Benchmarking compressors and turbines for production performance

In other cases, while the compressor trains achieved acceptable levels of reliability and availability, they failed to meet their design expectations. For example, in two completely separate installations, the compressor trains could not operate beyond 90% of rated speed due to uncontrollable vibrations and would trip if the speed exceeded 90%. These high vibrations were not connected to balancing or alignment issues, which were eliminated. In such cases the production levels were constrained and the operators had to accept a shortfall in production and revenue due to restricted compressor speed.

Underlying low train reliability and availability and poor performance are certain key machinery and compressor “design” parameters that affect the train’s performance as well as its reliability and availability. Some of these parameters are easy to measure, identify, and rectify; while other parameters are difficult to determine and require careful measurement, detailed calculation, and analysis.

Overall train reliability and availability benchmarking.

Benchmarking these key parameters to match best industry practices will help to maximize the asset’s performance and reliability, thus improving revenue and profitability. Although benchmarking will make the process easier to maintain and simpler to diagnose, it is not a “foolproof” system and still requires accurate measurement, correct analysis and interpretation of results by experienced engineers.

The compressor train in this context includes the gas compressor, the gas turbine and the associated process system, i.e. coolers, scrubbers and recycle and surge control systems. The train also includes the compressor and turbine ancillaries such as lube and seal systems; the sealing systems could be oil based systems or dry gas seal systems with their auxiliaries.

Developing benchmarking standards

The overall availability is a function of lost time for unscheduled shut downs, as well as scheduled shut downs for maintenance and inspection. While the scheduled maintenance period is decided by the plant owner, and is therefore less disruptive, machine reliability -- or the unscheduled shut down due to equipment failure -- is a major cause of poor plant availability.

The benchmarking standards are divided into reliability-related issues and performance-related factors.

The factors that impact on the reliability of a compressor train can be divided among the following:

• Design factors
• Quality of maintenance, inspection, and technical support
• Operation of compressor and gas turbines relative to their normal design conditions.

It is possible for a well designed compressor train, with proper maintenance and technical support, to reach high levels of reliability and availability when operated within its proper design envelope. Availability from 95% to 97% has been reported by many successful operating companies. However availability is not a fixed parameter but varies with time; it depends upon the interaction of equipment design and operational parameters.

Operating parameters in particular change with time and require careful attention. For example, changing reservoir conditions would impact on process conditions which may cause the compressor to operate outside its normal design envelope; it may surge or alternatively require too much power which the engine may not be able to deliver thus failing on overload, potentially resulting in production loss or surging of the compressor. If the benchmarking is carried out continuously the changes mentioned above are spotted in time and equipment failures avoided.

The following key parameters collectively identify the health and performance of the compressor train.

1. Radial and axial vibrations levels

2. Compressor operating speed

3. The position of the operating point on the compressor performance map

4. Real compressor performance (head and efficiency)

5. Compressor performance degradation with respect to the original design

6. The driver power (EGT) and power margin

7. Recycle (or anti surge) valve position

8. Process stability

9. Seal system performance.

It is possible to develop simplified graphs to demonstrate the above parameters which may be used collectively to monitor and benchmark the overall train performance. For example, the axial and radial vibrations are combined for a high speed compressor showing the desired operating conditions and the relative benchmarking standards.

Platform X performance & benchmarking standards.

The radial vibrations indicate the mechanical health of a rotor and bearing system. Low vibration levels (10-20 microns) indicate a good degree of balance and integrity of rotating elements and good alignment. A damaged rotor or fouled rotor would invariably result in a high degree of unbalance and suffer from high vibrations as a result. For high speed compressors, if the radial vibrations exceed 30 microns this is indicative of a “rough running” and should be investigated.

Similarly, axial displacement of +/- 200 microns would be acceptable but if they exceed +/- 250 microns they would signify axial misalignment of the rotor or thrust bearing failure. Thus by carefully monitoring these relatively simple measurements, the operators are able to keep close watch on the mechanical integrity of the machines.

Compressor aerodynamic performance

Compressor aerodynamic performance requires accurate measurement of a number of key parameters and the use of sophisticated computer programs using well proven thermodynamic equations of state. The use of equations of state requires accurate information about gas compositions. Use of incorrect gas molecular weight in the equations of state would result in incorrect volumetric flow, polytropic head and polytropic efficiency. This may appear as a performance degradation and would lead to incorrect diagnostics.

By monitoring the compressor performance on a regular basis the actual performance (head and efficiency versus flow) can be determined and plotted on the performance maps. The position of the operating point with respect to design point is crucial. If the operating point lies close to the original design the machine would perform as designed. However, if the operating point drifts away and moves close to surge or stonewall region this would signify that the machine is operating too far from the original design intent. Continuous operation close to surge or in stonewall would lead to operational issues; such as compressor surging or lack of sufficient engine power.

Gas turbine performance

The gas turbine delivers its maximum power at 100% engine speed; the engine power is lower at reduced speed. Gas turbine performance and delivered power is sensitive to a number of parameters including:

• Ambient temperature variations
• Fouling of air filters
• Fouling of air compressors
• Damage combustion chambers
• Erosion or damage to hp turbine blades
• Power turbine speed.

Gas turbine performance is a complex subject and whilst manufacturers provide performance data of their machinery at the design conditions, interpretation of conditions away from design is not straightforward. Many additional measurements and sophisticated computer programs are required to determine true gas turbine performance. A more practical alternative is to install a torque meter coupling between the gas turbine and the compressor which can provide a direct measurement of torque transmitted and engine power from the operating speed.

Operation of the compressor below 100% speed results in a reduction in net engine power. Therefore, due account must be taken of compressor operating speed when determining engine available power. Whenever the required compressor power exceeds the maximum power the engine can deliver, the machine will trip or operate in overload. Normally, the engine control system would limit the firing temperature by reducing fuel flow once the turbine exhaust temperatures or “EGTs” or “ECTs” as they are sometimes called, reach the maximum permissible limit.

Refurbishment or redesign

When the operating conditions change significantly from the original design, the compressor operates inefficiently and shows poor efficiencies requiring additional power.

An RB 211 24 B driven compressor train installed offshore after a redesign.

This is characterized by recurring problems causing low reliability or availability. In such instances, it is not enough to simply improve maintenance or opt for like-for-like refurbishment because the cause of the problem is either operational deviation from the initial design intent or a poor initial design.

Such issues can only be resolved by a redesign of the compression train. This task can be implemented readily in most cases with appropriate care and due diligence.

Modern gas compressor projects are now well equipped with advanced DCS systems backed by large historians capable of storing vast amount of data. Such systems provide the opportunity to carry out performance analysis of operating data stored in the historians. This action would capture the current performance of compressor and gas turbines and record the operating hours as well as reliability including failure frequency and time taken to restart the system. Reports generated will not only provide current performance but also trend availability and performance changes over a period of time.

Conclusion

A good and well maintained plant should aim to achieve at least 95% overall availability. Lower availability results from poor reliability of a compression train which is affected by the following:

• Design factors
• Maintenance, inspection and technical support
• Operation of compressors (and GT) relative to their design conditions.

By monitoring certain key parameters such as vibrations, compressor performance and its operating points, engine power against its base line performance, the health of the compression trains can be maintained at a high level thus increasing plant availability and reliability.

When reliability and availability problems persist despite good maintenance practices, the cause is likely to be design issues. This may be in the form of operations away from original design or a poor initial design. Such problems can only be resolved by a careful redesign which is readily implementable.

References:

(1) API specifications (617, 618, 614).

(2) Determining the Real Performance of Centrifugal compressors operating in the field. GPA Europe May 2006 Conference. By Dr. M Sib Akhtar.

(3) Benchmarking Performance & Reliability of Compressors and Gas Turbines for Oil and Gas Production Facilities. Offshore Middle East Conference 2010. By Dr M Sib Akhtar. A full version of this paper is available via email to [email protected] or by telephoning 01372 700760.