The Studying Method Of The Predication And Trouble Diagnosis Of Refrigeration

The refrigeration equipment is applied to such a lot of trades and departments as industrial and mining enterprises, commerce, medical treatment, service trade, and etc extensively, it is a prerequisite of guaranteeing the environmental quality of artificial refrigeration that the refrigeration system is operated safely and reliably. The traditional way to overhaul regularly has presented unnecessary shutting down too much and overhauled, which has caused a large amount of waste of manpower and material resources; On the other hand, with the popularization and development in such new technologies as information technology , network technology , digital communication ,etc., the failure predication and diagnosis of the refrigeration system have already incorporated the modern control technology, and such system can maintain and reduce energy consumption from the system emerging in this way, lengthen equipment service life . Though this technology is developed to some extent in the refrigeration field, people's attention degree is not high, so it is not perfect enough. study always solves the problem, to the above situation, I do the following research respectively. This text use trouble tree analytical theory combine refrigeration system general structure set up trouble tree of refrigeration of system, the branch, leaf incident are analyzed from two respects of frequency and maintenance cost in order to state the influence degree to the head incident separately, in the trouble frequency , the higher trouble frequency is manipulative component and electric apparatus component, they are hard troubles; The compressor has the highest maintenance cost according to maintenance cost, and the compressor breaks down the great part is the hard trouble, to the hard trouble we analyse through mechanism and improve technology in the designing for manufacturing to avoid ; the other way round the soft trouble because of concealing and changing is the emphases to study. Through analyzing seven kinds of soft accident, I draw the characteristic and state parameter relation. To these seven kinds of situations, this thesis uses greyly models, grey related analysis and the neural network model of BP respectively to predict and diagnose. About the method of parameter prediction in grey theory, the grey predicting model is set up t with cold water temperature .Through examining, the precision is higher. Through analyzing to the present and historical signal that is gathered, thus realized to the prediction of the parameter variation tendency of some time of future, give an alarm ahead of time, help the early prediction of the trouble and diagnose and avoid the emergence of the accident. This text sets up diagnose models of many parameters of refrigeration system greyly and relatively, according to the related size of degree, the typical trouble stated in the past is an example, will wait to examine diagnosing the mode belongs to with the related degree of biggest standard trouble modes and first-class, related degree heavy, degree of closeness heavy, according to related degree being all kinds of to can judge where kind of troubles are appeared. The result indicates the diagnosing method is relatively accurate, suitability is stronger This thesis uses the neural network model of BP to diagnose the common trouble of seven kinds of refrigeration systems, and has trained to the standard sample data. Diagnose result indicate output of network is consistent as to already sample knowledge of study with the expectation, so it is effectual to diagnose the refrigeration system though the neural network.