Machinery Maintenance Modeling & Diagnostic Advancements

Today, maintenance professionals are looking for better ways to evaluate machine conditions. One critical task for plant management is making the decision of whether to stop a machine for unplanned maintenance or run until the next planned plant outage. These evaluations and decisions are often based on expert experience alone, and therefore subjective.

Software can support and assist end users in making critical decisions. One example is a new software platform used in high-pressure compression plants. It is based on the concept of quantitative evaluation of machine performance.

Using the software, it has been possible to assess the centrifugal compressors’ health conditions through quantitative comparison of field operative parameters to conditions predicted from a machine-computer model.

Image 1. A typical process screen (Image and graphic courtesy of IPC)

This additional information is not intended to replace a human expert’s conclusions, but rather to support it. When applied along with mechanical best practices, vibrational machinery analysis, accurate mechanical inspections and high-quality repair processes and procedures, the experts are armed with data that helps them with decision-making.

This approach, aimed to exceed the actual best practices, supports plant management in maintenance decision-making steps and maximizes plant return on investment (ROI).

Method Description

Quantitative assessment of compressor performance and health status requires a comparison of measured performance to design, adjusted to actual operative conditions. Starting from OEM design or test performance maps, the software predicts expected compressor performance in the actual inlet conditions (actual suction temperature and pressure and gas mix composition).

The software calculates, at the actual running speed, the following expected curves, valid for actual inlet conditions:

• discharge pressure vs. suction flow
• discharge temperature vs. suction flow
• compressor efficiency vs. suction flow
• polytropic head vs. suction flow

This allows the calculation of actual flow, expected pressure, temperature, head and predicted efficiency, and then compares that information to the measured numbers. The comparison of these values allows the detection of deviations. The process provides quantitative information about the compressor health status and possible operational problems.

Table 1. Comparing the pressures and temperatures from PT/TT transducers to the value predicted by the software

The same comparison has been executed for head and polytropic efficiency. In Table 1, it can be noted that the maximum error calculated is less than 1 percent.

Measured errors have been considered tolerable for the purpose of the evaluation of compressor field performance. The analysis developed using the software allowed an evaluation of the efficiency deviation (difference between the actual compressor
efficiency and the expected efficiency in the actual operative conditions). Time trends of calculations provided a useful analytical basis for compressor maintenance decisions.

The method has been used to predict compressor performances and support planning machinery maintenance.

Conclusion

Numerical prediction of compressor performance allows a quantitative evaluation of compressor health status.

The applied methods also allowed:

• Prediction of the performance of a centrifugal compressor under varying thermodynamic conditions of the inlet gas. The prediction of compressor performance is accurate even at high pressures, where the ideal-gas-theory and simplified-formulas approach commonly used introduces considerable errors.
• Analysis of compressor performance during operation and comparison to expectations provided by the OEM
• Advanced protection from surge
• Indications of the health of the compressor (diagnostics) based on the capability to analyze the performance and efficiency of the machine in a simple and immediate way
• Reinforcement of decision making and maintenance planning along with specific recommendations for optimization of plant productivity