3 minute read

Pre-empting pump failures

Next Article
Brexit: what next?

Brexit: what next?

Aaron White, Product Manager (Technology) at Packaged Pumps Systems (PPS), explains how the company has developed a predictive maintenance solution that enables it to save its customers both time and money.

Traditionally pump failures have occurred with little or no warning and in some instances, can go unnoticed for some time. Once identified, the pump owner would call their service organisation who would despatch an engineer to locate and hopefully rectify the problem. This process often requires new parts to be ordered and so can take more than one visit to get a pump system up and running again.

Advertisement

More recently, predictive maintenance technologies have been brought to market to help determine the condition of in-service equipment and to estimate the best time to perform maintenance to help reduce the instances of unexpected system downtimes.

These predictive maintenance technologies also provide an opportunity for service organisations to work more closely with their customers. If the end-user is willing to share data with service providers, maintenance procedures can be made more efficient and streamlined.

“We believe that predictive maintenance offers an exciting opportunity for service providers like ourselves,” said Aaron White, Product Manager (Technology) at Packaged Pumps Systems (PPS). “Before embarking on any predictive maintenance project, however, we first needed to understand how industry could apply the technology. We needed to answer questions about how the data would be transmitted from the pump to us? What would need to be measured, and how would large amounts of data be processed.

The company set out to answer these questions. “From a physical ‘pump autopsy’ of 50+ failed pumps collected from various sites, we found 70% of failures could have been detected remotely and before they failed,” continued Aaron. The most significant cause of failure was found to be the deterioration of isolation between the windings and earth, which resulted in shorted and burnt windings. It is possible for this fault to be pre-detected by measuring the insulation resistance between the protective earth and the motor windings.

McKinsey & Company published a study showing that the use of predictive maintenance will have a 50% reduction in downtime due to equipment failures.

PROVIDING A SERVICE

Having done its research, PPS set about finding a solution which would allow it to collect and transmit data from customers’ pumps to provide a predictive maintenance service. “We established an Intelligent Service ecosystem (ISe), which consists of four main pillars,” explained Aaron. These are:

1. Internet of Things (IoT) compatible devices with automated insulation resistance testing every two days or on request.

2. Remote monitoring software that employs advanced algorithms to feed the data through a traffic light type threshold and trend system.

3. An Operations Centre for 24/7/365 monitoring that tracks both short- and long-term trends, while analysing events and actioning alerts. 4. A service engineer base to allow for onsite action.

The impact of this predictive maintenance service has been evaluated by analysing data collected over ten months. The number of breakdowns per 1,000 addresses on un-monitored sites equated to 211 per 1,000. With remote monitoring in place, this fell to 143 per 1,000. These figures demonstrate a decrease in breakdowns of 47% when using predictive maintenance.

“From data collected over 12 months, and a monthon-month comparison of over 13,000 data samples we now have knowledge about pump health deterioration that hasn’t been available before,” continued Aaron. “The average drop in insulation resistance is 4 MΩ per month. Armed with this big data statistic, we can start to extrapolate accurate pump life expectancy.

“McKinsey & Company recently published a study showing that the use of predictive maintenance will have a 50% reduction in downtime due to equipment failures and this figure is in line with our own findings.”

In conclusion, Aaron said: “Based on two years’ experience with remote monitoring and predictive maintenance, we firmly believe that this is the way forward. The cost of sensors and processing/storing data is reducing, the algorithms we use are improving every day, our operational knowledge is increasing, and market acceptance is growing.”

This article is from: