occurs in the central air conditioning system, a long time is often needed to determine the fault occurrence point and complete subsequent maintenance work, and this process causes unnecessary energy waste. According to incomplete statistics, in the process of troubleshooting a central air-conditioning system, the time for searching the cause of the failure generally accounts for over 50% of the total troubleshooting time.
Because the central air-conditioning system has the characteristics of nonlinearity, complexity, changeability, mutual coupling of a plurality of system parameters and the like, establishing an extremely complete and generalized universal central air-conditioning system fault diagnosis method at the present stage is difficult to realize. At present, a fault diagnosis method of a central air conditioner which is usually adopted is mainly based on historical experience data, and a fault diagnosis model is trained by utilizing a neural network. However, when a single learning algorithm is used for predicting the fault diagnosis model, the complexity of the prediction model is improved due to too many input data variables, so that overfitting of prediction output is caused, and the accuracy of model prediction is reduced.
Based on the technical problems, a new central air conditioner fault diagnosis method based on a stacking fusion algorithm needs to be designed.
Disclosure of Invention
The invention aims to provide a method for diagnosing faults of a central air conditioner based on a stacking fusion algorithm.
In order to solve the technical problem, the invention provides a method for diagnosing the fault of a central air-conditioning system based on a stacking fusion algorithm, which comprises the following steps:
s1, establishing a digital twin model of the central air-conditioning system by adopting a mechanism modeling and data identification method;
s2, acquiring state data of the central air-conditioning system during normal operation and different faults through a plurality of sensors, preprocessing the state data, extracting features of preprocessed data variables by adopting a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm, and selecting the extracted features according to a grey correlation algorithm to obtain a sample data set;
step S3, dividing the sample data set into a training data set and a testing data set, and simultaneously building a double-layer stacking model, determining the number of the base learners as m and the number of the secondary learners as 1;
step S4, training each base learner by adopting a k-fold cross validation method, and obtaining the prediction result of each base learner as a secondary training data set; when each base learner is trained, selecting a plurality of groups of different machine learning algorithms for combination to generate a plurality of groups of secondary training data sets under a combination mode;
step S5, inputting a plurality of groups of secondary training data sets into a secondary learner for training to obtain a plurality of central air-conditioning fault diagnosis models;
and step S6, evaluating the prediction performance of the plurality of central air-conditioning fault diagnosis models through the test data set, selecting the model with the best prediction performance as the optimal central air-conditioning fault diagnosis model, and diagnosing the faults of the central air-conditioning system through the optimal central air-conditioning fault diagnosis model.
Further, in step S1, establishing a digital twin model of the central air conditioning system by using a mechanism modeling and data identification method specifically includes:
constructing a physical model, a logic model and a simulation model of the central air-conditioning system; wherein the content of the first and second substances,
the construction of the physical model comprises the following steps: at least establishing a physical model of a water chilling unit, a chilled water circulation system and a cooling water circulation system, wherein the water chilling unit provides chilled water with a certain temperature for the tail end and consists of a compressor, an evaporator, a condenser and a throttle valve; the chilled water circulating system is used for transmitting chilled water to a cooling coil pipe to cool indoor return air and consists of a chilled water pump, a chilled water pipe and an air handling unit; the cooling water circulation system releases heat absorbed in a refrigerating fluid of the water chilling unit into the atmosphere and consists of a cooling water pump, a cooling water pipe and a cooling tower;
the construction of the logic model comprises the following steps: establishing a controllable closed-loop logic model according to a logic mechanism relation among all physical entities of the central air-conditioning system, and mapping the physical model to the logic model;
the construction of the simulation model comprises the following steps: building a simulation model of the central air-conditioning system based on the collected operation data, state data and physical attribute data of the central air-conditioning system;
carrying out virtual-real fusion on the physical model, the logic model and the simulation model to construct a system-level digital twin model of a physical entity of the central air-conditioning system in a virtual space;
and accessing multi-working-condition real-time operation data of the central air-conditioning system into the system-level digital twin model, and performing self-adaptive identification correction on the simulation result of the system-level digital twin model by adopting a reverse identification method to obtain the digital twin model of the central air-conditioning system after identification correction.
Further, the establishment of the water chiller model comprises:
neglecting the pressure loss of the suction and exhaust of the compressor and neglecting the heat exchange between the compressor and the environment, establishing a compressor model expressed as:
Figure BDA0003436111540000031
Figure BDA0003436111540000032
Figure BDA0003436111540000033
Figure BDA0003436111540000034
wherein m isrIs the refrigerant mass flow rate; vthThe theoretical gas transmission capacity of the compressor; v. of1The specific volume of the air suction of the compressor; xi is the gas transmission coefficient; pthsThe theoretical power consumption of the compressor in the isentropic compression process is realized; piThe power consumption of the actual compression process of the compressor is the indicated power; pelThe electric power required to be input for the actual compression process of the compressor is the power measured by the power meter; k is an isentropic compression index; peIs the evaporating pressure, i.e. the compressor suction pressure; pkIs the condensing pressure, i.e. the compressor discharge pressure; etaiIndicating efficiency for the compressor; etaelElectrical efficiency of the compressor; h is2Is the enthalpy of the refrigerant at the outlet of the compressor; h is1Is the enthalpy of the compressor inlet refrigerant;
the modeling of the condenser comprises:
neglecting the heat exchange between the condenser and the outside, and considering the flow of the refrigerant and the cooling water as a one-dimensional uniform flow, the heat exchange process in the condenser is obtained as follows:
Qc＝mw,ccp,w(two,c-twi,c)＝mr(hri,c-hro,c)；
Q1,c＝K1,cF1,cΔt1,c；
Q2,c＝K2,cF2,cΔt2,c；
Q3,c＝K3,cF3,cΔt3,c；
Figure BDA0003436111540000035
Figure BDA0003436111540000036
Figure BDA0003436111540000037
wherein Q iscThe total heat exchange capacity of the condenser; m isw,cIs the cooling water flow rate; c. Cp,wIs the constant pressure specific heat of water; t is twi,cIs the cooling water inlet temperature; t is two,cIs the cooling water outlet temperature; t is tri,cIs the refrigerant inlet temperature; t is tro,cIs the refrigerant outlet temperature; m isrIs the refrigerant flow rate; h isri,cA condenser inlet refrigerant enthalpy; h is a total ofro,cIs the enthalpy value of the outlet of the condenser; q1,c、Q2,c、Q3,cThe heat exchange capacity of a superheat zone, a two-phase zone and a supercooling zone of the condenser is respectively; f1,c、F2,c、F3,cThe heat exchange areas of a superheat area, a two-phase area and a supercooling area of the condenser are respectively; Δ t1,c、Δt2,c、Δt3,cThe heat exchange temperature difference of a superheat zone, a two-phase zone and a supercooling zone of the condenser is respectively; k1,c、K2,c、K3,cHeat transfer coefficients of a superheat zone, a two-phase zone and a supercooling zone of the condenser are respectively set;
the modeling of the evaporator comprises:
neglecting the heat exchange between the evaporator and the outside, and considering the flow of the refrigerant and the chilled water as a one-dimensional uniform flow, the heat exchange process in the evaporator is obtained as follows:
Qe＝mw,ecp,w(twi,e-two,e)＝mr(1-x)(hro,e-hri,e)；
Q1,e＝K1,eF1,eΔt1,e；
Q2,e＝K2,eF2,eΔt2,e；
Figure BDA0003436111540000041
Figure BDA0003436111540000042
wherein Q iseThe heat exchange capacity of the evaporator; m isw,eIs the flow rate of the chilled water; t is twi,eThe temperature of the inlet chilled water of the evaporator; t is two,eThe temperature of the chilled water at the outlet of the evaporator; t is tw1,eThe temperature of the chilled water at the inlet of the two-phase zone; h isri,eIs the evaporator inlet enthalpy; h is a total ofro,eIs the evaporator outlet enthalpy; x is the dryness of the refrigerant at the inlet of the evaporator; q1,eThe heat exchange quantity of the superheat zone of the evaporator is obtained;
Q2,ethe heat exchange quantity of the two-phase area of the evaporator; Δ t1,eHeat exchange temperature difference is carried out in an overheating area of the evaporator; Δ t2,eHeat exchange temperature difference of two phase regions of the evaporator; t is tro,eThe temperature is the suction temperature of the compressor, namely the temperature of the refrigerant at the outlet of the evaporator; t is tr,eIs the evaporation temperature;
the modeling of the throttle valve comprises:
the thermostatic expansion valve is formed by the pressure P of a temperature sensing medium in a temperature sensing bulbbProviding a valve opening force, by steam pressure PcAnd a spring force providing a closing force, the spring force being a minimum at Δ P when the valve is in the closed statemin(ii) a Displacement y and P of the valveb、Pc、ΔPminThe relationship between them is expressed as: k (P)b-Pc-ΔPmin) (ii) a k is the reciprocal of the spring coefficient;
the water pump model establishment comprises the following steps:
the rotation speed ratio f of the water pump is defined as the rotation speed n of the water pump motor and the rated motor rotation speed n0The ratio, expressed as:
Figure BDA0003436111540000043
the relationship between the pump lift and the pump efficiency and the flow and rotation ratio of the water pump is expressed as follows:
Figure BDA0003436111540000044
Figure BDA0003436111540000045
wherein HpuIs the pump lift; m iswMass flow of the water pump; etapuThe efficiency of the water pump; h is01、h02、h03、h11、h12、h13Are fitting coefficients.
Further, in step S2, acquiring state data of the central air conditioning system during normal operation and different faults through a plurality of sensors, and obtaining a sample data set after data preprocessing and feature extraction, specifically including:
acquiring the temperature of the inlet and outlet of the compressor, the temperature of the inlet and outlet of the evaporator, the temperature of the inlet and outlet of the condenser and the temperature of the inlet and outlet of the condenser at the time of normal operation and simulated fault by thermocouples arranged on the walls of the inlet and outlet pipes of the compressor, the water inlet and outlet of the evaporator and the water inlet and outlet of the condenser; the pressure at the inlet and the outlet of the compressor during normal operation and fault simulation is acquired by a pressure sensor arranged at the inlet and the outlet of the compressor; collecting cold water flow and cooling water flow during normal operation and simulated fault through flow sensors arranged on horizontal pipes at outlets of a cold water pump and a cooling water pump;
carrying out denoising processing, missing value filling, repeated invalid value deletion and normalization preprocessing on the acquired data;
taking the state characteristic parameters of the central air-conditioning equipment as independent variables and the fault label characteristics of the central air-conditioning equipment as dependent variables, extracting the characteristics of the preprocessed data variables by adopting a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm, selecting the extracted characteristics according to a grey correlation algorithm, establishing a mapping relation between the state characteristic parameters of the central air-conditioning equipment and the fault label characteristics, and obtaining a sample data set;
wherein the fault signature includes at least: the flow of cooling water is increased or reduced, the flow of chilled water is increased or reduced, the water inlet temperature of a condenser is overhigh, non-condensable gas exists in a refrigerant, and the refrigerant leaks; the characteristic variables correspondingly selected by the fault label characteristics at least comprise condenser water inlet temperature, condenser water outlet temperature, evaporator water inlet temperature and evaporator water outlet temperature.
Further, the performing feature extraction on the preprocessed data variable by using a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm includes: decomposing the preprocessed data variable by a wavelet packet decomposition algorithm into a plurality of wavelet values, reconstructing wavelet characteristics of the decomposed wavelet values by a wavelet packet reconstruction algorithm, and finally outputting a wavelet characteristic data set of the preprocessed data;
the selecting the extracted features according to the grey correlation algorithm comprises the following steps: and calculating the relevance values corresponding to the extracted features according to a grey relevance algorithm, sorting the relevance values corresponding to the extracted features in size, distinguishing the relevance degree between the state characteristic parameters of each central air-conditioning equipment and the fault label features, and taking the characteristic parameters with larger relevance degrees as sample data sets for fault diagnosis.
Further, in step S4, training each base learner by using a k-fold cross validation method, and obtaining a prediction result of each base learner as a secondary training data set; when training each base learner, select the different machine learning algorithm of multiunit to make up, generate the secondary training data set under the multiunit composite mode, specifically include:
determining the number of the base learners as m, and randomly dividing the sample data set D into m data sets D with equal size1、D2、D3……DjDefinition of DjAnd D-j＝D-DjJ is 1,2,3, … …, m is j-fold test data set and training data set of k-fold cross validation, respectively, in training data set D-jGet the model on the training base learner
Figure BDA0003436111540000061
Figure BDA0003436111540000062
m for test data set DjEach sample, model of
Figure BDA0003436111540000063
Will output a test result; when the cross validation process is finished, obtaining the output result of the base learner on the whole training data set;
converting the obtained output result of the base learner into a probability type result, keeping the results of the m base learners within a [0,1] interval, and splicing the probability type output result and the training set labels to form a new training set as a secondary training data set;
when the m base learners are trained, the selectable machine learning algorithm comprises the following steps: the method comprises the following steps of (1) SVM, BP neural network, random forest, GBDT model, XGboost model, Light GBM model, linear regression model, support vector machine and weighted extreme learning machine; and according to the prediction performances of different algorithms, fixing the value of m, selecting various different algorithms to combine to generate various combined working conditions of the base learner or performing different values on m, and selecting different algorithms to combine to generate various combined working conditions of the base learner.
Further, in step S5, inputting multiple sets of secondary training data sets into a secondary learner for training to obtain multiple central air-conditioning fault diagnosis models, including: sequentially inputting secondary training data sets obtained by training of all base learners under different combination working conditions into a secondary learner for training to obtain a plurality of central air conditioner fault diagnosis models; the machine learning algorithm used by the secondary learner is one of SVM, BP neural network, random forest, GBDT model, XGboost model, Light GBM model, linear regression model, support vector machine and weighted extreme learning machine.
Further, in step S6, the evaluating the prediction performance of the multiple central air-conditioning fault diagnosis models through the test data set, selecting the model with the best prediction performance as the optimal central air-conditioning fault diagnosis model, and performing fault diagnosis of the central air-conditioning system through the model includes:
calculating the mean absolute error value MAE, the root mean square difference value RMSE and the fitting degree R of the model2As the evaluation standard of the central air-conditioning fault diagnosis model, the better the model performance is, the smaller the average absolute error value MAE and the root mean square difference value RMSE are, and the fitting degree R is2The larger;
Figure BDA0003436111540000064
Figure BDA0003436111540000065
Figure BDA0003436111540000066
wherein, yi、
Figure BDA0003436111540000067
Respectively an actual value, a predicted value and a mean value of the sample; n is the size of the test data set.
Further, a base learner in the double-layer stacking model is a weighted base learner, each base learner is endowed with a weight by introducing a weight formula based on a G-mean value, an output result is corrected according to the weight, and the corrected result is fused into a secondary training data set and is input into a secondary learner to obtain a final central air-conditioning fault diagnosis model;
wherein, the weight calculation formula of each base learner is as follows:
Figure BDA0003436111540000071
αiis the output weight; GM (GM)iG-mean values for the ith base learner for the input sample set;
and the secondary learner in the stacking model is a secondary learner introducing an attention mechanism, when the n-dimensional input variable in the secondary training data set is input to the secondary learner at the time t, the weight of the n-dimensional features is calculated firstly, then normalization processing is carried out on the obtained weight to obtain the weight importance degree ratio of different features, finally, the obtained weight and the weight ratio are weighted to obtain a final feature vector, and the central air-conditioning fault diagnosis model is optimized and output.
The second aspect of the present invention further provides a fault diagnosis device for a central air conditioning system based on a stacking fusion algorithm, where the fault diagnosis device for the central air conditioning system includes:
a digital twin model building module: establishing a digital twin model of the central air-conditioning system by adopting a mechanism modeling and data identification method;
a sample data acquisition module: acquiring state data of a central air-conditioning system in normal operation and different faults through a plurality of sensors, and performing data preprocessing and feature extraction to obtain a sample data set;
a stacking model building module: dividing a sample data set into a training data set and a testing data set, simultaneously building a double-layer stacking model, and determining that the number of base learners is m and the number of secondary learners is 1;
a base learner training module: training each base learner by adopting a k-fold cross validation method, and obtaining a prediction result of each base learner as a secondary training data set; when each base learner is trained, selecting a plurality of groups of different machine learning algorithms for combination to generate a plurality of groups of secondary training data sets under a combination mode;
a secondary learner training module: inputting a plurality of groups of secondary training data sets into a secondary learner for training to obtain a plurality of central air conditioner fault diagnosis models;
the fault diagnosis model evaluation module: and evaluating the prediction performance of the plurality of central air-conditioning fault diagnosis models through the test data set, selecting the model with the best prediction performance as the optimal central air-conditioning fault diagnosis model, and performing fault diagnosis on the central air-conditioning system through the model.
The invention has the beneficial effects that:
(1) the invention establishes a central air-conditioning system digital twin model by adopting a mechanism modeling and data identification method, performs virtual simulation mapping on an actual central air-conditioning system, inputs actual measurement data for identification and correction, improves the precision of the model, provides a basis for subsequently establishing a central air-conditioning fault diagnosis model, realizes the prediction of the fault diagnosis model based on the digital twin model, and makes a fault diagnosis decision based on the model prediction;
(2) the method comprises the steps of carrying out denoising processing, missing value filling, repeated invalid value deleting and normalization preprocessing on collected data; taking the state characteristic parameters of the central air-conditioning equipment as independent variables and the fault label characteristics of the central air-conditioning as dependent variables, extracting the characteristics of the preprocessed data variables by adopting a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm, selecting the extracted characteristics according to a grey correlation algorithm, establishing a mapping relation between the state characteristic parameters of the central air-conditioning and the fault label characteristics, obtaining a sample data set, extracting the characteristics of the input characteristic parameters and analyzing the correlation, screening out important characteristic parameters, and reducing the influence of irrelevant factors;
(3) according to the method, a double-layer stacking model is built, the number of the base learners is determined to be m, and the number of the secondary learners is 1; training each base learner by adopting a k-fold cross validation method, and obtaining a prediction result of each base learner as a secondary training data set; when training each base learner, selecting a plurality of groups of different machine learning algorithms for combination to generate a plurality of groups of secondary training data sets under a combination mode; inputting a plurality of groups of secondary training data sets into a secondary learner for training to obtain a plurality of central air conditioner fault diagnosis models; the prediction performance of the multiple central air-conditioning fault diagnosis models is evaluated through the test data set, the model with the best prediction performance is selected as the optimal central air-conditioning fault diagnosis model, and multiple models are fused through stacking, so that the prediction error can be effectively reduced and the prediction precision can be improved compared with a single model;
(4) the method comprises the steps of setting a base learner in a double-layer stacking model as a weighted base learner, endowing each base learner with a weight by introducing a weight formula based on a G-mean value, correcting an output result according to the weight, fusing the corrected result into a secondary training data set, inputting the secondary training data set into a secondary learner to obtain a final central air-conditioning fault diagnosis model, setting a weight for the base learner according to the quality of the classification effect of the base learner, correcting the classification result, fusing the corrected classification result, and making a final decision for the whole integrated learner model by the corrected secondary learner to have a positive effect, so that the distribution characteristics of output information of the base learner are optimized; the secondary learner is a secondary learner with an attention mechanism, so that the utilization effect of the secondary learner on the characteristics of the base learner is enhanced, and the prediction accuracy of the fault diagnosis model is improved.
Additional features and advantages will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart of a central air conditioning system fault diagnosis method based on a stacking fusion algorithm of the present invention;
FIG. 2 is a schematic diagram of the central air conditioning system according to the present invention;
FIG. 3 is a schematic diagram of the heat exchange of the central air conditioning system of the present invention;
FIG. 4 is a block diagram of the stacking algorithm of the present invention;
fig. 5 is a schematic structural diagram of a central air conditioning system fault diagnosis device based on a stacking fusion algorithm.
Detailed Description
To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
Fig. 1 is a flow chart of a central air conditioning system fault diagnosis method based on a stacking fusion algorithm according to the present invention.
As shown in fig. 1, this embodiment 1 provides a method for diagnosing a fault of a central air conditioning system based on a stacking fusion algorithm, where the method for diagnosing a fault of a central air conditioning system includes:
s1, establishing a digital twin model of the central air-conditioning system by adopting a mechanism modeling and data identification method;
s2, acquiring state data of the central air-conditioning system during normal operation and different faults through a plurality of sensors, preprocessing the state data, extracting features of preprocessed data variables by adopting a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm, and selecting the extracted features according to a grey correlation algorithm to obtain a sample data set;
step S3, dividing the sample data set into a training data set and a testing data set, and simultaneously building an improved double-layer stacking model, and determining that the number of the base learners is m and the number of the secondary learners is 1;
step S4, training each base learner by adopting a k-fold cross validation method, and obtaining the prediction result of each base learner as a secondary training data set; when each base learner is trained, selecting a plurality of groups of different machine learning algorithms to be combined to generate a plurality of groups of secondary training data sets under the combined mode;
step S5, inputting a plurality of groups of secondary training data sets into a secondary learner for training to obtain a plurality of central air-conditioning fault diagnosis models;
and step S6, evaluating the prediction performance of the plurality of central air-conditioning fault diagnosis models through the test data set, selecting the model with the best prediction performance as the optimal central air-conditioning fault diagnosis model, and diagnosing the faults of the central air-conditioning system through the model.
Fig. 2 is a schematic structural diagram of a central air conditioning system according to the present invention.
Fig. 3 is a schematic diagram of heat exchange of the central air conditioning system according to the present invention.
As shown in fig. 2 and 3, in this embodiment, in step S1, the establishing a digital twin model of a central air conditioning system by using a mechanism modeling and data identification method specifically includes:
constructing a physical model, a logic model and a simulation model of the central air-conditioning system; wherein the content of the first and second substances,
the construction of the physical model comprises the following steps: at least establishing physical models of a water chilling unit, a chilled water circulation system and a cooling water circulation system entity, wherein the water chilling unit provides chilled water with a certain temperature for the tail end and consists of a compressor, an evaporator, a condenser and a throttle valve; the chilled water circulating system is used for transmitting chilled water to a cooling coil or a tail end fan coil in an air processor to cool indoor return air and consists of a chilled water pump, a chilled water pipe and an air processing unit or a fan coil; the cooling water circulation system releases heat absorbed in a refrigerant of the water chilling unit into air through a cooling tower and consists of a cooling water pump, a cooling water pipe and the cooling tower;
the construction of the logic model comprises the following steps: establishing a controllable closed-loop logic model according to a logic mechanism relation among all physical entities of the central air-conditioning system, and mapping the physical model to the logic model;
the construction of the simulation model comprises the following steps: building a simulation model of the central air-conditioning system based on the collected operation data, state data and physical attribute data of the central air-conditioning system;
carrying out virtual-real fusion on the physical model, the logic model and the simulation model to construct a system-level digital twin model of a physical entity of the central air-conditioning system in a virtual space;
and accessing multi-working-condition real-time operation data of the central air-conditioning system into the system-level digital twin model, and performing self-adaptive identification correction on the simulation result of the system-level digital twin model by adopting a reverse identification method to obtain the digital twin model of the central air-conditioning system after identification correction.
In this embodiment, the establishment of the water chiller model includes:
neglecting the pressure loss of the suction and exhaust of the compressor and neglecting the heat exchange between the compressor and the environment, establishing a compressor model expressed as:
Figure BDA0003436111540000101
Figure BDA0003436111540000102
Figure BDA0003436111540000103
Figure BDA0003436111540000104
wherein m isrIs the refrigerant mass flow rate; vthFor theoretical gas transport of compressorsAn amount; v. of1The specific volume of the air suction of the compressor; xi is the gas transmission coefficient; pthsThe theoretical power consumption of the compressor in the isentropic compression process is realized; piThe power consumption of the actual compression process of the compressor is the indicated power; p iselThe electric power required to be input for the actual compression process of the compressor is the power measured by the power meter; k is an isentropic compression index; peIs the evaporating pressure, i.e. the compressor suction pressure; pkIs the condensing pressure, i.e. the compressor discharge pressure; etaiIndicating efficiency for the compressor; etaelElectrical efficiency of the compressor; h is2Is the enthalpy of the refrigerant at the outlet of the compressor; h is1Is the enthalpy of the compressor inlet refrigerant;
the condenser model building comprises the following steps:
neglecting the heat exchange between the condenser and the outside, and considering the flow of the refrigerant and the cooling water as a one-dimensional uniform flow, the heat exchange process in the condenser is obtained as follows:
Qc＝mw,ccp,w(two,c-twi,c)＝mr(hri,c-hro,c)；
Q1,c＝K1,cF1,cΔt1,c；
Q2,c＝K2,cF2,cΔt2,c；
Q3,c＝K3,cF3,cΔt3,c；
Figure BDA0003436111540000111
Figure BDA0003436111540000112
Figure BDA0003436111540000113
wherein Q iscThe total heat exchange capacity of the condenser; m isw,cFor cooling water flow；cp,wIs the constant pressure specific heat of water; t is twi,cIs the cooling water inlet temperature; t is two,cIs the cooling water outlet temperature; t is tri,cIs the refrigerant inlet temperature; t is tro,cIs the refrigerant outlet temperature; m isrIs the refrigerant flow rate; h isri,cA condenser inlet refrigerant enthalpy; h isro,cIs the enthalpy value of the outlet of the condenser; q1,c、Q2,c、Q3,cThe heat exchange capacity of a superheat zone, a two-phase zone and a supercooling zone of the condenser is respectively; f1,c、F2,c、F3,cThe heat exchange areas of a superheat area, a two-phase area and a supercooling area of the condenser are respectively; Δ t1,c、Δt2,c、Δt3,cThe heat exchange temperature difference of a superheat zone, a two-phase zone and a supercooling zone of the condenser is respectively; k1,c、K2,c、K3,cHeat transfer coefficients of a superheat zone, a two-phase zone and a supercooling zone of the condenser are respectively set;
the evaporator model building comprises:
neglecting the heat exchange between the evaporator and the outside, and considering the flow of the refrigerant and the chilled water as a one-dimensional uniform flow, the heat exchange process in the evaporator is obtained as follows:
Qe＝mw,ecp,w(twi,e-two,e)＝mr(1-x)(hro,e-hri,e)；
Q1,e＝K1,eF1,eΔt1,e；
Q2,e＝K2,eF2,eΔt2,e；
Figure BDA0003436111540000114
Figure BDA0003436111540000115
wherein QeThe heat exchange capacity of the evaporator; m isw,eIs the flow of chilled water; t is twi,eThe temperature of the inlet chilled water of the evaporator; t is two,eThe temperature of the chilled water at the outlet of the evaporator;tw1,ethe temperature of the chilled water at the inlet of the two-phase zone; h isri,eIs the evaporator inlet enthalpy; h isro,eIs the evaporator outlet enthalpy; x is the dryness of the refrigerant at the inlet of the evaporator; q1,eThe heat exchange quantity of the superheat zone of the evaporator is obtained; q2,eThe heat exchange quantity of the two-phase area of the evaporator; Δ t1,eHeat exchange temperature difference is carried out in an overheating area of the evaporator; Δ t2,eHeat exchange temperature difference of two phase regions of the evaporator; t is tro,eThe temperature is the suction temperature of the compressor, namely the temperature of the refrigerant at the outlet of the evaporator; t is tr,eIs the evaporation temperature;
the throttle valve model building comprises the following steps:
the thermostatic expansion valve is formed by the pressure P of a temperature sensing medium in a temperature sensing bulbbProviding a valve opening force, by steam pressure PcAnd a spring force providing a closing force, the spring force being a minimum at Δ P when the valve is in the closed statemin(ii) a Displacement y and P of the valveb、Pc、ΔPminThe relationship between them is expressed as: k (P)b-Pc-ΔPmin) (ii) a k is the reciprocal of the spring coefficient;
the water pump model establishment comprises the following steps:
the rotation speed ratio f of the water pump is defined as the rotation speed n of the water pump motor and the rated motor rotation speed n0The ratio, expressed as:
Figure BDA0003436111540000121
the relationship between the pump lift and the pump efficiency and the flow and rotation ratio of the water pump is expressed as follows:
Figure BDA0003436111540000122
Figure BDA0003436111540000123
wherein HpuIs the pump lift; m iswMass flow of the water pump; etapuThe efficiency of the water pump; h is01、h02、h03、h11、h12、h13Are fitting coefficients.
In this embodiment, in step S2, acquiring state data of the central air conditioning system during normal operation and different faults through a plurality of sensors, and obtaining a sample data set after performing data preprocessing and feature extraction, specifically including:
acquiring the temperature of the inlet and outlet of the compressor, the temperature of the inlet and outlet of the evaporator, the temperature of the inlet and outlet of the condenser and the temperature of the inlet and outlet of the condenser at the time of normal operation and simulated fault by thermocouples arranged on the walls of the inlet and outlet pipes of the compressor, the water inlet and outlet of the evaporator and the water inlet and outlet of the condenser; the pressure at the inlet and the outlet of the compressor during normal operation and fault simulation is acquired by a pressure sensor arranged at the inlet and the outlet of the compressor; collecting cold water flow and cooling water flow during normal operation and simulated fault through flow sensors arranged on horizontal pipes at outlets of a cold water pump and a cooling water pump;
carrying out denoising processing, missing value filling, repeated invalid value deletion and normalization preprocessing on the acquired data;
taking the state characteristic parameters of the central air-conditioning equipment as independent variables and the fault label characteristics of the central air-conditioning equipment as dependent variables, extracting the characteristics of the preprocessed data variables by adopting a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm, selecting the extracted characteristics according to a grey correlation algorithm, establishing a mapping relation between the state characteristic parameters of the central air-conditioning equipment and the fault label characteristics, and obtaining a sample data set;
wherein the fault signature includes at least: the flow of cooling water is increased or reduced, the flow of chilled water is increased or reduced, the water inlet temperature of a condenser is overhigh, non-condensable gas exists in a refrigerant, and the refrigerant leaks; the characteristic variables correspondingly selected by the fault label characteristics at least comprise condenser water inlet temperature, condenser water outlet temperature, evaporator water inlet temperature and evaporator water outlet temperature.
The fault simulation method for increasing or decreasing the flow rate of the cooling water includes: when the water chilling unit stably runs under a rated working condition, a gate valve on a cooling water pipe and a stop valve on a bypass pipe are adjusted; the fault simulation method for increasing or decreasing the cold water quantity comprises the following steps: when the water chilling unit stably runs under a rated working condition, a gate valve on a cold water pipe and a stop valve on a bypass pipe are adjusted; the fault simulation method for the overhigh water inlet temperature of the condenser comprises the following steps: turning off or reversing a fan of the cooling tower; the fault simulation method for the presence of the non-condensable gas in the refrigerant comprises the following steps: when the water chilling unit stably runs under a rated working condition, nitrogen is flushed from a fluoridizing port of the water chilling unit; the refrigerant fault simulation method comprises the following steps: the evaporator outlet temperature increases.
In this embodiment, the performing feature extraction on the preprocessed data variable by using the wavelet packet decomposition algorithm and the wavelet packet reconstruction algorithm includes: decomposing the preprocessed data variable by a wavelet packet decomposition algorithm into a plurality of wavelet values, reconstructing wavelet characteristics of the decomposed wavelet values by a wavelet packet reconstruction algorithm, and finally outputting a wavelet characteristic data set of the preprocessed data;
the selecting the extracted features according to the grey correlation algorithm comprises the following steps: and calculating the relevance values corresponding to the extracted features according to a grey relevance algorithm, sorting the relevance values corresponding to the extracted features in size, distinguishing the relevance degree between the state characteristic parameters of each central air-conditioning equipment and the fault label features, and taking the characteristic parameters with larger relevance degrees as sample data sets for fault diagnosis.
Fig. 4 is a block diagram of the improved stacking algorithm according to the present invention.
As shown in fig. 4, in the present embodiment, in step S4, each base learner is trained by using a k-fold cross validation method, and a prediction result of each base learner is obtained as a secondary training data set; when training each base learner, select the different machine learning algorithm of multiunit to make up, generate the secondary training data set under the multiunit composite mode, specifically include:
determining the number of the base learners as m, and randomly dividing the sample data set D into m data sets D with equal size1、D2、D3……DjDefinition of DjAnd D-j＝D-DjJ is 1,2,3, … …, m is k fold and crossCross-validation of the jth test data set and the training data set in the training data set D-jGet the model on the training base learner
Figure BDA0003436111540000131
Figure BDA0003436111540000132
m for test data set DjEach sample, model of
Figure BDA0003436111540000133
Will output a test result; when the cross validation process is finished, obtaining the output result of the base learner on the whole training data set;
converting the obtained output result of the base learner into a probability type result, keeping the results of the m base learners within a [0,1] interval, and splicing the probability type output result and the training set labels to form a new training set as a secondary training data set;
when the m base learners are trained, the selectable machine learning algorithm comprises the following steps: the method comprises the following steps of (1) SVM, BP neural network, random forest, GBDT model, XGboost model, Light GBM model, linear regression model, support vector machine and weighted extreme learning machine; and according to the prediction performances of different algorithms, fixing the value of m, selecting various different algorithms to combine to generate various combined working conditions of the base learner or performing different values on m, and selecting different algorithms to combine to generate various combined working conditions of the base learner.
In this embodiment, in step S5, inputting multiple sets of secondary training data sets into a secondary learner for training to obtain multiple central air-conditioning fault diagnosis models, where the method includes: sequentially inputting secondary training data sets obtained by training of all base learners under different combination working conditions into a secondary learner for training to obtain a plurality of central air conditioner fault diagnosis models; the machine learning algorithm used by the secondary learner is one of SVM, BP neural network, random forest, GBDT model, XGboost model, Light GBM model, linear regression model, support vector machine and weighted extreme learning machine.
In practical application, the learning effect of the base learner is analyzed by observing the learning curve of the base learner, whether the phenomenon of over-fitting or under-fitting occurs is judged, and the parameter value of the base learner is adjusted, so that the model has a better prediction effect. And calling the selected classification learning algorithm in the second layer model, performing fusion learning on the prediction result of the first layer model, training to obtain a strong classifier, and setting the parameter setting value of the selected classification algorithm.
In this embodiment, in step S6, the evaluating the prediction performances of the multiple central air-conditioning fault diagnosis models through the test data set, selecting the model with the best prediction performance as the optimal central air-conditioning fault diagnosis model, and performing fault diagnosis on the central air-conditioning system through the model includes:
calculating the average absolute error value MAE, the root mean square difference value RMSE and the fitting degree R of the model2As the evaluation standard of the central air-conditioning fault diagnosis model, the better the model performance is, the smaller the average absolute error value MAE and the root mean square difference value RMSE are, and the fitting degree R is2The larger;
Figure BDA0003436111540000141
Figure BDA0003436111540000142
Figure BDA0003436111540000143
wherein, yi、
Figure BDA0003436111540000144
Respectively an actual value, a predicted value and a mean value of the sample; n is the size of the test data set.
In this embodiment, the base learners in the double-layer stacking model are weighted base learners, a weight is given to each base learner by introducing a weight formula based on a G-mean value, an output result is corrected according to the weight, and the corrected result is fused into a secondary training data set and input into a secondary learner to obtain a final central air-conditioning fault diagnosis model;
wherein, the weight calculation formula of each base learner is as follows:
Figure BDA0003436111540000145
αiis the output weight; GM (GM)iG-mean values for the ith base learner for the input sample set;
and after the n-dimensional input variable in the secondary training data set is input to the secondary learner at the moment t, firstly calculating the weight of the n-dimensional features, then carrying out normalization processing on the obtained weight to obtain the weight importance degree ratio of different features, finally obtaining a final feature vector by weighting the obtained weight and the weight ratio, and optimizing and outputting the central air-conditioning fault diagnosis model.
Example 2
Fig. 5 is a schematic structural diagram of a central air conditioning system fault diagnosis device based on a stacking fusion algorithm according to the present invention.
As shown in fig. 5, in the present embodiment, a second aspect of the present invention further provides a central air conditioning system fault diagnosis device based on a stacking fusion algorithm, where the central air conditioning system fault diagnosis device includes:
a digital twin model building module: establishing a digital twin model of the central air-conditioning system by adopting a mechanism modeling and data identification method;
a sample data acquisition module: acquiring state data of a central air-conditioning system during normal operation and different faults through a plurality of sensors, and obtaining a sample data set after data preprocessing and feature extraction;
a stacking model building module: dividing a sample data set into a training data set and a testing data set, simultaneously building a double-layer stacking model, and determining that the number of base learners is m and the number of secondary learners is 1;
the base learner training module: training each base learner by adopting a k-fold cross validation method, and obtaining a prediction result of each base learner as a secondary training data set; when each base learner is trained, selecting a plurality of groups of different machine learning algorithms to be combined to generate a plurality of groups of secondary training data sets under the combined mode;
a secondary learner training module: inputting a plurality of groups of secondary training data sets into a secondary learner for training to obtain a plurality of central air conditioner fault diagnosis models;
the fault diagnosis model evaluation module: and evaluating the prediction performance of the plurality of central air-conditioning fault diagnosis models through the test data set, selecting the model with the best prediction performance as the optimal central air-conditioning fault diagnosis model, and performing fault diagnosis on the central air-conditioning system through the model.
The invention establishes a digital twin model of the central air-conditioning system by adopting a mechanism modeling and data identification method, performs virtual simulation mapping on the actual central air-conditioning system, inputs the actual measurement data for identification and correction, improves the precision of the model, provides a basis for subsequently establishing a central air-conditioning fault diagnosis model, realizes the prediction of the fault diagnosis model based on the digital twin model, and makes fault diagnosis decision based on the model prediction.
The method comprises the steps of carrying out denoising processing, missing value filling, repeated invalid value deleting and normalization preprocessing on collected data; the method comprises the steps of taking a state characteristic parameter of central air-conditioning equipment as an independent variable and a fault label characteristic of the central air-conditioning equipment as a dependent variable, extracting the characteristic of a preprocessed data variable by adopting a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm, selecting the extracted characteristic according to a grey correlation algorithm, establishing a mapping relation between the state characteristic parameter of the central air-conditioning and the fault label characteristic, obtaining a sample data set, extracting the characteristic of the input characteristic parameter and analyzing the correlation, screening out important characteristic parameters, and reducing the influence of irrelevant factors.
According to the method, a double-layer stacking model is built, the number of the base learners is determined to be m, and the number of the secondary learners is 1; training each base learner by adopting a k-fold cross validation method, and obtaining a prediction result of each base learner as a secondary training data set; when training each base learner, selecting a plurality of groups of different machine learning algorithms for combination to generate a plurality of groups of secondary training data sets under a combination mode; inputting a plurality of groups of secondary training data sets into a secondary learner for training to obtain a plurality of central air conditioner fault diagnosis models; the prediction performance of the multiple central air-conditioning fault diagnosis models is evaluated through the test data set, the model with the best prediction performance is selected as the optimal central air-conditioning fault diagnosis model, and multiple models are fused through stacking, so that the prediction error can be effectively reduced and the prediction precision can be improved compared with a single model.
The method comprises the steps of setting a base learner in a double-layer stacking model as a weighted base learner, endowing each base learner with a weight by introducing a weight formula based on a G-mean value, correcting an output result according to the weight, fusing the corrected result into a secondary training data set, inputting the secondary training data set into a secondary learner to obtain a final central air-conditioning fault diagnosis model, setting a weight for the base learner according to the quality of the classification effect of the base learner, correcting the classification result, fusing the corrected classification result, and making a final decision for the whole integrated learner model by the corrected secondary learner to have a positive effect, so that the distribution characteristics of output information of the base learner are optimized; the secondary learner is a secondary learner with an attention mechanism, so that the utilization effect of the secondary learner on the characteristics of the base learner is enhanced, and the prediction accuracy of the fault diagnosis model is improved.
In the several embodiments provided in the present application, it should be understood that the disclosed system and method may be implemented in other ways. The system embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, the functional modules in the embodiments of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In light of the foregoing description of the preferred embodiment of the present invention, many modifications and variations will be apparent to those skilled in the art without departing from the spirit and scope of the invention. The technical scope of the present invention is not limited to the content of the specification, and must be determined according to the scope of the claims.

Claims (10)

1. A fault diagnosis method of a central air-conditioning system based on a stacking fusion algorithm is characterized by comprising the following steps:
s1, establishing a digital twin model of the central air-conditioning system by adopting a mechanism modeling and data identification method;
s2, acquiring state data of the central air-conditioning system during normal operation and different faults through a plurality of sensors, preprocessing the state data, extracting features of preprocessed data variables by adopting a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm, and selecting the extracted features according to a grey correlation algorithm to obtain a sample data set;
step S3, dividing the sample data set into a training data set and a testing data set, and simultaneously building a double-layer stacking model, determining the number of the base learners as m and the number of the secondary learners as 1;
step S4, training each base learner by adopting a k-fold cross validation method, and obtaining a prediction result of each base learner as a secondary training data set; when training each base learner, selecting a plurality of groups of different machine learning algorithms for combination to generate a plurality of groups of secondary training data sets under a combination mode;
step S5, inputting a plurality of groups of secondary training data sets into a secondary learner for training to obtain a plurality of central air-conditioning fault diagnosis models;
and step S6, evaluating the prediction performance of the plurality of central air-conditioning fault diagnosis models through the test data set, selecting the model with the best prediction performance as the optimal central air-conditioning fault diagnosis model, and diagnosing the faults of the central air-conditioning system through the optimal central air-conditioning fault diagnosis model.
2. The method for diagnosing faults of a central air-conditioning system based on a stacking fusion algorithm as claimed in claim 1, wherein in the step S1, establishing a digital twin model of the central air-conditioning system by using a mechanism modeling and data identification method comprises:
constructing a physical model, a logic model and a simulation model of the central air-conditioning system; wherein the content of the first and second substances,
the construction of the physical model comprises the following steps: at least establishing a water chilling unit, a chilled water circulating system and a cooling water circulating system; the water chilling unit comprises a compressor, an evaporator, a condenser and a throttle valve; the chilled water circulating system comprises a chilled water pump, a chilled water pipe and an air handling unit; the cooling water circulation system comprises a cooling water pump, a cooling water pipe and a cooling tower;
the construction of the logic model comprises the following steps: establishing a controllable closed-loop logic model according to a logic mechanism relation among all physical entities of the central air-conditioning system, and mapping the physical model to the logic model;
the construction of the simulation model comprises the following steps: building a simulation model of the central air-conditioning system based on the collected operation data, state data and physical attribute data of the central air-conditioning system;
carrying out virtual-real fusion on the physical model, the logic model and the simulation model to construct a system-level digital twin model of a physical entity of the central air-conditioning system in a virtual space;
and accessing multi-working-condition real-time operation data of the central air-conditioning system into the system-level digital twin model, and performing self-adaptive identification correction on the simulation result of the system-level digital twin model by adopting a reverse identification method to obtain the digital twin model of the central air-conditioning system after identification correction.
3. The method for diagnosing the fault of the central air conditioning system based on the stacking fusion algorithm as claimed in claim 2, wherein the modeling of the water chilling unit comprises:
neglecting the pressure loss of the suction and exhaust of the compressor and neglecting the heat exchange between the compressor and the environment, establishing a compressor model expressed as:
Figure FDA0003436111530000021
Figure FDA0003436111530000022
Figure FDA0003436111530000023
Figure FDA0003436111530000024
wherein m isrIs the refrigerant mass flow rate; vthThe theoretical gas transmission capacity of the compressor; v. of1The specific volume of the air suction of the compressor; xi is the gas transmission coefficient; pthsThe theoretical power consumption of the compressor in the isentropic compression process is realized; p isiThe power consumption of the actual compression process of the compressor is the indicated power; pelThe electric power required to be input for the actual compression process of the compressor is the power measured by the power meter; k is an isentropic compression index; peIs the evaporating pressure, i.e. the compressor suction pressure; pkIs the condensing pressure, i.e. the compressor discharge pressure; etaiIndicating efficiency for the compressor; etaelElectrical efficiency of the compressor; h is2Is the enthalpy of the refrigerant at the outlet of the compressor; h is1Is the enthalpy of the compressor inlet refrigerant;
the modeling of the condenser comprises:
neglecting the heat exchange between the condenser and the outside, and considering the flow of the refrigerant and the cooling water as a one-dimensional uniform flow, the heat exchange process in the condenser is obtained as follows:
Qc＝mw,ccp,w(two,c-twi,c)＝mr(hri,c-hro,c)；
Q1,c＝K1,cF1,cΔt1,c；
Q2,c＝K2,cF2,cΔt2,c；
Q3,c＝K3,cF3,cΔt3,c；
Figure FDA0003436111530000031
Figure FDA0003436111530000032
Figure FDA0003436111530000033
wherein Q iscThe total heat exchange capacity of the condenser; m isw,cCooling water flow rate; c. Cp,wIs the constant pressure specific heat of water; t is twi,cIs the cooling water inlet temperature; t is two,cIs the cooling water outlet temperature; t is tri,cIs the refrigerant inlet temperature; t is tro,cIs the refrigerant outlet temperature; m isrIs the refrigerant flow rate; h isri,cA condenser inlet refrigerant enthalpy; h isro,cIs the enthalpy value of the outlet of the condenser; q1,c、Q2,c、Q3,cThe heat exchange capacity of a superheat zone, a two-phase zone and a supercooling zone of the condenser is respectively; f1,c、F2,c、F3,cThe heat exchange areas of a superheat area, a two-phase area and a supercooling area of the condenser are respectively; Δ t1,c、Δt2,c、Δt3,cThe heat exchange temperature difference of a superheat zone, a two-phase zone and a supercooling zone of the condenser is respectively; k1,c、K2,c、K3,cHeat transfer coefficients of a superheat zone, a two-phase zone and a supercooling zone of the condenser are respectively set;
the modeling of the evaporator comprises:
neglecting the heat exchange between the evaporator and the outside, and considering the flow of the refrigerant and the chilled water as a one-dimensional uniform flow, the heat exchange process in the evaporator is obtained as follows:
Qe＝mw,ecp,w(twi,e-two,e)＝mr(1-x)(hro,e-hri,e)；
Q1,e＝K1,eF1,eΔt1,e；
Q2,e＝K2,eF2,eΔt2,e；
Figure FDA0003436111530000034
Figure FDA0003436111530000035
wherein Q iseThe heat exchange capacity of the evaporator; m isw,eIs the flow rate of the chilled water; t is twi,eThe temperature of the inlet chilled water of the evaporator; t is two,eThe temperature of the chilled water at the outlet of the evaporator; t is tw1,eThe temperature of the chilled water at the inlet of the two-phase zone; h isri,eIs the evaporator inlet enthalpy; h isro,eIs the evaporator outlet enthalpy; x is the dryness of the refrigerant at the inlet of the evaporator; q1,eThe heat exchange quantity of the superheat zone of the evaporator is obtained; q2,eThe heat exchange quantity of the two-phase area of the evaporator; Δ t1,eHeat exchange temperature difference is carried out in an overheating area of the evaporator; Δ t2,eHeat exchange temperature difference of two phase regions of the evaporator; t is tro,eThe temperature is the suction temperature of the compressor, namely the temperature of the refrigerant at the outlet of the evaporator; t is tr,eIs the evaporation temperature;
the modeling of the throttle valve comprises:
the thermostatic expansion valve is formed by the pressure P of a temperature sensing medium in a temperature sensing bulbbProviding a valve opening force, by steam pressure PcAnd a spring force providing a closing force, the spring force being a minimum at Δ P when the valve is in the closed statemin(ii) a Displacement y and P of the valveb、Pc、ΔPminThe relationship between them is expressed as: k (P)b-Pc-ΔPmin) (ii) a k is the reciprocal of the spring coefficient;
the modeling of the water pump comprises:
the rotation speed ratio f of the water pump is defined as the rotation speed n of the water pump motor and the rated motor rotation speed n0The ratio, expressed as:
Figure FDA0003436111530000041
the relationship between the pump lift and the pump efficiency and the flow and rotation ratio of the water pump is expressed as follows:
Figure FDA0003436111530000042
Figure FDA0003436111530000043
wherein HpuIs the pump lift; m iswMass flow of the water pump; etapuThe efficiency of the water pump; h is01、h02、h03、h11、h12、h13Are fitting coefficients.
4. The method for diagnosing faults of a central air-conditioning system based on a stacking fusion algorithm according to claim 1, wherein in the step S2, the method for acquiring state data of the central air-conditioning system during normal operation and different faults through a plurality of sensors, and obtaining a sample data set after data preprocessing and feature extraction comprises:
acquiring the temperature of the inlet and outlet of the compressor, the temperature of the inlet and outlet of the evaporator, the temperature of the inlet and outlet of the condenser and the temperature of the inlet and outlet of the condenser at the time of normal operation and simulated fault by thermocouples arranged on the walls of the inlet and outlet pipes of the compressor, the water inlet and outlet of the evaporator, the water inlet and outlet of the condenser and the refrigerant pipe of the inlet and outlet of the condenser; the pressure at the inlet and the outlet of the compressor during normal operation and fault simulation is acquired by a pressure sensor arranged at the inlet and the outlet of the compressor; collecting cold water flow and cooling water flow during normal operation and simulated fault through flow sensors arranged on horizontal pipes at outlets of a cold water pump and a cooling water pump;
carrying out denoising processing, missing value filling, repeated invalid value deletion and normalization preprocessing on the acquired data;
taking the state characteristic parameters of the central air-conditioning equipment as independent variables and the fault label characteristics of the central air-conditioning equipment as dependent variables, extracting the characteristics of the preprocessed data variables by adopting a wavelet packet decomposition algorithm and a wavelet packet reconstruction algorithm, selecting the extracted characteristics according to a grey correlation algorithm, establishing a mapping relation between the state characteristic parameters of the central air-conditioning equipment and the fault label characteristics, and obtaining a sample data set; wherein the content of the first and second substances,
the fault signature includes at least: the flow of cooling water is increased or reduced, the flow of chilled water is increased or reduced, the water inlet temperature of a condenser is overhigh, non-condensable gas exists in a refrigerant, and the refrigerant leaks; the characteristic variables correspondingly selected by the fault label characteristics at least comprise condenser water inlet temperature, condenser water outlet temperature, evaporator water inlet temperature and evaporator water outlet temperature.
5. The method for diagnosing the fault of the central air-conditioning system based on the stacking fusion algorithm as claimed in claim 4, wherein the performing the feature extraction on the preprocessed data variable by using the wavelet packet decomposition algorithm and the wavelet packet reconstruction algorithm comprises: decomposing the preprocessed data variable by a wavelet packet decomposition algorithm into a plurality of wavelet values, reconstructing wavelet characteristics of the decomposed wavelet values by a wavelet packet reconstruction algorithm, and finally outputting a wavelet characteristic data set of the preprocessed data;
the selecting the extracted features according to the grey correlation algorithm comprises the following steps: and calculating the relevance values corresponding to the extracted features according to a grey relevance algorithm, sorting the relevance values corresponding to the extracted features in size, distinguishing the relevance degree between the state characteristic parameters of each central air-conditioning equipment and the fault label features, and taking the characteristic parameters with larger relevance degrees as sample data sets for fault diagnosis.
6. The method for diagnosing faults of a central air conditioning system based on a stacking fusion algorithm as claimed in claim 1, wherein in step S4, each base learner is trained by a k-fold cross validation method, and a prediction result of each base learner is obtained as a secondary training data set; when training each base learner, selecting a plurality of groups of different machine learning algorithms to combine, and generating a secondary training data set under a plurality of groups of combination modes, wherein the method comprises the following steps:
determining the number of the base learners as m, and randomly dividing the sample data set D into m data sets D with equal size1、D2、D3……DjDefinition of DjAnd D-j＝D-DjJ is 1,2,3, … …, m is j-fold test data set and training data set of k-fold cross validation, respectively, in training data set D-jGet the model on the training base learner
Figure FDA0003436111530000051
k 1,2,3, … …, m for test dataset DjEach sample, model of
Figure FDA0003436111530000052
Will output a test result; when the cross validation process is finished, obtaining the output result of the base learner on the whole training data set;
converting the obtained output result of the base learner into a probability type result, keeping the results of the m base learners within a [0,1] interval, and splicing the probability type output result and the training set labels to form a new training set as a secondary training data set;
when the m base learners are trained, the selectable machine learning algorithm comprises the following steps: the method comprises the following steps of (1) SVM, BP neural network, random forest, GBDT model, XGboost model, Light GBM model, linear regression model, support vector machine and weighted extreme learning machine; and according to the prediction performances of different algorithms, fixing the value of m, selecting various different algorithms to combine to generate various combination working conditions of the base learner or performing different values on m, and selecting different algorithms to combine to generate various combination working conditions of the base learner.
7. The method for diagnosing faults of a central air-conditioning system based on a stacking fusion algorithm according to claim 1, wherein in the step S5, inputting multiple sets of secondary training data sets into a secondary learner for training to obtain multiple central air-conditioning fault diagnosis models, including: sequentially inputting secondary training data sets obtained by training of all base learners under different combination working conditions into a secondary learner for training to obtain a plurality of central air conditioner fault diagnosis models; the machine learning algorithm used by the secondary learner is one of SVM, BP neural network, random forest, GBDT model, XGboost model, Light GBM model, linear regression model, support vector machine and weighted extreme learning machine.
8. The method for diagnosing faults of a central air-conditioning system based on a stacking fusion algorithm as claimed in claim 1, wherein in step S6, the method for diagnosing faults of a central air-conditioning system based on a stacking fusion algorithm comprises the steps of:
calculating the mean absolute error value MAE, the root mean square difference value RMSE and the fitting degree R of the model2The expression is:
as the evaluation standard of the central air-conditioning fault diagnosis model, the better the model performance is, the smaller the average absolute error value MAE and the root mean square difference value RMSE are, and the fitting degree R is2The larger;
Figure FDA0003436111530000061
Figure FDA0003436111530000062
Figure FDA0003436111530000063
wherein, yi、
Figure FDA0003436111530000064
Respectively the fact of the sampleA boundary value, a predicted value and a mean value; n is the size of the test data set.
9. The method for diagnosing the faults of the central air-conditioning system based on the stacking fusion algorithm is characterized in that a base learner in the double-layer stacking model is a weighted base learner, each base learner is endowed with a weight by introducing a weight formula based on a G-mean value, output results are corrected according to the weights, and the corrected results are fused into a secondary training data set;
wherein, the weight calculation formula of each base learner is as follows:
Figure FDA0003436111530000065
wherein alpha isiIs the output weight; GM (GM)iG-mean values for the ith base learner for the input sample set;
and the secondary learner in the stacking model is a secondary learner introducing an attention mechanism, when the n-dimensional input variable in the secondary training data set is input to the secondary learner at the time t, the weight of the n-dimensional features is calculated firstly, then normalization processing is carried out on the obtained weight to obtain the weight importance degree ratio of different features, finally, the obtained weight and the weight ratio are weighted to obtain a final feature vector, and the central air-conditioning fault diagnosis model is optimized and output.
10. A central air-conditioning system fault diagnosis device based on a stacking fusion algorithm is characterized by comprising the following components:
the digital twin model establishing module is used for establishing a digital twin model of the central air-conditioning system by adopting a mechanism modeling and data identification method;
the system comprises a sample data acquisition module, a central air-conditioning system and a data processing module, wherein the sample data acquisition module acquires state data of the central air-conditioning system in normal operation and different faults through a plurality of sensors, and acquires a sample data set after data preprocessing and feature extraction;
the stacking model building module is used for dividing a sample data set into a training data set and a testing data set, building a double-layer stacking model at the same time, and determining that the number of the base learners is m and the number of the secondary learners is 1;
the base learner training module is used for training each base learner by adopting a k-fold cross validation method to obtain a prediction result of each base learner as a secondary training data set; when each base learner is trained, selecting a plurality of groups of different machine learning algorithms for combination to generate a plurality of groups of secondary training data sets under a combination mode;
the secondary learner training module inputs a plurality of groups of secondary training data sets into a secondary learner for training to obtain a plurality of central air conditioner fault diagnosis models;
and the fault diagnosis model evaluation module is used for evaluating the prediction performance of the plurality of central air-conditioning fault diagnosis models through the test data set, selecting the model with the best prediction performance as the optimal central air-conditioning fault diagnosis model, and diagnosing the faults of the central air-conditioning system through the model.
CN202111614025.7A 2021-12-27 2021-12-27 Central air conditioner fault diagnosis method and device based on stacking fusion algorithm Active CN114484731B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111614025.7A CN114484731B (en) 2021-12-27 2021-12-27 Central air conditioner fault diagnosis method and device based on stacking fusion algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111614025.7A CN114484731B (en) 2021-12-27 2021-12-27 Central air conditioner fault diagnosis method and device based on stacking fusion algorithm

Publications (2)

Publication Number Publication Date
CN114484731A true CN114484731A (en) 2022-05-13
CN114484731B CN114484731B (en) 2023-05-16

Family

ID=81495832

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111614025.7A Active CN114484731B (en) 2021-12-27 2021-12-27 Central air conditioner fault diagnosis method and device based on stacking fusion algorithm

Country Status (1)

Country Link
CN (1) CN114484731B (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114754973A (en) * 2022-05-23 2022-07-15 中国航空工业集团公司哈尔滨空气动力研究所 Wind tunnel force measurement test data intelligent diagnosis and analysis method based on machine learning
CN115127192A (en) * 2022-05-20 2022-09-30 中南大学 Semi-supervised water chilling unit fault diagnosis method and system based on graph neural network
CN115512717A (en) * 2022-11-21 2022-12-23 东北石油大学三亚海洋油气研究院 Two-stage single-sensor pipeline leakage positioning method based on ensemble learning
CN115773562A (en) * 2022-11-24 2023-03-10 杭州经纬信息技术股份有限公司 Unified heating ventilation air-conditioning system fault detection method based on federal learning
CN116681291A (en) * 2023-08-02 2023-09-01 杭州小策科技有限公司 Wind control prediction method and system based on integrated model
CN116678072A (en) * 2023-06-05 2023-09-01 耀昶嵘相变材料科技(广东)有限公司 Fault processing method and terminal of central air conditioning system and central air conditioning system
CN116975502A (en) * 2023-09-21 2023-10-31 天津津轨汇海科技发展有限公司 Energy efficiency monitoring management system for electromechanical equipment of subway train air conditioning system
CN117232097A (en) * 2023-11-09 2023-12-15 上海轻环能源科技有限公司 Central air conditioner refrigerating station optimal control method and system based on self-learning fusion model
CN117704880B (en) * 2023-12-13 2024-05-07 中建-大成建筑有限责任公司 Energy-saving temperature monitoring method and system for evaporator

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000346496A (en) * 1999-06-09 2000-12-15 Hitachi Ltd Air conditioner for railway vehicle
CN102538143A (en) * 2012-02-06 2012-07-04 广东美的电器股份有限公司 Intelligent phonic search engine air-conditioning system and control method thereof
CN102563814A (en) * 2012-02-08 2012-07-11 广东志高空调有限公司 Cloud air conditioning system based on cloud computation technology
JP2015182821A (en) * 2014-03-20 2015-10-22 三菱重工業株式会社 Method for determining index of article temperature, temperature sensor, air conditioner, delivery vehicle, and stockroom
CN204880544U (en) * 2014-08-01 2015-12-16 三菱电机株式会社 Remote operation device of air conditioner
CN106547858A (en) * 2016-10-21 2017-03-29 珠海格力电器股份有限公司 Big data analysis method for air conditioning unit and air conditioning unit
CN207570077U (en) * 2017-12-05 2018-07-03 南京贝龙通信科技有限公司 Remote elevator air-conditioner control system based on GSM/GPRS communications
US20180238772A1 (en) * 2017-02-20 2018-08-23 Gray Matter Systems LLC System for abnormal condition detection using nearest neighbor
CN109766952A (en) * 2019-01-21 2019-05-17 福州大学 Photovoltaic array fault detection method based on Partial Least Squares and extreme learning machine
AU2020101535A4 (en) * 2020-07-28 2020-09-03 Guangdong Ocean University Fault Diagnosis Method for Ship Power System Based on Integrated Learning
CN112257779A (en) * 2020-10-22 2021-01-22 重庆中源绿蓝环境科技有限公司 Method for acquiring self-learning working condition parameters of central air conditioner
CN113203589A (en) * 2021-04-29 2021-08-03 华中科技大学 Distributed fault diagnosis method and system for multi-split air conditioning system

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000346496A (en) * 1999-06-09 2000-12-15 Hitachi Ltd Air conditioner for railway vehicle
CN102538143A (en) * 2012-02-06 2012-07-04 广东美的电器股份有限公司 Intelligent phonic search engine air-conditioning system and control method thereof
CN102563814A (en) * 2012-02-08 2012-07-11 广东志高空调有限公司 Cloud air conditioning system based on cloud computation technology
JP2015182821A (en) * 2014-03-20 2015-10-22 三菱重工業株式会社 Method for determining index of article temperature, temperature sensor, air conditioner, delivery vehicle, and stockroom
CN204880544U (en) * 2014-08-01 2015-12-16 三菱电机株式会社 Remote operation device of air conditioner
CN106547858A (en) * 2016-10-21 2017-03-29 珠海格力电器股份有限公司 Big data analysis method for air conditioning unit and air conditioning unit
US20180238772A1 (en) * 2017-02-20 2018-08-23 Gray Matter Systems LLC System for abnormal condition detection using nearest neighbor
CN207570077U (en) * 2017-12-05 2018-07-03 南京贝龙通信科技有限公司 Remote elevator air-conditioner control system based on GSM/GPRS communications
CN109766952A (en) * 2019-01-21 2019-05-17 福州大学 Photovoltaic array fault detection method based on Partial Least Squares and extreme learning machine
AU2020101535A4 (en) * 2020-07-28 2020-09-03 Guangdong Ocean University Fault Diagnosis Method for Ship Power System Based on Integrated Learning
CN112257779A (en) * 2020-10-22 2021-01-22 重庆中源绿蓝环境科技有限公司 Method for acquiring self-learning working condition parameters of central air conditioner
CN113203589A (en) * 2021-04-29 2021-08-03 华中科技大学 Distributed fault diagnosis method and system for multi-split air conditioning system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
王党琴: "暖通空调***故障诊断技术的应用分析", 大众标准化 *

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115127192B (en) * 2022-05-20 2024-01-23 中南大学 Semi-supervised water chilling unit fault diagnosis method and system based on graph neural network
CN115127192A (en) * 2022-05-20 2022-09-30 中南大学 Semi-supervised water chilling unit fault diagnosis method and system based on graph neural network
CN114754973A (en) * 2022-05-23 2022-07-15 中国航空工业集团公司哈尔滨空气动力研究所 Wind tunnel force measurement test data intelligent diagnosis and analysis method based on machine learning
CN115512717A (en) * 2022-11-21 2022-12-23 东北石油大学三亚海洋油气研究院 Two-stage single-sensor pipeline leakage positioning method based on ensemble learning
CN115773562A (en) * 2022-11-24 2023-03-10 杭州经纬信息技术股份有限公司 Unified heating ventilation air-conditioning system fault detection method based on federal learning
CN116678072A (en) * 2023-06-05 2023-09-01 耀昶嵘相变材料科技(广东)有限公司 Fault processing method and terminal of central air conditioning system and central air conditioning system
CN116678072B (en) * 2023-06-05 2024-05-10 耀昶嵘相变材料科技(广东)有限公司 Fault processing method and terminal of central air conditioning system and central air conditioning system
CN116681291A (en) * 2023-08-02 2023-09-01 杭州小策科技有限公司 Wind control prediction method and system based on integrated model
CN116681291B (en) * 2023-08-02 2023-11-07 杭州小策科技有限公司 Wind control prediction method and system based on integrated model
CN116975502A (en) * 2023-09-21 2023-10-31 天津津轨汇海科技发展有限公司 Energy efficiency monitoring management system for electromechanical equipment of subway train air conditioning system
CN116975502B (en) * 2023-09-21 2023-12-19 天津津轨汇海科技发展有限公司 Energy efficiency monitoring management system for electromechanical equipment of subway train air conditioning system
CN117232097B (en) * 2023-11-09 2024-02-20 上海轻环能源科技有限公司 Central air conditioner refrigerating station optimal control method and system based on self-learning fusion model
CN117232097A (en) * 2023-11-09 2023-12-15 上海轻环能源科技有限公司 Central air conditioner refrigerating station optimal control method and system based on self-learning fusion model
CN117704880B (en) * 2023-12-13 2024-05-07 中建-大成建筑有限责任公司 Energy-saving temperature monitoring method and system for evaporator

Also Published As

Publication number Publication date
CN114484731B (en) 2023-05-16

Similar Documents

Publication Publication Date Title
CN114484731A (en) Method and device for diagnosing faults of central air conditioner based on stacking fusion algorithm
Li et al. An explainable one-dimensional convolutional neural networks based fault diagnosis method for building heating, ventilation and air conditioning systems
Li et al. Data partitioning and association mining for identifying VRF energy consumption patterns under various part loads and refrigerant charge conditions
Fan et al. Feasibility and improvement of fault detection and diagnosis based on factory-installed sensors for chillers
Zhou et al. A comparison study of basic data-driven fault diagnosis methods for variable refrigerant flow system
CN110555479A (en) fault feature learning and classifying method based on fusion of 1DCNN and GRU
CN113792762A (en) Water chilling unit fault diagnosis method, system and medium based on Bayesian optimization LightGBM
Zhou et al. Data-driven fault diagnosis for residential variable refrigerant flow system on imbalanced data environments
Li et al. Review on fault detection and diagnosis feature engineering in building heating, ventilation, air conditioning and refrigeration systems
CN1201888A (en) System for monitoring outdoor heat exchanger coil
Li et al. Interpretation of convolutional neural network-based building HVAC fault diagnosis model using improved layer-wise relevance propagation
Li et al. Diagnosis for multiple faults of chiller using ELM-KNN model enhanced by multi-label learning and specific feature combinations
Zhang et al. Fault detection and diagnosis for the screw chillers using multi-region XGBoost model
Zhang et al. Deep learning in fault detection and diagnosis of building HVAC systems: A systematic review with meta analysis
CN112766327B (en) Air conditioner fault prediction method, electronic equipment and storage medium
CN114370696A (en) Method for controlling outlet water temperature of cooling tower of central air conditioner based on D-S evidence theory
Zhang et al. Imbalanced data based fault diagnosis of the chiller via integrating a new resampling technique with an improved ensemble extreme learning machine
Ding et al. New fault diagnostic strategies for refrigerant charge fault in a VRF system using hybrid machine learning method
Li et al. Machine learning based diagnosis strategy for refrigerant charge amount malfunction of variable refrigerant flow system
CN111723925A (en) Method, device, equipment and medium for fault diagnosis of on-road intelligent train air conditioning unit
CN112990258A (en) Fault diagnosis method and system for water chilling unit
Li et al. Across working conditions fault diagnosis for chillers based on IoT intelligent agent with deep learning model
Zhang et al. Integrated generative networks embedded with ensemble classifiers for fault detection and diagnosis under small and imbalanced data of building air condition system
CN113822359B (en) Heat pump system self-adaptive fault diagnosis method based on residual data scaling strategy
Li et al. Improved convolutional neural network chiller early fault diagnosis by gradient-based feature-level model interpretation and feature learning

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
PE01 Entry into force of the registration of the contract for pledge of patent right

Denomination of invention: A Fault Diagnosis Method and Device for Central Air Conditioning Based on Stacking Fusion Algorithm

Effective date of registration: 20230926

Granted publication date: 20230516

Pledgee: Guotou Taikang Trust Co.,Ltd.

Pledgor: Zhejiang Yingji Power Technology Co.,Ltd.

Registration number: Y2023980058937

PE01 Entry into force of the registration of the contract for pledge of patent right