Improving Golf and Business With Predictive Asset Maintenance
Brittany Detamore 270006ARTU email@example.com | 2013-07-23 11:05:37.0 | Tags:  predictive-analytics ibm-cognos big-data predictive-maintenance analytics ibm-spss | 1 Comments | 9,417 Visits
Guest post by Ishan Sehgal, Program Director, IBM Predictive Maintenance and Quality Product Management
Follow Ishan on Twitter @ishansehgal
As I was still contemplating how after hitting such a great drive, that my subsequent approach shot missed the green, my playing partner reminded me how almost every time I had a GIR, I made a par or better. It occurred to me that while I may not have wanted to hear it, I could not deny that the chances of me winning this hole were looking bleak. While it would be nearly impossible for me to process much more data that can fit on my relatively neat scorecard, the insight was evident – I needed to work on my irons.
Data in an industrial environment has evolved from the days of hand managed clipboards. Machines and processes have become instrumented with thousands of events being measured and captured every second. Combined with associated environmental data such as operator, weather, supplier data, it is sometimes difficult to determine where to get started.
Ironically it can be this abundance and sometimes messy data that can help make predictive analytics in an industrial environment especially compelling. I also see an increasing acceptance of data correlation in helping mangers in taking proactive actions to improve asset and production process performance. With the proper application of the aptly termed ‘big data’ technologies and real-time analytics, industrial customers are now in a position to leverage the convergence of machine instrumentation, increased enterprise data accessibility and access to powerful statistical modeling.
Predictive asset maintenance is based on the notion that increasing amounts of data are being generated about the performance of equipment and systems, but often this data is only being used for short term reactive needs. This data is discarded soon after its initial generation; however, in aggregate, this data can provide a rich reserve for data mining purposes in unearthing patterns and trends for predictive analysis. Producing predictive models from this amalgamation of industrial data is only of value if ways can be found to act upon the predictive insights gained such that they can yield an operational benefit.
For analytics to be operationalized, data integration capabilities should be tightly coupled with analytic tools. These predictive, descriptive, and prescriptive tools along with data curation techniques are required. Select data is staged and aggregated in a centralized relational data analytics store where an asset’s pre-calculated Key Performance Indicators (KPIs) are persisted in summary profiles that enable real-time analytics.
IBM Predictive Maintenance and Quality (PMQ) brings all this together in a single, packaged software offering. IBM SPSS software (predictive analytics) and IBM Cognos software (business intelligence) provide the necessary analytics while IBM data tools from Websphere, DB2, and Information Management wire these capabilities to a real world industrial data environment.
PMQ enables predictive insights to become prescriptive actions that can be routed in a variety of ways: They can be viewed as a report, delivered as an email, streamed to a mobile device or even sent back to an enterprise asset management system, such as IBM® Maximo Asset Management (EAM), where work orders can be launched based on predictive insights.
Let me know what business problems you are looking to address with this capability – be sure to include any advice for me and my analytics-minded golf foursome!
- Check out this webinar on how Israel Electric Corporation utilized the technologies to predict asset failure
- See if this golf fix helps you – I found it useful
For more on this topic, join a live videochat Friday at 2pm ET http://ibm.co/BDBytes