Bayesian network method for fault diagnosis of civil aircraft environment control system

Abstract

This paper addresses the issues of fault diagnosis of the environmental control system of a certain commercial aircraft model of which the environmental control system has a high failure rate in field and causes many unplanned maintenance events. Because of the complexity and reciprocal compensation mechanism of the environmental control system, it is difficult to carry out fault isolation timely once the failure occurred during aircraft turnaround time, which thus may cause flight delay or even cancelation. The original contribution of this work is to propose a Bayesian network–based fault diagnosis method for commercial aircraft environmental control system where a multi-information fusion mechanism is used to incorporate the system first principle, expert experience and condition monitoring data. It incorporates extraction technology of sensor feature parameters and the structural learning of Bayesian network to realize the effective diagnosis of multiple faults. A case study is conducted based on a data set from a commercial aircraft fleet. The results show that the fault isolation ratio of this method is greater than 89%. The proposed Bayesian fault diagnosis network method can be used as a troubleshooting tool for airline maintenance technicians in fault isolation of environmental control system, reducing the time spent on-line troubleshooting and aircraft downtime.

Aircraft system-level diagnosis with emphasis on maintenance decisions

“…As regards system-level diagnostic studies that consider the simultaneous existence of multiple faults, Liu et al 21 developed a diagnostic technique that is able to diagnose multiple faults in an ECS. This is done by a Bayesian network, which is constructed based on first principles, expert knowledge, and maintenance data and is designed to detect faults in the following components: heat exchanger, air cycle machine, temperature control valve and ram air actuator.…”

“…According to Liu et al 21 the most frequent ECS faults appear in the Primary or Secondary Heat Exchangers (PHX – SHX), the TCV, and the ram air door actuator an the ACM. Furthermore, as it has been reported by industry experts, the ACM is the most frequently replaced component.…”

“…As regards the ECS, the most notable studies discussing multiple faults include Hare et al 22 who used neural networks to identify the simultaneous existence of heat exchanger, air cycle machine, and sensor faults. Also, as described previously, Liu et al 21 developed a Bayesian network capable of diagnose simultaneous faults in four components. Finally, Palmer et al 27 developed a technique that is able to diagnose air cycle machine degradation, pipe corrosion, and degradation in the temperature control valve.…”

Multi Sources Information Fusion Based on Bayesian Network Method to Improve the Fault Prediction of Centrifugal Compressor

“…Many works developed in this same reflection have been made, but each of these works has its particularity. [10] Has shown that information fusion by the Bayesian network applied to a control system of the commercial environment aircraft has allowed fault isolation at a rate higher than 89%. Other works concern the information fusion by Bayesian networks to improve the maintenance that has been developed in recent years in the use of reliability information, vibration signals, the current signals, or the sensor data [11][12][13][14].…”

Anomaly detection based on multivariate data for the aircraft hydraulic system

“…[5][6][7] Considering the application of PHM methods on the aviation, the complex flight environments and the multiple monitoring parameters are the huge challenge. PHM methods, such as anomaly detection, 8 fault diagnosis, 9,10 and prognosis 11 for aircraft, have been deeply studied based on hypothesis, simulation, or laboratory data, while the modeling and learning on actual flight data are still in the exploratory stage. On one hand, in actual flight collected by the airborne quick access recorder (QAR), the proportion of normal and abnormal samples is unbalanced that why most supervised methods are not available for the aircraft monitoring and diagnosis.…”