Pumps & Valves - Process Machinery - a troubleshooting guide01 October 2004

Every piece of process machinery is purchased to be installed, operated, and play its part in making a profit for the plant operator. When a problem arises with a piece of equipment that causes concern to the operator and starts to affect safety or profitability, the techniques of effective troubleshooting have to be employed. The results of effective troubleshooting are to understand the problem that has caused concern, make any corrective modifications, and return the equipment to a safe and effective profitable condition.

Troubleshooting is the logical analysis of history. Problems can be resolved by examining the history of the equipment and its application. As troubleshooting is linked to history, which is a function of time, the first step is to determine exactly when the problem first manifested itself, and who recorded the event. From this point backwards, events need to be analysed to determine if they could singularly, or in a compound fashion, have initiated the problem.

With newly installed equipment this historical analysis may go back as far as original installation, prompting the question 'has it ever operated correctly?' On established pieces of equipment, the analysis of historical data may only have to go back to the last major overhaul when the condition of the equipment was known and hopefully recorded. If the last event was a major overhaul then what took place will need to be analysed to ensure that an operation (or the omission of an operation) during the overhaul was not the instigator of the problem.

History at a price

The biggest problem with collating the historical record of a piece of equipment is getting information that is reliable, factual and historically correct. Many troubleshooting scenarios have failed to find the root cause of the problem because of faulty information. Hearsay, 'he told me,', 'I thought I saw,', 'we always,', and past practice, are not normally associated with reliable, factual pieces of data related to the problem that is to be resolved.

Because of the problems in getting good reliable data, an initial step is to clearly define what the problem is that has to be resolved, then to define what resource can be allocated to its resolution.

Has the piece of equipment stopped unexpectedly? Become very noisy? Failed to perform to specification or expectations? Failed in a disastrous manner that has caused injury? Is the problem of such a magnitude that an army of people can be employed to resolve the problem? Or is it simply a matter of getting someone who knows, opening a valve, and the problem going away?

With resource comes the cost: no troubleshooting effort is free. The fact that there is a problem to be resolved will have an effect on the profitability of the equipment and the associated application.

Experience

In parallel with looking at the operating history of the piece of equipment, and its associated elements in the process where it is employed, one should also look for other applications of the equipment in similar circumstances. Ideally these pieces of equipment should be as close to identical as possible in design and application. If the other pieces of equipment are successful then the task is to understand what the differences are between the successful pieces and the failed one under scrutiny. This analysis may reveal that there are no other applications of the equipment that work, and that the problem is fundamental to the design of the equipment. This is particularly true of pumps and pumping applications.

A pump is designed and manufactured to a specification. It is then operated in the vendor's test facility and invariably performs faultlessly. The pump is then transported to site and installed in its application. It may be after a short time or after many years of operation that the pump fails.

Troubleshooting the problem in a logical manner will determine if:

- The pump has a basic design problem and would have failed in any application,
- The pump in the specific application is wrongly applied, or
- The way the pump was being operated in the application was at fault.

When analysing the information about an application the question is 'has the equipment ever worked according to the customer's requirements?' If the answer is no, the next question is 'are the operating requirements the same as those for which the equipment was specified and designed for?'

Specification

The most important thing to establish very early on in troubleshooting a machinery problem is whether it has ever run correctly and whether it is running within the design parameters defined by the vendor and used by the equipment designers. If it is, and the specification has not changed, then the problem is not one of basic design.

There are many instances of pump failure where it is subsequently found that the pump was handling a process fluid that it had not been specified for. This is particularly true where the operating temperature and pressure at the pump suction force the liquid to cross its vapour pressure curve and flash into vapour. Pumps are normally designed for a fully filled liquid stream at inlet. If this is no longer the case and the inlet flow to the pump has become a mixed liquid and vapour stream, then there will be hydraulic disturbance at the pump inlet.

This then induces variable axial and radial hydraulic loads on the pump shaft, which distress the bearings and disturb the mechanical seal. The resulting seal failure is often the first sign of distress and the main reason for initiating a seal failure troubleshooting exercise.

An in-depth study of a mechanical seal failure will rarely expose pump cavitation as the initiator, let alone the reason for the changed inlet fluid conditions.

Maintenance

Poor maintenance is often blamed as the root cause of problems associated with machinery. Rarely has this been the case in my experience. Over-maintenance has more often been a contributor by damaging spigot registers and overzealous cleaning of flat faces so that they are no longer flat. This over-maintenance then induces internal misalignment within a machine. The lack of correct fitting and internal alignment then induces stresses and strains on the machine components that then cause faster deterioration, requiring further maintenance. This is particularly true where rolling element bearings are concerned, as they operate with very fine internal clearances. Over-maintenance develops into a spiral of deterioration with shorter and shorter run lengths between machine outages leading to eventual failure. Logical analysis of the maintenance history, from new, will show this trend of reducing operating periods. The solution is to return the machine to an as-new condition, or to replace the machine or component with a new one.

A word of caution, when replacing components not supplied by the manufacturer: two identical-looking parts may not be identical if not manufactured to the original specification drawings. Manufacturing machining dimension tolerances, material specifications and heat treatments are often not fully complied with on pirate parts.

If troubleshooting uncovers a cyclic failure mode at regular intervals, then returning the equipment to the condition it was in after the last problem outage is not a solution: the failure will probably occur again at the same time interval for the same root cause. Once a problem has been analysed and the root cause determined then a change must be made if the interval to the next outage is to be extended.

The one maintenance-related element that has a profound effect on equipment failure is dirt. No piece of machinery is ever designed to have dirt in it. It may be designed to handle a dirty fluid or a dirty gas stream, but not to have dirt in places such as lubrication systems, rolling element bearings and control systems.

Operation

In any application one must never lose sight of the fact that most equipment is operated by humans. A simple maloperation can often cause the bestdesigned piece of equipment, correctly applied, to fail. Getting to the truth can be a monumental task if known maloperation is the root cause. Walking through the operating steps prior to failure with the operators, explaining why each step is taken and what effect it can have on the equipment, often helps to educate and extract any wrongly held concepts on how the equipment was thought to operate.

An example of this was a centrifugal pump that failed repeatedly on start-up. The operators did not appreciate that starting a pump with an empty discharge line caused the pump to run out on its head flow characteristic, resulting in a momentary very high flow through the pump and hence suction system, which in turn caused a low suction pressure and cavitation in the inlet of the pump. The cavitation caused the mechanical seal to fail. Starting the pump with a closed discharge was equally disastrous as it induced high radial loads on the shaft and again caused the seal to fail. Starting the pump with a partially open discharge solved the problem and educated the operators.

The process of troubleshooting is to repeatedly ask the question 'why?' If an operating valve position was changed, and the answer to 'why?' is not a justifiable reason, then the real reason is to be further investigated. The response to a rising oil temperature through an oil cooler is often to change the cooling water flow without asking why the temperature has changed in the first instance.

Records

As has been stated, troubleshooting is the logical analysis of history. Good plant records can help enormously with back tracking what has changed. The first record should be the design/application information. Any major overhauls, modifications or changes to specification should be recorded. The little black books that many maintenance technicians use are often an invaluable source of information when troubleshooting.

After a major troubleshooting exercise, a report should form part of the plant records. There is no value in repeating an investigation after the second failure.

Paul C Barnard CEng MIMechE spent over 33 years in the oil and chemical industry covering projects, troubleshooting, turnarounds and teaching throughout Europe and North America. Now a consultant, he also gives lectures on troubleshooting for AEA Technology.

SOE

This material is protected by MA Business copyright
See Terms and Conditions.
One-off usage is permitted but bulk copying is not.
For multiple copies contact the sales team.