Myth: Compressor Transient Surge Analysis is Really Accurate

The transient behavior of most physical systems is often fundamentally different from their static behavior. This principle ties back to Newton’s second law, which incorporates acceleration as a second derivative term. For example, the movement of a race car accelerating through a curve cannot be analyzed by superimposing its stationary positions over time. This approach neglects the 3D inertial forces acting on the vehicle. However, for the sake of simplicity, quasi-static analysis is typically employed in engineering tools. Such an approach simplifies complex problems by treating transient phenomena as a series of steady-state conditions. This method is sufficient for many slow-moving physical processes, as transient effects can often be disregarded when inertial forces are minimal. Unfortunately, this simplification can lead to the misapplication of analysis methods when transient problems are analyzed using tools or software designed for steady-state scenarios.

For example, the characteristics of centrifugal or axial compressors are usually described with a head-flow map that assumes all operating points are stable and that thermal transfer is negligible; however, unless a compressor operates under a single condition for an extended period, allowing all heat-soaking effects to reach equilibrium, most real operating points are inherently transient. This is particularly true during startup and shutdown phases, especially in emergency-related fast shutdowns.

We previously discussed the phenomenon of surge in our Myth Busters column: Surge is a violent physical occurrence in centrifugal compressors that can potentially cause significant damage. It occurs when a compressor operates at the low-flow end of its range, where the impellers’ aerodynamic capability falls below the threshold needed to produce sufficient head to overcome the pressure differential between high discharge and low suction pressure. This imbalance reverses flow, temporarily rebalancing the pressures until the compressor generates forward flow and head.

Surge is a periodic and cyclic phenomenon that occurs at characteristic frequencies of the compressor and piping system, independent of the running speed. The full-flow reversals inherent to surge create significant forces that can severely damage various internal components. While the phenomenon has been thoroughly documented in existing literature and will not be explained here, it is essential to recognize that all compression systems are susceptible to surge. Correspondingly, surge control systems are designed to ensure compressors operate away from the surge line, avoiding the dynamically unstable region.

When assessing the operating stability of centrifugal compressors within complex plant piping systems—particularly their ability to avoid surge—transient surge analysis is often conducted. This analysis involves a detailed, one-dimensional evaluation of the compressor and its associated components, including valves, pipes, vessels, and the control system behavior. The goal is to evaluate the system’s capability to prevent surge under four transient scenarios: startup, process control, controlled shutdown, and emergency shutdown. Emergency shutdowns present the greatest challenge, as they lead to notable process change orders faster than any other situation. Failure of a compression train in a station—with multiple trains in operation—is another situation with rapid changes in operating conditions. Therefore, the surge control system must respond rapidly, opening the recycle valve to prevent the compressor from crossing the surge line as it decelerates quickly under a high-pressure differential.

Two key factors determine how close a system can operate to the surge line during an emergency shutdown:

• the system’s response, including surge control capabilities
• the physical impact of surge on the compressor

The latter is often overlooked. Many robust low-energy compressors could theoretically operate in surge indefinitely without sustaining damage, as long as the temperatures stay reasonable. Extending the normal operating range closer to the surge line is acceptable for such machines. Conversely, more sensitive or process-critical compressors require a safe surge margin to ensure, through a properly conducted dynamic analysis, that the system remains out of surge.

Unfortunately, many fixed and variable operating parameters and fluid properties are unknown before a machine begins operation. The analysis itself is complicated, requiring fluid dynamics modeling across the entire system under rapidly changing boundary conditions. Factors such as driver shutdown dynamics—involving electric motors, steam turbines, or gas turbines—are seldom modeled accurately. Even the compressor’s mechanical behavior and performance map pose challenges, as these models are based on steady-state assumptions. For example, standard compressor head-flow performance maps derive from measurements taken under steady conditions with the compressor in thermal equilibrium. During quickly changing process conditions, thermal equilibrium doesn’t exist.

The ability to prevent surge during an emergency shutdown is typically determined using simplified estimation tools. Two primary methods are employed: experience-based non-dimensional numbers—which consider piping volumes, inertia, and compressor power—and low-fidelity dynamic simulations. The latter combines lumped representations of the system’s physics with experimental data, and generally, the speed decay of the compressor must be estimated.

When the shutdown behavior of the driver, train inertia, system volumes, or valve-opening characteristics are known, these simplified tools can offer insights into dynamic system behavior and aid in conceptual decisions, such as determining the necessary number of valves early on in a project. However, non-dimensional number models lack sufficient accuracy for these analyses. Most compressor manufacturers mandate that surge control systems prevent short transient surges during rapid emergency shutdowns. This is prudent for the operator since repeat surges may degrade compressor performance. Increasing compression ratios and demanding operating conditions have made using hot-gas bypass valves and the surge control valve to mitigate compressor surges during emergency shutdowns more common.

Achieving satisfactory performance requires a carefully designed and optimized surge control system, which depends on detailed and accurate engineering analysis. Simplified rules-of-thumb, non-dimensional models, or basic lumped volume models cannot accomplish this. A well-implemented, full-time-domain numerical one-dimensional dynamic compressor analysis, calibrated with real test data, can achieve modeling uncertainties as low as 10 - 20%. This assumes that the driver behavior, heat transfer, dynamic compressor behavior, and control system behavior are correctly implemented. It also assumes that the instrumentation that feeds data into the control system is calibrated correctly and accurately.

The industry’s transient one-dimensional dynamic analysis tools have significantly improved over the past five to 10 years and have been repeatedly validated with test data. However, any engineering tools with uncertainty over 10% cannot be used to reduce design safety margins (such as head rise to surge or turndown) that are lower than that or, as an alternative, require increasingly expensive valves. Thus, even with these advancements, analysis results must be interpreted conservatively to facilitate sound surge control system design decisions. When analysis tools with significant inherent uncertainties are used, achieving a reasonable design is impossible when critical parameters demand accuracy within a few percent.

Clearly, the physics of compressor shutdowns is highly complex, leaving many real-world effects unaccounted for. The accuracy of any complex analysis is contingent on the quality of its assumptions and boundary conditions. In transient surge analysis, these assumptions and boundary conditions are often poorly understood, leading to inaccurate results. Most commercially available transient surge analysis software relies on significant assumptions about one-dimensional versus 3D effects and presumes steady-state compressor maps.

Additionally, these tools frequently fail to accurately model driver dynamics, transient pressure waves, and surge control system behavior, which can lead to substantial errors. In many cases, transient surge analysis results have such high uncertainties that overly conservative surge margins, restricting a compressor’s operating range, are required.

An engineering analysis is only as good as its assumptions. This is especially true for transient surge analysis where physical modeling is challenging, boundary conditions are often gross estimates, and required design margins for safe operation are very tight. Consequently, transient surge analysis outcomes are often so generic that they lack practical value or drive cost for the yard valves.

About the Authors

Klaus Brun is the Director of R&D at Ebara Elliott Energy. He is also the past Chair of the Board of Directors of the ASME International Gas Turbine Institute and the IGTI Oil & Gas Applications Committee.

Rainer Kurz is a recent retiree as Manager of Gas Compressor Engineering at Solar Turbines Inc. in San Diego, CA. He is an ASME Fellow and has published over 200 articles and papers in the turbomachinery field.