IMC-2016 Reliability Engineering for Maintenance - 39:54
by Rick Crory, Crory & Associates, Inc. and Samuel Paske, Metropolitan Council Environmental Services
Despite impressive gains increasing plant reliability in most industries, organizations still struggle with the raw material of reliability improvement - failure data. This paper uncovers two linked reasons: Poorly understood benefits and excessively high costs. The benefits of collecting and acting on good reliability information are not clearly understood. Similarly, the right level of information has not been established. Reliability data can be so detailed that it requires significant effort to create and manage, or it is so high level that it offers little actionable insight. There seems to be no middle ground of both actionable insight and reasonable effort.
Until now. Starting with clearly identified benefits, businesses can clearly make the case to staff and management for the extra effort good failure data requires. Businesses are concerned with 3 things related to physical assets: performance, risk and cost. Good failure data can be used to maximize performance, manage risk and control costs by optimizing maintenance activities and identifying assets for capital investment. In both cases, spending is used to sustain plant profitability rather than the other way around.
Currently most organizations chose to align collection of failure data to each asset class/type. This results in tens if not hundreds of failure classes/codes when in truth assets typically fail for one of three reasons; electrically, mechanically or structurally. Additionally, best in class CMMSs tend to tie failure data to the asset (e.g. problem, cause, remedy) ignoring component which is key to determining the actual failure mode for reliability analysis. In order to resolve this gap organizations incorporate components into the failure data further ballooning the amount of data needing to be created and thus maintained. The end result with so many selections to chose from maintenance workers in the field are overwhelmed and in many cases pick what ever values come up first thus incorrect data is collected. This vast amount of data impacts mobile applications as well when trying to download thousands if not tens of thousands of codes onto a handheld device. Bottom line to the organization is high cost due to vast amounts of data creation/maintenance, incorrect/inaccurate data to engineers and valuable time lost in getting this data downloaded for use in the field.