Improving Uptime and Operational Performance with LTTS Avertle

Summary

A digital twin is a dynamic virtual representation of a physical entity using real-world data.  The most common application uses real-time data and analytics for predictive maintenance (PdM) to prevent unplanned downtime for critical assets.  PdM improves business performance including revenue, asset longevity, and maintenance costs.

As ARC Advisory Group learned in a recent briefing, the L&T Technology Services (LTTS) Avertle solution is designed to deploy PdM in manufacturing and other industrial environments.  Its edge computing is well suited for high-speed processes and conveyor systems.  Avertle’s case stories demonstrate a proven track record, making it a viable option for PdM applications.

Predictive Maintenance Benefits and Project Justification

Unplanned downtime negatively impacts the typical metrics for maintenance.  These include asset uptime, asset longevity, cost control, and safety.  Improvements start with reducing unplanned downtime, such as when equipment stops working at the most critical time while it is making product for sale. 

Digital Twin

Research indicates that preventive maintenance and periodic inspections are usually ineffective.  This is because 82 percent of assets have a random failure pattern.[1]  Digital twins monitor asset health continuously and warn when it deteriorates so the asset can be repaired before failure occurs.  Many of these digital twin applications for PdM have evolved to include monitoring operational performance of an asset (throughput, quality, yield).

Getting Executive Support

Your CEO and other senior executives focus on their metrics like revenue and return on assets (ROA).  Unfortunately, most project justifications for predictive maintenance focus on reducing maintenance costs, which often is not adequate to get executive attention and support.  For nearly all C-level executives, their key performance indicators (KPIs) are in the profit and loss (P&L) statement and balance sheet.  PdM helps prevent downtime, which increases operating time, production volume, revenue, and profits.  The justification for a PdM project should start with the increased revenue.

High-speed Processes and Edge Analytics

Several industries have high-speed processes where the data acquisition and associated analytics require edge computing to help prevent unplanned downtime and improve operational performance.  A cloud-only solution becomes impractical due to data communication capacity limits, slow response time, network delays, and/or outages.  Examples of high-speed processes are found in:

• Discrete industries like automotive and CPG
• Food and beverage packaging lines
• Mining and metals conveyors and processing

• Pumps in water and wastewater plants and distribution/collection systems

Avertle Briefing from LTTS

Recently, LTTS executives briefed ARC on its Avertle solution for condition-based maintenance (CBM).  As we learned, the solution includes data acquisition, analytics, visualization, and alerts. 

Avertle Architecture and Key Functions

The Avertle solution has these principle components:

• Edge Gateway device for data acquisition and processing.  This includes prebuilt fault signatures and edge analytics to identify anomalies.  It incorporates both first principle models based on the machine’s design and pre-built machine learning models for various equipment types.  It provides high-speed processing and near-term failure prediction at the edge.
• Analytics Server (cloud or on-premise) filters the edge failure predictions to identify operating anomalies worthy of an alert.  It also uses artificial intelligence to tune the models using real operational data.  The analytics server has prebuilt fault signature libraries for over 35 asset classes like electric motors, chillers, diesel generators, and HVAC.  Each fault signature contains a multi-variate model used in machine learning algorithms.  The analytics server provides more in-depth analysis for improved failure and performance prediction to reduce false positives and provide more advance notice of an issue.
• Visualization services for assets including alarms, events, current status, and an asset health index.  Its equipment and plant asset hierarchy can be imported from another asset management application and used to navigate and display an asset’s condition. 

Typical User of Avertle

An alert typically needs some level of evaluation to determine an appropriate action.  Several questions need to be answered.  Is it valid?  Does the operator need additional training?  Do we have bad materials?  Perhaps the process control system needs adjustment.  If it is a maintenance issue, is an electrician, mechanic or other skill needed?  This type of evaluation is usually performed by a reliability, process, or control system engineer who would typically use Avertle to review alerts and determine the next steps.

Case Examples

Avertle has successfully reduced unplanned downtime as shown in these customer case stories provided by LTTS.

Bottling Lines In the USA

A leading food and beverage company had excessive unplanned downtime for the filling line in its bottling plant.  This is a critical set of assets since downtime directly impacts plant performance, KPIs, and revenue.  The existing preventive maintenance strategy was not effective and the decision was made to move to predictive maintenance using Avertle with 24x7 monitoring. 

Initial success led to a rollout across eight plants in US for 57 assets.  The asset types include blow molder, filler & capper, filler & seamer, labeler, packer & palletizer, ammonia chiller, and compressor.  The “small data” approach acquires a manageable set of parameters, including current, temperature, and vibration sensors for the specific assets being monitored.

Early failure detection with edge processing reduced data congestion and latency.  Increased uptime of the production lines provided a 3 percent increase in production (cases handled).  Also, maintenance costs were reduced by 7 percent.

Bottling Lines in Europe

A leading global beverage maker needed to reduce the unplanned downtime of critical assets at the filling station.  A failure among the conveyor motors caused an average production downtime of 90 minutes.  In addition to reduced production, the downtime impacted the process which led to product losses due to contamination of the liquid. 

Avertle was applied in filling stations for the three bottling machines.  Sensors on the filler conveyor motor, spray pump, and delivery belt provided current, temperature, and vibration to monitor equipment health.  Analytics on the edge for reduced data congestion and latency enabled early failure detection.

The Avertle solution reduced unplanned downtime for these critical assets and avoided production losses with improved product quality and bottling volume.  Benefits included improved asset uptime from 80 percent to 93 percent, reduced maintenance cost by 15 percent, and improved production.

Conclusion

Unplanned downtime for critical assets has significant negative business impact on revenue, on-time shipments, asset longevity, and maintenance costs.  Effective PdM helps prevent unplanned downtime and improves business performance while also increasing executive KPIs. 

The typical LTTS Avertle digital twin applications deploy PdM to help prevent unplanned downtime.  As we learned, its edge computing is well suited for high-speed processes and conveyor systems.  After installation and some artificial intelligence learning, Avertle applications have achieved 85 to 90 percent of valid alerts.  These alerts usually go to engineering to determine an appropriate next step.  Avertle has a proven track record and should be considered for preventive maintenance applications.

ARC Advisory Group clients can view the complete report at  ARC Client Portal 

If you would like to buy this report or obtain information on how to be-come a client, please  contact us

Keywords: Predictive Maintenance, Edge Analytics, Avertle, L&T Technology Services, ARC Advisory Group.