AspenMtell, AspenProMV, machine learning

Detecting Asset Failure Early for Food and Beverage Companies

September 24, 2020

In the first blog in this series, we talked about some of the challenges facing food and beverage manufacturers today, including how to improve product and production quality, meet strict food safety requirements and reduce unplanned downtime — all while managing the largely still unknown long-term impacts of the COVID-19 pandemic. We also discussed the critical need of companies in the sector to step up the pace of their digitalization transformation in order to effectively handle these issues, and any others, while leveraging opportunities spurred by technological innovations in the industry.


Predictive Maintenance for the Food and Beverage Industry 

In this second installment in the series, we take a look at how predictive maintenance can play an important role in helping food and beverage manufacturers improve production quality and reduce unplanned downtime. Our focus is the Aspen Mtell® solution, already proven with businesses in the energy, chemicals, mining, pharmaceuticals, food and beverage, and other industries, and its new Maestro features, with capabilities that go beyond preventative maintenance.

Using machine learning to recognize precise patterns in operating data that indicate degradation and impending failure before it happens, Aspen Mtell delivers the earliest (average 45 days in advance), most accurate (91%) warning of equipment failures and can prescribe detailed actions to mitigate or solve problems. For food and beverage manufacturers, more accurate failure detection (with a curtailment of over/under maintenance of assets) means:

  • fewer false alarms and safety incidents
  • avoidance or reduction in unplanned downtime
  • significantly lower repair or lifecycle costs (30% on average) 
  • increased net production output

With Aspen Mtell, producers can self-sufficiently and accurately predict time to failure so they know precisely WHEN and HOW a failure will occur and WHAT to do about it. This allows manufactures adequate time to schedule repairs vs. unavoidably address them.


Finding Agents Faster and More Easily with Maestro for Aspen Mtell

Maestro is a new, breakthrough collection of features within Aspen Mtell, which assists users in building prediction Agents. Maestro can save users a significant amount of time by helping them manage the three biggest barriers in successful model building: data selection, data cleaning and creating context by incorporating domain expertise.  Combined, these three tasks can consume up to 50% of users’ time.

Aspen Mtell is equipment and process agnostic, which means it analyzes at both the asset and process level. The software looks upstream and downstream of equipment to determine potential failure. This means it does not depend on traditional measurement and trending of vibration and other parameters to identify degradation and functional failure. In fact, when analyzing equipment failure in this way in most cases the damage is already done. Therefore, an immediate cost savings opportunity exists in our client’s digitalization strategy because they can forgo adding sensors and third-party monitoring. 

Two types of “Agents” drive Aspen Mtell – Anomaly and Failure Agent. These Agents are designed to do more than only Anomaly detection to determine “normal” behavior. This is because they detect the actual behavioral patterns that begin early in root cause conditions and lead to specific failures.

Because these patterns are not typically unique to a single piece of equipment, Aspen Mtell is able to learn on one asset and share its learnings with many similar assets via transfer learning. This sharing of learning means Aspen Mtell is able to rapidly scale across a manufacturing site or enterprise (i.e. 30 assets in 30 days). Its scalability and rapid time to value make Aspen Mtell a game changer for users looking to accelerate their digitalization journeys. 

Other benefits of Aspen Mtell:

  • It can be deployed quickly, with an average installation time of four weeks. 
  • It is simple to learn and use.
  • Generally, users can be independently monitoring assets and building “Agents” in two to three days. 

In addition to its many visible benefits, there are still other benefits of using Aspen Mtell, perhaps less visible to your organization, including:

  • transparency
  • sustainability
  • cost savings
  • supportive of corporate and investor goals
Keep your eyes out for the next and final installment of our three-part series on digitalization for food and beverage manufacturers. In the meantime, check out our new video on Maestro for Aspen Mtell.




There was a problem storing your subscription

Leave A Comment