Research On Performance Diagnosis Method Of Central Air Conditioning System Based On Feature Recognition Model

Air conditioning systems often cause faults due to the aging of equipment,natural wear,unreasonable design,unreasonable control or inadequate maintenance,resulting in deterioration of the system operating efficiency and indoor air quality.Air conditioning system faults are generally divided into hard fault and soft fault.The hard fault is a fault that causes a device or instrument to fail completely.The soft fault is the performance fault,which means a fault that causes the performance of the instrument and the instrument to gradually decrease and partially fail.Performance fault are generally generated gradually and are not easily noticeable,but their impact on system performance is long-term and harmful.Therefore,the study of performance fault diagnosis and evaluation of air conditioning system is of great significance to maintain efficient and energy-saving operation of the system and reduce operation and maintenance costs.Based on ASHARAE 1043-RP data,this paper establishes a simulation model of air conditioning system based on feature recognition,and uses this model to study the diagnosis method of chiller performance fault and transmission system performance fault.The performance fault diagnosis model of the air conditioning system can be established by using a small amount of measured data without the need to obtain the structural parameters of the air conditioning system and a large amount of measured data.On-line fault detection and diagnosis of chillers and transmission and distribution system are achieved.The accuracy and reliability of the feature recognition model and performance fault diagnosis model established in this paper are verified by the measured data of the central air conditioning system of a project.The main work of the thesis is as follows:(1)The simulation model of air conditioning system based on feature recognition is established.Through the theoretical analysis and derivation,the unknown structural parameters of each component of the air conditioning system are lumped to form the characteristic parameters,and a simulation model of air conditioning system based on feature recognition is established by using a small amount of measured data to obtain the characteristic parameters.The model needs no structural parameters,has good generality and high prediction precision,and can be applied to air conditioning system performance fault detection and diagnosis.(2)The strategy of fault diagnosis of the chiller plant based on feature recognition model are put forward.In this paper,based on the feature recognition model,the deviation degree and trend of each performance index of chillers under fault are studied,and the influence mechanism of the fault on each index is analyzed.The box-line diagram is used to study the distribution of the deviation degree of diagnostic indicators under different working conditions,and the fault judgment rule table and fault determination strategy with clear physical meaning and high diagnostic accuracy are determined,and the fault detection and diagnosis process of the chiller based on the feature recognition model is formed.And then the success rate of the fault diagnosis strategy in this paper is verified by using the ASHRAE 1043-RP fault data.It is found that the diagnostic success rate can reach 88%-100% after reaching fault level 3.The overall distribution of chillers' performance indexes(refrigerating capacity or COP)with the main diagnostic indexes was studied,and the deviation degree of each diagnostic index was determined when the chillers' performance declined to a certain extent,which was taken as the threshold of fault diagnostic indexes of chillers.(3)A fault diagnosis strategy of transmission and distribution system based on feature recognition model and pattern search optimization algorithm is proposed.Combining the pattern search optimization algorithm and the air conditioning system simulation model to optimize the operating conditions of the air conditioning system,compare the chilled water,cooling water flow and system comprehensive efficiency COPs of the current operating conditions and optimal working conditions,and diagnose performance faults caused by mismatched flow of the distribution system.The study found that under typical working conditions,the COPs can be improved by more than 10% through fault diagnosis of the transmission and distribution system and timely adjustment of the transmission and distribution system to eliminate the performance fault.(4)The proposed air conditioning system performance fault diagnosis method is used to diagnose the performance faults of an actual central air conditioning system chiller and transmission and distribution system.And the effectiveness of the method in the actual air conditioning system performance fault diagnosis is verified.