Capabilities that transform manufacturing operations

Identify stations & machines that have critical predictions

Among the tens and hundreds of stations, machines and devices, the plant management can identify those stations that have potential for failure in the next few shifts and drill down to each machine and device to identify the probable reason for adverse prediction. The failure include machine failure, process failure and performance failure.

Get quantified score of productivity risks

Get a snapshot of predicted availability, performance and quality scores for each station and the trend for the next few shifts. Drill down to any station / machine to visualize the devices that cause adverse prediction. For every adverse prediction get probability of the event occurring, predicted time to occurrence, MTTR for the issue and confidence score.

Get optimized recommendations

Each adverse predictions are fed through an optimization model to derive the best time repair the asset. The results are consolidated to form an optimized maintenance plan. Based on the historical actions taken for an issue, prescriptions are made for the best course of action for any given issue.

Subscribe to prioritized predictive maintenance alerts

Plant Maintenance Supervisors get predictive maintenance alerts when the probability of machine failure is greater than set threshold. Each alerts come with a recommended time to repair based on optimization model results. Plant MEs are able to add notes, collaborate with technicians to assess the prediction results.

Create EAM ticket with click of a button

Two way integration with EAM system is possible. Plant Maintenance Supervisors are able to send EAM tickets after evaluating predictive maintenance alerts. They are able to monitor the status of the maintenance ticket throughout the maintenance workflow. The dashboard also gives consolidated metrics on the maintenance ticket status of all the work orders sent.

Do survival analysis and root cause analysis of assets

All the predictors responsible for prediction of machine failure can be analyzed over multi-variant graph. Survival analysis of an asset based on past predictions are also possible with graph of historic prediction data. Plant Maintenance Engineers can use this capability to decide on maintenance strategy for assets.

Monitor and analyze subset of processes for anomalies

Process Engineers are able to take a subset of assets and processes and analyze them for anomalies. System consolidates anomaly score and trends for the subset of assets and devices to enable process engineers take decision on process improvement.

Features that support the role of multiple plant personas

While the objective of PPA is to help the plant to achieve throughput to potential, the application enables plant managers, plant engineers, plant maintenance engineers / supervisors, process engineers and quality engineers to get unique, relevant predictive and prescriptive insight that aids their role in achieving the plant objective

Use ISA95 information model to roll-up prediction scores

PPA has ISA 95 information model implemented to represent the plant organization. This enables prediction scores to be rolled up from device level to station/machine level, line and plant level. Therefore, plant management can get quantified visibility into the predicted KPIs at plant / line / station level and quantified risks to productivity at various levels.

Pre-built advanced analytics pipelines

PPA has pre-built advanced analytical pipeline for early anomaly detection, anomaly scoring, regression/prediction, auto classification and optimization optimized for various use case and processes in discrete manufacturing plants. This reduces upfront data scientist’s effort and accelerates time to value.

How customers use it

  • Reduce production loss due to unplanned downtime

    Reduce production loss due to unplanned downtime

    Problem

    Unplanned downtime can be the most significant factor affecting overall equipment effectiveness. Estimates indicate approximately 89% of machinery failures are random, yet many organizations still practice "run to fail" or "time based maintenance."

    Solution

    Asset instrumentation, connectivity, IoT, and analytics capabilities provide visibility into unplanned downtime to significantly reduce impact through predictions that allow operations and maintenance to proactively address impending asset failure.

  • Reduce loss due to poor quality

    Reduce loss due to poor quality

    Problem

    Even in a thriving company cost of poor quality is estimated to be 10 to 15 % of operations. The earlier quality problems are identified in the manufacturing process the greater positive impact on overall equipment effectiveness and cost savings.

    Solution

    Identify when production equipment is trending away from prescribed calibration(s) and is in danger of producing defective parts. Identify the source of the problem and recommend steps to mitigate process variances.

Technical details

Software requirements

A workstation that runs one of the supported web browsers.

    Hardware requirements

    This solution is hosted in IBM's secure cloud data centers.

    • IBM provides cloud infrastructure as a service from data centers around the world.

    You may also be interested in

    IBM Visual Insights

    Identify defects via AI-powered visual inspection to improve quality.

    Learn more

    IBM Prescriptive Quality

    Identify process and product quality issues earlier

    Learn more

    See how it works

    Buy now and get started