Several user discussion groups took place during the 40th Turbomachinery Symposium in Houston, Texas. For example, one was held on the subject of centrifugal pump operation, maintenance and reliability. Few users, it turned out, were conducting risk-based analysis. In fact, there appeared to be confusion as to what was a risk-based approach.
One user complained that he spends 18 months planning an outage and then a few weeks before the outage, a problem crops up which the company is not prepared for. Another user observed a conflict between an OEM’s recommendation to conduct work at 30,000 hours, while risk models might suggest that 45,000 hours would be better.
A survey of the room revealed that compressor maintenance intervals varied from three years up to 13 years. One user said he had moved overhaul intervals from eight to ten years. However, he backs this up with thorough steampath audits whenever a machine comes down. The company goes to great lengths to find and fix anything that might be wrong when it is offline. Further, he suggests that replacement parts should be on hand for equipment deemed critical to operations. If anything goes wrong, he said, you don’t waste weeks while losing millions per day ordering a new machine.
Another user agreed with this approach. He said a recent unexpected failure would have taken eight weeks to fix. With huge revenue losses at stake, the company had managed to borrow a replacement from a nearby company. Company policy is now to have replacement gear on site for high-revenue items.
Robert Perez, Staff Reliability Engineer for Enterprise Products Inc., presented on the subject of engineering ethics. He outlined case studies showing examples of ethical and unethical handlings due to failures and other problems.
“Following the letter of the law only gives you so much protection,” said Perez. “Even following a standard may not be enough as the standard takes years to develop and may already be out of date.”