The Bukit Panjang Light Rail
Transit (BPLRT) operates on a three-Phase 600-VAC traction power system. The
light rail vehicle (LRV) run on elevated guideways along a centrally-mounted power
and ATO rails shown in Figure 1. The rails are arranged in a five-conductor
rail configuration, specifically phases A, B, C, ATO/Earth Rail, and Earth Rail
/ATO. Each LRV has two sets of current collector assemblies (CCAs), each of
which has eight sets of current collector devices (CCDs). The CCDs may be
misaligned and dislodged due to operational wear, which can lead to unexpected
inter-phase shorts and consequent power supply trips and service disruptions.
Figure 1. Five-conductor power and ATO/Earth rail arrangement and CCDs.
The objective of this project is
to detect early signs of CCD abnormality by using a non-intrusive condition
monitoring system based on a patented inductive coupling technology. The proposed
system monitors the electrical contact between CCDs and power rails when the
LRV is moving under its usual operation condition. By injecting a well-defined
signal of known frequency and magnitude,
and monitoring the same signal in real-time, it is possible to ascertain
the quality of electrical contact between CCD and power rail and, as a
consequence, the condition of CCD.
As a proof of concept, several trials
have been conducted on the BPLRT track. The results collected have the ability
to differentiate new, used and worn CCDs, as shown in Figure 2.
Figure 2. Received signal level versus time for new, used, and worn CCDs.
Figure 2 shows that the variation
of the received signal level increases consistently from top to bottom, with
the new and worn CCDs exhibiting the lowest and highest variation, respectively.
Consequently, the proposed condition monitoring system has the ability to
detect the deterioration of CCDs due to wear and aging.
Further analyses of spread of the
sampled data from the trials allow us to quantify the CCD condition in terms of
the statistical standard deviation (s)
of the data collected, as illustrated by the green, yellow, and red bands in Figure
3 for the new, used, and worn CCDs.

Figure 3. Collected data standard deviation (s) and CCD condition for new, used, and worn CCDs.
According to Figure 3, s can be used a parameter
for threshold setting for the replacement of CCD before it deteriorates
further. In conclusion, the proposed condition monitoring system is capable of
identifying CCD and power rail deterioration caused by wear and aging. Discussions
are under way to consider the possibility of mounting the system on a BPLRT train
on a 6-month comprehensive field trial to both refine the system and to make it
even more robust in identifying threshold levels for predictive maintenance.