Setting up a condition monitoring system

Modern condition monitoring methods are used to operate turbo-machineries reliably and efficiently. Condition monitoring involves the use of sensors to gather data and identify trends. Dedicated parameters are monitored, baseline (normal condition) is defined and the trend of data is captured to identify condition changes.

It is important to obtain baseline information as soon as possible after the start-up of turbo-machineries. Too often, baseline conditions have ignored and it badly has affected condition monitoring. Without a baseline, there is no reference for the comparison and interpretation of data.

Major Components of Turbomachinery

For an effective condition monitoring, all abnormal conditions should be identified compared to normal (or baseline) conditions. However, it is not possible for all and every component and parameters of a complex turbo-machinery. Therefore, optimum number of parameters and components should be selected for an effective condition monitoring approach. Major machinery components considered in commonly-used condition monitoring systems are:

1.        Bearings: Including radial and thrust bearings.

2.        Seal and packing.

3.        Rotors.

4.        Auxiliaries: such as lubrication oil system, seal system, cooling system, etc.

There may be other key items in each specific turbo-machineries such as those in gas turbines, steam turbines, and others which need dedicated condition monitoring. However, the abovementioned is a generalized list applicable to nearly all turbo-machineries.

Modern Integrated Condition Monitoring

A very important development in modern condition monitoring of turbo-machineries is the concept of integrated condition monitoring. This means combining and interpretation of all monitoring data available, for example, from online vibration monitoring, thermal imaging and so on, all together to see if they all pointed the same direction.

For most problems in turbo-machineries, a single monitoring method might not be effective for an early prediction of a problem. There have been many shortfalls and ineffectiveness when using just one (single) monitoring method.

However, there are great benefits from having different condition monitoring methods combined to detect any developing problem at an early stage. There is a greater degree of confidence when two (or more) condition monitoring technologies reveal the same problem.

There are some obstacles which should be properly overcome in order to have an integrated condition monitoring system. Unfortunately, in some plants, condition monitoring systems such as vibration analysis, oil analysis, and others are typically carried out by different groups within the organization. Another obstacle, some condition monitoring systems such as some oil analysis programs usually consist of submitting occasional samples to an outside laboratory and results look more like chemistry than machinery condition monitoring. Integrating all these condition monitoring methods/data have been challenging. However, such an integration is the key for success in modern condition monitoring.

The latest generation of vibration are very capable and advanced. However, despite such capabilities, not even the most sophisticated vibration monitoring system can successfully predict developing problems in all cases. For instance, if vibration measurements are high then it could indicate that machinery is not in good condition. On the other hand, there were cases that high vibration was just misleading and there was not a serious problem. There have been cases where the vibration has been increased, but the machinery was healthy and it continued to work for a relatively long time. Another issue is for some problems and malfunctions, it should actually be developed to a certain degree to increase the vibration above noticing points. In other words, the early waring could be missed using only vibration monitoring.

This can be generalized when using only single monitoring method. When only one method or one set of sensor(s) is utilized, the prediction is not usually reliable. The solution is to use different/independent techniques to achieve reliable and early warning signals. For example, when three different monitoring methods, such as online vibration monitoring, thermal imaging and on-site wear debris analysis are used simultaneously, the monitoring quality is significantly improved and an accurate prediction can be obtained. If all these methods indicate a developing malfunction/problem in a component (for instance, a bearing) such a prediction can be reliable.

Radial Bearings & Thrust (Axial) Bearings

Radial bearings of modern turbo-machineries are usually sized based on rotordynamics. They should not often be affected by minor machinery load changes or slightly modified operating conditions unless there is a link to lubrication oil conditions, machinery rotordynamics, etc. On the other hand, thrust bearings have usually been designed based on machinery loads (with respect to balance piston(s) and impeller arrangements) and they can be affected by machinery load changes and new operating conditions.

General Guidelines on Set Points

Alarm and trip set points are very important for efficient and reliable operation of turbo-machineries. These set-points should be properly selected to avoid unnecessary alarm or shutdown (trip) in transient but safe situations. On the other hand, malfunctions and problems should be identified in early stages to avoid catastrophic failures. Set-points should be selected case by case considering all involved factors such as process/operating conditions, machinery details, shop test results, performance test data, baseline, etc, but some rule of thumb and practical recommendations are noted as follows.

As a very rough indication, for rolling-element (anti-friction) bearings, housing vibration (peak) and bearing housing temperature limits are recommended around 8 mm/s and 85°C, respectively. For hydro-dynamic (or tilting-pad) bearings, housing vibration (peak to peak) and bearing housing temperature limits are recommended around 60 microns and 110°C, respectively. Most often limits are higher for hydro-dynamic bearings compared to rolling-element (anti-friction) ones. Recommendations regarding the maximum temperature at hydro-dynamic (or tilting-pad) bearings are varying. Even there have been recommended limited more than the above-mentioned. However, any temperature spikes and fluctuations (or vibration spikes and fluctuations) could be an indication of an abnormal behaviour which should be investigated.

As another set of rough indications, for rolling-element (anti-friction) and hydro-dynamic thrust (axial) bearings, axial velocity and displacement limits are around 1 mm/s and 0.5 mm, respectively. Lubrication oil supply and return temperatures are expected to be around 60°C and 90°C, respectively.

Lubrication Oil Analysis

A lubrication oil analysis is very effective tool to evaluate health of bearings. As a general rule, the reduction of lubrication oil viscosity to less than a 60% (compared to lubrication oil producer specification), particle size larger than 20 micron or water content above 150 ppm need careful attention. These may be considered as alarm limits.

Performance Analysis: An Effective Monitoring Tool

The analysis of performance and operation has always been an effective monitoring tool. As an indication, for turbo-machineries, if for a given flowrate and shaft speed, produced head falls below 90% or 85% of expected (certified and tested) value (based on certified and tested curve), the turbomachinery should be investigated, even if possible and appropriate, inspected at the first opportunity.

Significant changes in operating conditions can cause damages to turbo-machineries or their components. More complex the relationship between the turbo-machinery and overall process/operation of plant, more difficult to predict and monitor the situation. Turbo-compressors and gas turbines usually offer complex operational behaviors.