Tip:
Highlight text to annotate it
X
Predictive maintenance is all about predicting when a component will fail, or require service,
before that failure happens. How do you handle maintenance today? Reactively, replacing components
after they fail? Or scheduling, by replacing components on a pre-defined schedule? A predictive
maintenance solution goes well beyond that -- using historical data to predict when a
failure is likely to occur, so you can take action and avoid costly downtime.
The IBM solution is called IBM Predictive Maintenance and Quality. Here are the six
steps that make it work.
The first is loading master data. Before you can start making predictions you'll need to
load some master data into the analytic data store, which is a DB2 database. This master
data includes a list of all the devices to be monitored by the solution.
Let's use a scenario to illustrate this. Consider a manufacturing firm that's having difficulty
maintaining the power transformer that feeds electricity to their production lines. If
they wanted to predictively maintain the power transformer they would need to provide master
data about the transformer device itself.
Step two is loading and storing events. The analytic data store needs to receive events
from monitored devices. So for our power transformer we'd provide events for observing temperature
and current load measurements. The Predictive Maintenance and Quality solution receives
these events in real time or batch and uses IBM WebSphere Message Broker to transform
the message format sent by each device into a common format.
Third is performing aggregation. Events are aggregated into key performance indicators
using measurement types and profile variables. A measurement type defines how to interpret
a particular device reading (so a reading of "107" is understood to be a temperature
reading and not something else). An example of a profile variable is to calculate the
average temperature of the transformer and its current load.
The fourth step is scoring. This is where the magic happens! Predictive models are created
in IBM SPSS Modeler. These predictive models use historical data to determine the probability
of certain future outcomes. For example, we could create a model based on historical data
regarding transformer temperature, current load, and occurrences of failure. The score
that is returned can be thought of as an estimate of the likelihood that the transformer will
fail within a designated period of time, based on the most recent readings.
With the scores calculated it's time to start making decisions, That's step five - decision
management. With SPSS Decision Management, rules can be authored, tested, optimized,
and deployed. When a predictive score shows a particularly high probability of failure,
the action may be to transfer the load to another device and shut down the transformer
for inspection.
Finally the sixth step is dashboards and delivering recommendations. Communicating recommended
actions, such as to perform an on-site inspection, can be accomplished by creating work orders
in Maximo. The accumulated KPIs and current profile values (such as the average temperature
of the transformer) can be viewed in IBM Cognos Business Intelligence reports.
The entire IBM Predictive Maintenance and Quality solution is fully described in this
IBM Redpaper. The paper covers each of these steps in technical detail and is aimed at
anybody looking to understand the power of predictive maintenance and quality solutions.
Check it out!