Understanding pump curves – watching Bad Actors

A Wider Monitoring Philosophy for the Oil, Gas and Process Sectors

By Ron Astall, United Pumps Australia

Long-time loyal followers of Pump Industry are likely acquainted with Ron Astall’s well-received series of articles on comprehending pump curves. In this edition, Ron returns to introduce us to the idea of Bad Actors and what to do if you have a Bad Actor.

With this article I am digressing somewhat from my previous focus on curves, their operating regimes, selection practices, suction conditions and system interactions. However, an important common thread remains – the effects of operating conditions on pump reliability.

Oil and Gas pumping applications are critical for safety and reliability; because of this, equipment in these services routinely has sophisticated condition monitoring regimes, with associated alarm and shutdown settings.

This is a discussion on how the way we operate, repair and maintain our pumps can impact reliability. From this perspective, we will consider the concept of ‘Lifestyle Monitoring’ as an adjunct to conventional machine condition monitoring.

My focus will be on centrifugal pumps in the chemical and petrochemical industries, but I believe the concepts can be applied to other applications. Indeed, my first example is from irrigation.

First, a short story

Back in the mists of time, I was sent to a customer’s site just outside Sydney to troubleshoot a 175kW single stage split case pump in irrigation service. This pump fed recycled water through a large sprinkler system, where it simultaneously watered and fertilised some lush pastures.

The problem was that the pump’s shaft would break every few months. The failure fracture surface was consistent with bending fatigue and pressure and power measurements showed that the pump was running reasonably close to Best Efficiency Point (BEP). Under these conditions, high radial shaft loads would not be an issue and there was nothing immediately obvious to explain these failures.

Having taken our measurements, the operator explained that since there was no check valve, the discharge valve had to be closed before switching the pump off. The discharge valve was a hundred metres from the pump house. Once the valve was closed, the operator went back to the pump house and switched off the pump. He explained that he did the reverse during start-up; starting the pump then walking over to the valve and opening it.

Anecdotally, we learned that a cigarette or two was also involved in many of these journeys. This pump was operating at zero flow for up to fifteen minutes per shift, several times per day, while the operator moved back and forth.

The solution was to install a remote on/off switch right next to the valve – no more hiking expeditions and no more pump problems!

What is a Bad Actor?

No, not Russell Crowe singing in Les Misérables – a Bad Actor is a repeat offender, an unreliable machine with poor Mean Time Between Failure (MTBF) numbers.

A Bad Actor will be well known to the maintenance team and will almost certainly be a money pit. How often do we come across a story where the bearings are failing every three months – the bearings are replaced only to fail again, get replaced, fail again, etc? Or a particular machine has been reliable for years, but now fails every six months?

It is not difficult to identify a Bad Actor – sadly, often, a proper root cause analysis does not happen.

Does the Process team blame the Reliability team for the ongoing problems, while at the same time the Reliability team are blaming the Process people for trashing their machines?

While busy apportioning the blame, no one addresses the root causes. From our example above, the operator was not aware of the dangers of running a pump dead headed, nor did the system designers foresee the problems. It is tempting and easy to just blame the machine. Some Bad Actors may indeed be a substandard design, but more likely the poor old pump is being blamed for being misapplied, poorly operated, or poorly maintained.

When a Bad Actor is identified, I recommend a basic Pump Audit, such as that in Figure 1.

Whilst most pump failures are seal failures or bearing failures, as a basic starting point, it is always good to document the current actual operating conditions: is this the type of pump you would select now for this service?

What’s in a Pump Audit?

A Pump Audit (Fig. 2) may highlight root causes that might be otherwise missed.

For the purposes of this discussion, let us assume that root causes have been established and addressed. Let’s park this for now and examine how we can optimise ongoing machine health.

In this article, I really want to focus on mitigation principles and the opportunities for a wider, preventative approach to machine health monitoring.

Health monitoring – what are the inputs/what are the results?

For the rest of this discussion, I am going to focus on how machine health monitoring can maximise machine reliability.

Most monitoring is logically Condition Monitoring with associated set points for alarms and when things get worse, trips or shutdowns. The parameters monitored can include bearing temperature, vibration, seal leakage, flow, differential pressure and power consumption. This monitoring will always be of critical importance for safety and plant integrity.

I am not suggesting that conventional Condition Monitoring is not important. However, Condition Monitoring normally only provides a warning when there is already a problem, or a problem is already developing.

How do we lavish love, TLC and attention on this baby? Can we identify and monitor any issues that may cause problems and thus avoid them developing. Let’s draw some parallels with our personal health.

Personal health monitoring – what are the inputs/what are the results?

Our inputs are our genetics and our lifestyle; the outputs are our general health indicators such as blood pressure, cholesterol, blood sugar, liver function, lung function, triglycerides, etc. These are analogous to machine condition monitoring.

In recent years, nearly everyone I know took up wearing some kind of health monitoring, heart rate, step counter, blood O₂, GPS, sleep tracker, cardio burn zone, kilojoule, floors-climbed activity gizmo. Many also tell the time, make phone calls, surf the ‘net, arrange Taylor Swift tickets and perhaps make you a cappuccino. These aspects are what I call the Lifestyle inputs to our general health.

We humans have become increasingly enthusiastic about monitoring health inputs. If we measure it, we will control it. This applies not only to our physical wellbeing, but also our behaviour.

An excellent example is the experience of one of our gas producers at their remote desert site. Vehicle accidents and associated personal injury rates were terrible – the desert environment inspired many drivers to think they were competing in the Dakar Rally.

What did they do? They made it compulsory for all vehicles to have real-time vehicle and driver behaviour monitoring. The incident rate immediately plummeted: problem solved and lives saved.

It is this monitoring of Lifestyle inputs that I am suggesting can foster a preventative approach to Machine Health Monitoring. In many instances, much of the relevant data is already being recorded. The next step is to use and interpret this information as a reliability data input.

Yes, the process is sacrosanct. Ultimately, production is the goal. But in the same way drivers can drive more sensibly, reduce fuel consumption and still arrive at their destination on time; with more operator awareness of how we are treating our machines, there must be scope to meet production targets without compromising machine health.

As a parallel to evaluating our personal health inputs, I have created an “Inputs” and “Results” chart for machine health.

Machine health monitoring – the inputs and the results

For a given pumping installation, not all the items in Figure 3 will be relevant. If the process is always constant, we may not bother to monitor process related aspects. Similarly, we may not need to monitor operating data for seals or auxiliary systems. It may merely be sufficient to confirm maintenance practices.

Of course, we still need our traditional Condition Monitoring safeguards. The aim is to acknowledge that how we operate and maintain our machines will impact their reliability and to consider what we might monitor to improve the Lifestyle of our pumps.

By Lifestyle Monitoring, I do not mean that our machines ought to take a yearly Palm Cove vacation, sit in a deck chair, drink pina coladas and watch the younger units frolic on the sand. Intriguing imagery though! By Lifestyle Monitoring I mean being aware of, and tracking the operating conditions and scenarios that may needlessly compromise reliability.

Case study – mechanical seals

To illustrate the concept of Lifestyle monitoring I have outlined an approach for developing a program to improve mechanical seal reliability.

The starting point is to identify why seals fail (see Figure 4). Then we can develop our Lifestyle Monitoring program (see Figure 5).

To keep this discussion manageable, let’s drill down on just a few of the above and consider parameters we might monitor. The pump or seal manufacturer’s assistance may be needed for some of these aspects:

Temperature Monitoring: ensuring the process liquid, coolant and barrier fluid temperatures stay within guidelines will assist in maintaining a stable fluid film at the seal faces
Pressure Monitoring: as with temperature, we need to ensure sufficient pressures to prevent flashing and dry running at the seal faces, maintain barrier fluid properties and capture over-pressurisation incidents that will overload the seal faces. Ideally, pressures in the seal chamber itself could be measured. If Seal Chamber pressures cannot be easily measured, the pump or seal manufacturer ought to be able to advise the predicted pressures and in some instances how these might be controlled or varied. Where there are fancy secondary barrier fluid systems, these can
usually be instrumented as required to monitor pressures and temperatures
Coolant Flow: if there is a cooling water harness it is good to know that it is consistently flowing through the heat exchanger. If coolant flow fails, seal failure will follow Pump Flow rate (operating point): a pump running significantly away from its BEP may vibrate, surge and suffer shaft deflection or breakage – none of which will be helpful for the seals. A pump running at very low or zero flow will generate damaging high temperatures. See Figure 6 below for more detail on this aspect

Remember our problem pump at the beginning of this discussion? The issue was operation at zero flow.

Figure 6 below illustrates how the Pump Operating Point will affect the pump, the loads on the pump shaft and the likely subsequent vibration and shuttling that will upset the mechanical seal.

OK, so we’ve collected all this data. What will I do with it? Why not make it a reliable KPI? Figure 6 shows the concept for the pump operating point.

Guidelines can be established for time out of range and monthly figures published. If things are moving in the wrong direction, it will be noticed and hopefully questions will be asked.

Quite a few years ago, BP’s Dr Richard Brodzinski established a similar parameter in its control system to record pump operating flow versus design (rated) flow. They found that less than ten per cent of their pumps were operating within ten per cent of their design flow.

What did they do? They tightened their procedures around establishing pump duty points when purchasing new equipment. No arbitrary process margins, thank-you!

From my perspective in the industry, most recently-selected pumps now seem to be operated reasonably close to the specified design conditions.

Summary – you have a Bad Actor?

More often than not, the Bad Actor is bad because it is not the right pump for the job or it is being mistreated. The starting point is to do a pump audit and establish the root causes, address the problems and fix them. I did not say it would be easy.

To maintain ongoing reliability, I have proposed an additional monitoring focus on the pump’s Lifestyle. By Lifestyle Monitoring I mean being aware of and tracking operating
conditions and scenarios that may needlessly compromise reliability. Then use and interpret this information as a reliability data input.

By monitoring and recording this Lifestyle data we will have an opportunity to head off problems before they occur.

Albert Einstein once said; “Life is like a piano, what you get out of it depends on how you play it.”