Reliability Data Analysis Case Studies

Large Pharmaceutical Plant

A large pharmaceutical plant had to dispose of an expensive batch of intermediate products because it couldn’t prove to inspectors that temperatures had been maintained during the batch production process. In that case, they actually had the data, but they couldn’t extract it in time to satisfy their inspectors. Millions of dollars were lost on just one batch! How could they avoid having that happen again?

They knew they had problems in the process and data collection equipment. To solve it they had considered initiating a Reliability Centered Maintenance (RCM) project but it was going to be more effort and time consuming than they wanted. Their reliability engineer (now our consultant) had extensive experience in military Failure Reporting and Corrective Action System (FRACAS) programs based on MIL HDBK 2155. FRACAS is a system for reporting, classifying, analyzing failures, and planning corrective actions in response to them. It’s designed for solving problems like this in operation plants and systems.

He input and analyzed plant data using a database he had created specifically for that purpose and used elsewhere before. Various failure codes, compliant with ISO 14224, were used to classify data points. Data consistency checks were carried out. Various recurring breakdowns were identified, revealing problems in devices and systems that had previously been unknown. The repeat offenders (bad actor equipment) were dealt with. Also from the data, it was possible to spot a number of procedural shortcuts that were increasing risks and putting employees in danger. From those, the need for additional bench testing and calibrations were identified and risks were reduced.

A number of data inconsistencies, and “holes” (incomplete records) were identified. Knowing where those occurred, the plant could address their data collection methods to improve data quality in the future. Some of those problems were also resolved in interviews with technicians. Although their data input wasn’t good, their memories could close some of those information gaps. More problems began to emerge from the additional cleansed data and could now be resolved.

Large Refinery

A large California refinery appointed a new Reliability Engineer to chair their “Failure Analysis Team” (FAT).

The engineer (now one of our consultants) had done considerable work for a large defense contractor using their FRACAS (Failure Analysis and Corrective Action System). He combined the structure and coding found in ISO-14224 (which arose for the oil and gas industry) with FRACAS concepts (from defense), and built a database of failures that fit the ISO structure based on a variety of previous projects.

With “Superuser Rights” on the computerized maintenance management system (CMMS) he was able to work with work-order data. He compared the plant data with records in the database he had created. That comparison revealed a number of observations on the current PM program, and on the maintenance corrective actions being taken. Ineffective PMs were identified. Corrective actions taken could be improved upon.

Reports that were generated for review and customer team discussion/action included:

• Various charts showing data relationships
• New work orders analyzed this period
• All work orders analyzed by equipment ID including any that saw action since last report
• Work Orders with Unknown Causes – these require discussion with your team and possibly deeper digging
• All Work Orders By Codes
• Data entry problems – e.g.: WOs that were written against a room, facility, or too high in the hierarchy as opposed to a specific productive maintenance significant asset from the hierarchy