An inverse fault detection and diagnosis (IFDD) strategy for practical application on chiller product
International Journal of Refrigeration(2022)
Abstract
Current research on chiller fault detection and diagnosis (FDD) is usually based on the multi-classification method for mixed multiple fault types. The establishment of a multi-classification FDD model requires obtaining extensive test data of normal operation and multiple common faults at different fault levels, so the cost-to benefit ratio will not be accepted. In addition, because of the information interference between different types of fault data, FDD model is hard to avoid over-fitting or under-fitting, and there may be an intolerable high false alarm rate existing in product application. Therefore, it is difficult to promote FDD in actual chiller product applications. This paper proposes an inverse fault detection and diagnosis (IFDD) strategy, which is based on the characteristics of the target fault (the fault to be concerned) to establish a two-classification fault diagnosis model. Compared with the multi-classification method, the amount of fault test data required by the two classification methods to realize fault diagnosis can be greatly reduced. With IFDD strategy, the detection and diagnosis are completed simultaneously, while fast calculation speed, high diagnosis accuracy rate and low false alarm rate are achieved. The core of the IFDD strategy is to concentrate resources to capture the characteristics of the target fault, and no longer classify other types of faults and normal operating conditions, so it can avoid the influence of information interference between multiple types of fault data in the fault diagnosis model. In order to verify the feasibility of this strategy, a 200 ton variable-speed screw chiller has been tested on the laboratory, and a database of normal operation, refrigerant leakage, refrigerant overcharge faults, and a new operating condition of insufficient cooling water flow is initially established. Based on the sensors configured by the chiller unit and using the exponentially weighted moving-average (EWMA) control chart for IFDD, the model's fault diagnosis accuracy for the target fault reaches 100%, and no false alarm appears after the transition period.
MoreTranslated text
Key words
Inverse,FDD,Strategy,Practical application,Chiller product,EWMA
AI Read Science
Must-Reading Tree
Example
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined