276 - Driving Asset Answers with Suspect Data with Manjish and Yanpei

Rooted in Reliability: The Plant Performance Podcast

Driving Asset Answers with Suspect Data with Manjish and Yanpei

Welcome our guests to the podcast. We are going to discuss Asset Answers in today's episode. However, briefly about our guests:

Yanpei has a background in telecom and smart grid. Currently is working as a product leader at Asset Answers.

Manjish is the service lead for Asset Answers and has been in the maintenance and reliability for 15 years building CMMS and analytic software on top of the CMMS)

In this episode we covered:

• Many organizations have data but not information. Have you witnessed such?
• What are the main causes of suspect data?
• Does the data have to be perfect to start?

Many organizations have data but not information. Have you witnessed such?

In many industries. They do not have information on how a failure was detected, what kind of failure it was and, was it a people, system, or process related failure?

What are the main causes of suspect data?

The causes include:

1. The people- incomplete data, or inaccuracies in recording. The system might not be user-friendly if it has many (perhaps 100) failure modes. It might be too tedious to use.
2. The system Mismatch of what the tool is designed for and what its current use at the site. The users later realize there aren't the right fields that suit their use
3. The data collection process is not standardized such that it varies from site to site in the same organization.

The sites might have different functional location hierarchies I.e. areas, sites, systems, units such that: what might be a unit is considered a system in another.

Does the data have to be perfect to start?

No it doesn't. The solution we provide is to focus on the purpose of the data. You can calculate key performance metrics using the small available data.

How do organizations know the specific data points needed?

First, understand the organization's goals– perhaps to optimize resources or reduce equipment failures. Determine:

• What metrics are going to measure performance e.g. cost of maintenance, failure count, mean time before failure etc.
• Understand the quality of the data e.g. how many work orders have costs on them?
• Take a parallel approach i.e. While working to improve the data, use the good available data and measure pertinent KPIs.

Can we still gain insight into an asset using things like the Law of Big Numbers?

You can work backwards once you have decided and defined the purpose of the data. You do not need absolute numbers. For instance, measuring the value of allocating resources on working on one equipment versus another.

How do organizations select the right asset to work on once they have good data?

There are multiple ways of selecting a bad actor. For instance:

At Asset Answers, we have peer benchmarking analytics to compliment the internal view of key performance metrics. It provides an external view of the businesses in the same industry producing the same product.

How does peer benchmarking analytics work?

Asset Answers has been built by:

• Collection of maintenance data, CMMS data, and associated work orders for over 10 years
• Used the data to standardize all the codes like functional location hierarchy across different companies.
• Doing a data quality analysis
• Identifying the metrics upon which decisions can be made with high confidence given the available data
• Measure the metrics
• Providing the company's versus peer benchmarking view of all the metrics.
• Identify opportunities to focus on given current industry practices

It is basically the aggregation of data to analyze it at different levels (site, area, equipment class etc). The biggest challenge is standardization. It is an easy way to find bad actors.

What action drive improvements on identified bad actors?

Asset Answers tells you where to focus on. Actions depend on the asset strategy in place. We provide another tool called PM Diagnostics where both instances of reactive and proactive are counted.

The two are sketched on an axis (proactive maintenance on X-axis and reactive on Y-axis) to create 3 categories:

1. Insufficient proactive maintenance- assess the current asset strategy especially in terms of cost.
2. Suboptimal proactive maintenance- doing a lot of proactive maintenance yet too few failures. Is it necessary?
3. Ineffective proactive- doing a lot of proactive maintenance but experiencing a lot of failures. Identify failure modes not being addressed.

Can we calculate ROI for all these using Asset Answers?

We installed an equipment tracker that stores data on historical (before) and after improvement activities to show ROI.

Does it consider downtime avoidance?

Depends on the data being tracked in the CMMS and its quality. Can be used to measure maintenance versus downtime cost.

Key takeaway

• Data quality issues is normal and should not stop the use of available data in making practical decisions
• Peer benchmarking analytics provides perspectives essential for resource management.

Eruditio Links:

• Eruditio
• HP Reliability
• James Kovacevic's LinkedIn
• Reliability Report

Manjish and Yanpei Links:

• Manjish Naik Linkedin
• Yanpei Chao
• Asset Answers – GE Digital
• Past epsiode with Manjish
• SMRP.org
• SMRP 2021 Conference
• Maintenance Reliability Metrics/KPI’s 101 Keeping it Simple

Download RSS iTunesStitcher

Rooted In Reliability podcast is a proud member of Reliability.fm network. We encourage you to please rate and review this podcast on iTunes and Stitcher. It ensures the podcast stays relevant and is easy to find by like-minded professionals. It is only with your ratings and reviews that the Rooted In Reliability podcast can continue to grow. Thank you for providing the small but critical support for the Rooted In Reliability podcast!

The post 276 – Driving Asset Answers with Suspect Data with Manjish and Yanpei appeared first on Accendo Reliability.