PRODUCTS & SERVICES
Smarter data management
More intelligent approaches to remote data logging systems overcome big data transfer
limitations by avoiding the need for more capacity and demand on post-processing
Data logging systems are
used for on-road fleet testing
applications to store data locally.
Most loggers support standalone
operation, which includes local
intelligence to store data on events,
to perform calculations such as
rainflow classifications and also
to calculate data from a large range
of math and logical operations.
Today, measurement
configurations for thermo
management and powertrain testing
are composed of between 1,000
to 3,000 measurement channels.
Apart from the primary inputs
from the ECUs and analogue
module, a general rule is that 10%
new calculation channels are
implemented, including specific
statistics such as totalizer, averages
and min/max calculations.
Data loggers store a range of
di erent data formats, including
time-based data in various sample
rates from 1Hz up to 100 kHz, which
uses a lot of storage capacity
especially when it comes to large
channel count road load data
applications (RLDA) and high-speed
acoustic measurements. In addition
to this, event-based data from CAN,
LAN, and Ethernet tra ic is also
stored. As the tra ic data is rich
in content, and therefore high in
volume, the interpretation can only
be done when it is processed and
decoded with corresponding
description files on a server.
However, the largest amount of
internal storage capacity of a data
logger is taken up by recorded video
data. Even when the stream-based
data is stored in the H.264 codec,
rather than dedicated JPEG images
per frame, the data storage volume
is higher compared to both time and
tra ic-based data formats.
In a typical fleet management
application, approximately 10 to
20GB of data is collected over the
course of one day. Should a user
then store all this data then
downloading it all at once, either via
Ethernet, Wi-Fi or cellular networks,
it causes a bottleneck. It is common
for engineers to store as much data
as possible in order to double-check
full CAN tra ic and Ethernet PCAP
tra ic on top of the configuration.
The problem with this approach is
high data volumes with transfer
times taking several hours as well
as the post-processing system
requirements on servers, the CPU
and RAM resources.
With IPEcloud MDM, it o ers
features to process measurement
188 // July 2019 // www.electrichybridvehicletechnology.com
Loggers provide dependable data acquisition in hybrid and electric vehicles
data and is designed for large
volume data file management,
analysis and reporting. However, the
data processing and management
can only take place after it is
transferred with all timing delays
and the associated costs. When the
data is transferred, it is put through
a standard processing routine to
check for anomalies and fault codes.
As the amount of vehicle data
produced and collected has
increased over the last decade,
the old-fashioned mentality of
storing as much as possible is no
longer a viable method due to the
demands it puts on post-processing
resource. However, technology is
available today that avoids having to
add more capacity to data loggers,
instead it adds more intelligence.
The progress of intelligent data
analysis functions using MATLAB
tool boxes and MATLAB scripting
solutions, which are traditionally
processed on the server, can now
be executed on the data logger.
Therefore, the typical data
evaluation and interpretation can
take place on the logger side.
With the new generation of
IPEmotion RT data logger software,
users can deploy processing
intelligence to the front-end
acquisition system and save
significantly on both storage space
and transfer time.
A measurement data management system for data loggers organizes, processes, and evaluates test benches or large fl eet data
FREE READER INQUIRY SERVICE
To learn more about Ipetronik,
visit: www.magupdate.co.uk/PEHV
/www.electrichybridvehicletechnology.com
/PEHV