Predictive and cognitive combine to reduce maintenance time
Predictive maintenance analytic models and algorithms, enabled though instrumentation and connectivity of critical equipment, are allowing asset intensive organizations to identify impending equipment degradation or failure well in advance of the actual event and thereby proactively remedy the problem, often without impact to production schedules.
Some of the more sophisticated multivariate analysis capabilities help pinpoint source of the problem – possible generator failure caused by a combination of coolant leak and unusually high ambient temperature, and predict time to failure – within 72 hours based upon current operating conditions. Such prognostications can help maintenance personnel prioritize, plan, and schedule appropriate resources to effect repairs as efficiently as possible.
A wealth of maintenance and repair information in digital form
Many routine repairs and maintenance procedures are common knowledge to experienced maintenance personnel, retirement of senior personnel. Work orders, repair logs, and notes contained in enterprise asset management systems are also important sources of maintenance information. For newer equipment multiple sources of digital information – technical documentation, schematics, discussion forums, “how-to” videos, blogs, and discussion groups are becoming increasingly available. This wealth of digital information can be employed to effect better maintenance practices. The challenge is aggregation and analysis of relevant information to make it readily available in support of effecting repair/maintenance procedures. Here, cognitive capabilities can be applied to achieve this goal.
Cognitive capabilities consolidate, classify, and recommend
Cognitive capabilities can consolidate, analyze, and classify relevant digital sources of equipment-specific information to identify topics, methods, and procedures with the goal of recommending specific resources to maintenance personnel to assist in executing a specific repair or maintenance procedure. For example, a maintenance engineer encountering a fault code could query the cognitive system using the equipment model number and the fault code and be provided with the page in the technical document that explains the code, a video that shows how to reset and recalibrate the equipment, and a link to a discussion forum that explains potential problems (and solutions) in the event the engineer encounters difficulties.
Predictive and cognitive for optimized maintenance
The combination of predictive and cognitive analytics can optimize many facets of maintenance activities. Accurate identification of cause and time of predicted asset degradation or failure helps maintenance and operations determine potential impact on production schedules, scheduling and prioritization of maintenance activity, and resources needed to remedy the problem. Type of failure can then be given to the cognitive analysis system so that it can recommend the appropriate information resources to aid in timely and successful repair.
IBM IoT Equipment Advisor
IBM makes predictive and cognitive capabilities as described above available via the complementary capabilities of IBM Predictive Maintenance and IBM IoT Equipment Advisor. Together, these solutions help asset-intensive organizations improve equipment performance, accelerate time to repair, reduce maintenance costs, and implement more effective and efficient operations and maintenance strategies for critical assets.
Prepare for the benefits of predictive and cognitive
Predictive and cognitive have the potential to significantly transform the manner in which asset intensive organizations conduct maintenance activities. To take advantage of these capabilities, compare your current environment against this brief list of requirements needed to facilitate the transformation.
- Critical equipment is identified, instrumented, and currently producing operational data (if not see how to bring the benefit of IoT to older manufacturing equipment)
- A digital repository of operational data, including failures, for critical equipment exists and can be accessed and analyzed for development of predictive models
- Work orders, maintenance and repair data in a digital format are available, ideally residing in an EAM system such as IBM Maximo
- Sources of asset-specific technical, operational, repair, and procedural information in digital format have been identified to be analyzed and classified using cognitive capabilities to provide equipment-specific recommendations in support of maintenance and repair activities.
- If the potential of predictive and cognitive to improve maintenance practices is of further interest we invite you to explore the topic further via this solution brief.
- Learn more about how IBM Maximo Asset Management can help you manage your enterprise assets more effectively, contact your IBM representative or IBM Business Partner.