Pumps & Systems, September 2008
Why is "run to fail" a common operational strategy at water/wastewater organizations? Explore this phenomenon and other results from a maintenance methodology survey.
One of the most common questions from water and wastewater utilities when they are asked about their maintenance practices in lift stations is: "What do other organizations do?"
Global maintenance methodologies vary from one end of the spectrum to the other. Maintenance practices for utilities include run to fail, preventive maintenance on a schedule of time or time in service and predictive maintenance by monitoring leading indicators of pump and motor problems. Most people will be familiar with these terms, but it is worth a brief review.
Run to fail implies that the organization will expect inevitable failures and will have sufficient spare parts and staff on hand to keep the downtime below organizational requirements. The allowable downtime is usually calculated on a station-by-station basis.
Preventive maintenance and predictive maintenance were neatly contrasted in a recent article by Thompson and Granger:
"By applying advanced diagnostic technologies to the issue of equipment maintenance, the probability of any one piece of equipment continuing to operate effectively can be accurately determined. Repair crews can check on advance warnings, not the calendar, before going out to inspect, tear down (if necessary), and repair a piece of machinery. This is the essence of predictive maintenance, and it's far less expensive than the old fashioned preventive maintenance so many of us grew up believing in.
For years companies have been performing the preventive maintenance recommended by manufacturers without question. The reasoning being that manufacturers have done all the research needed to ensure their equipment will operate properly in any environment. Most companies also want to ensure that equipment warranties are maintained during the initial installation period. Once these maintenance actions are entered into the CMMS or work routine, no one challenges their validity or periodicity because ‘that is the way we have always done it.' This can be a very costly way of doing business."1
Run to Fail
Based on survey results, the most common methodology across the majority of utilities in the United States and Australia is run to fail. Manpower shortages and/or lack of a budget for better monitoring and control areoverwhelmingly the most common explanation for this choice.
One very large utility had a stated policy of run to fail, but this was a deliberate policy based on its own assessment of the costs and benefits of alternatives. This policy had the benefit of simplicity. It had to maintain enough store locations to keep an adequate amount of spare parts, and it had to have enough trucks and maintenance crews so it could replace any pump within a set number of hours.
This example was definitely an exception. Many smaller organizations were divided on their "de facto" policy of run to fail. Some felt that although there should be a better solution, the current scheme was working "because we don't get that many failures." Those organizations that believed their policy was not optimal were experiencing continual failures and breakdowns, with staff performing only reactive "fire fighting."
Preventive Maintenance
More common among medium-sized and larger organizations was a level of preventive maintenance, e.g., servicing the impellers and seals within a set timeframe or amount of run time (e.g. every six months, every 5,000 hours, etc).
While easy to fault-why not one time period than another?-organizations with preventive maintenance procedures believed that run to fail was flawed and there was a "duty of care" responsibility to keep the assets in a good condition. Most utilities with this methodology acknowledged that they did notknow whether they were doing too much maintenance. However, in most cases they were comfortable with their level of spending and were not experiencing a high level of failures. Moreover, their comfort with their level of spending was usually because they had become accustomed to it.
It was common to find this group of users beginning to use predictive maintenance to some level, particularly manually meggering motors.
Predictive Maintenance
What predictive maintenance is possible and practical on submersible centrifugal pumps, the pump choice for lift stations?
- Insulation Resistance Testing, also known as Meggering the motor, tests the integrity of the motor windings, which gradually break down. Between half to three-quarters of motor failures in submersible pumps are caused by the insulation breaking down. With this method, users must choose between a manual test on a periodic basis or an automatic test.
- Pump Flow Rates test the hydraulic performance of the pump. Since most wastewater lift stations do not have flowmeters, this is usually not monitored; instead, the proxy of hours run is commonly used. However, hours run can increase because of higher throughput or deterioration of pumps.
- Pump Efficiency measures volume pumped per energy unit (KWHr or KVAH). This is a more effective measurement as it also provides the financial impact of deteriorating pumps, letting the organization know the cost of the problems.
A few other methods are worth mentioning:
- Seal Faults are common, as most pump vendors provide them as standard or as an option. Since they give an early warning of water leakage, they provide predictive maintenance.
- Three-Phase Currents-Most U.S. lift stations do not have current meters, but they are standard in the UK and Australia. Typically, they have a line drawn on them for the expected current draw. As a local indication, it is a limited tool, but providing the three-phase current remotely provides another predictive maintenance indicator.
- Three-Phase Supply is rarely available remotely, but if it is, it will supply information for predictive maintenance.
Regional Perspectives on Predictive Maintenance
In the UK, due to the large size of each of the water/wastewater utilities, almost all have a plan to introduce predictive maintenance, or condition-based monitoring. However, due to the cost and complexity currently involved in implementation, the UK is still mostly a mix of run to fail and preventive maintenance.
The main perspective in the UK is that the organizations are "asset managers," and as asset managers they cannot afford to "fly blind" but require real data to know the status of all of their assets. There is also a major focus on energy costs and the associated "carbon footprint," with this factor becoming as important as pump and motor lifetime.
One large U.S. utility has a team of electricians who visit every site every three months to test the motors for insulation resistance-an expensive exercise. Across the United States, this practice is found in less than 10 percent of utilities.
Challenges
Run to Fail
The major problem with run to fail is that it is a choice to fly blind as to the state of the assets. While there are many reasons why pumps and motors can deteriorate faster than historical data shows, there is not space in this article to detail all of them. However, an excellent example is given in "Unbalanced Voltages and Electric Motors" (Pumps & Systems, June 2008) by Thomas H. Bishop.
Bishop's article outlines the effects of slight voltage imbalances on pump lifetime. There are similar effects on lifetime for small under voltages.
A practical example of this occurred at a large regional municipal utility that introduced a predictive maintenance product into the lift stations; at that time they discovered why so many pumps were failing before seven to eight years in one area instead of the expected 25 years. The three-phase supply was around 10 percent below nominal, and because the three-phase supply was not remotely monitored, no one was aware of it. The supply reduction was causing higher running currents, but not enough to trip any panel components or the pump thermistor. Still, the pumps were running too hot, causing a big reduction in lifetime.
Preventive Maintenance
The major problem with preventive maintenance is that it is expensive-and hard to justify:
"The challenge is to find the correct time interval: some machines will be dismantled unnecessarily, yet others will fail because they were not inspected often enough, others will fail after maintenance work because some human-induced error has been made."2
However, in the absence of a cost-effective predictive maintenance tool, many organizations are using preventive maintenance in an attempt to be better asset managers.
Predictive Maintenance
The major problem with Predictive Maintenance has been cost. Meggering motors was formerly done by sending electricians to each site, which is costly.
Measuring flow by installing flowmeters at each wastewater lift station is also expensive. A number of products and algorithms are available to do calculated flow via draw-down tests, but typically they only measure outflow at the station and do not break that down into individual pumps. Algorithms over SCADA face problems, including that U.S. communications protocols (Modbus is the most common) mostly do not give a date/time stamp, so there is some uncertainty as to when the pumps started and stopped.
Most products do not measure pump efficiency (gals or liters per kWHr).
Finally, using predictive maintenance (and preventive maintenance) still requires spare parts and response crews, although the investment should be significantly less than in the case of run to fail.
New Technology for Predictive Maintenance
In 2006, the first pump station manager was released, making cost-effective predictive maintenance viable.
A pump station manager combines a sophisticated pump controller, PLC, RTU and user interface into one unit which also measures three-phase supply, three-phase currents, power, energy, flow rates per pump, pump efficiency and insulation resistance testing (meggering). All of the data is available remotely, as the pump station manager can be connected to any open SCADA system.
Pump station managers have allowed predictive maintenance to emerge as a best practice for lift stations.
References
1. What Price Preventive Maintenance? T. Thompson and M. Granger, 2004
2. Predictive Maintenance of Pumps Using Condition Monitoring, Beebe, Raymond S., 2004