Wednesday, August 30, 2017

In a previous blog, Safe Operating Limits – Part 1, we provided a heuristic for defining the safe operating limits:

• Obtain the operating envelopes for all modes of operation
• Identify key limit values for the process parameter
• Add limit values commensurate with complexity and risk
• Drive the Safe Operating Limits as close to the operating envelope as reasonably practical
• Evaluate consequences of deviations

In Part 1, we used a simple example for upper and lower temperature limits, and in Part 2, we explored simple upper pressure limits.  In this continuation, we want to delve deeper into the identification of key limiting values based on the capabilities of the overpressure protection systems.

To start, we must select the process parameter(s) that require limitation in the first place.  Since the goal of setting limits is to maximize process safety, and ultimately the safety of personnel, we select parameters that we can reasonably expect to help prevent a loss of containment or mitigate its consequences.  Pressure and temperature are obvious such properties. In addition, OSHA’s PSM standard suggests flow rate and composition¹.  We have encountered other variables that are dependent on the process itself, such as liquid level, pH, and contaminant concentration.

One basis for the key limitations for pressure and temperature is the actual strength of the material used to construct the processing equipment and/or piping.  These are direct limitations that are provided in the code of construction for the equipment, and are commonly identified, such as the embrittlement transition temperature for a lower temperature limit, and the maximum allowable working pressure (MAWP) for an upper pressure limit of a pressure vessel built to ASME Section I or Section VIII.  Note in Part 2 of this series, we made a case for the maximum allowable operating pressure for a pressure relief device as another upper pressure limit.

The MAWP of a pressure vessel is determined at a ‘designated coincident temperature’², which we often find to be the maximum design temperature.  This maximum design temperature then becomes a key limitation for an upper temperature.  This is one of the most common examples of a dependency of one limit on another.

Other common dependencies are those related to the design of the overpressure protection system, particularly for the other process parameters of flow and composition.  Both the determination of the credibility of potential overpressure scenarios and the evaluation of the required relief rate for those scenarios deemed credible can be based on key limitations that should be reflected in the safe operating limits.

Overpressure scenarios.  There are several overpressure scenarios in which a process parameter is an input to the determination of the credibility of the scenario³.  In these cases, we are interested specifically in situations in which a scenario is deemed non-credible for the purposes of designing the relief system.  This may also take the form of acceptance of measures like administrative controls and mechanical locking elements in the elimination of a scenario.

Closed outlets.  Most closed outlets cases are determined based on an upstream limiting pressure, often mechanically limited.  This may be a relief device set pressure, which we already have as a limit.  In some cases, the upstream limiting pressure is based on the capabilities of a fluid driver (like a pump), making the suction pressure a potential limit for an upstream system.

Overfilling.  The maximum liquid level in the equipment is sometimes used in the determination of the overfilling scenario, particularly in cases where operator intervention or instrumented systems are relied upon for prevention.

Failure of automatic controls.  Most situations involving an opening of an inlet control valve or its bypass are determined based on a maximum operating pressure upstream, resulting in a direct correlation to the upper safe operating pressure in the upstream system.  In addition, situations involving the potential for gas breakthrough (for example, a liquid level control valve from a separator) are dependent on the liquid level in the upstream system.

Abnormal process heat or vapor input.  Considerations for cases involving inadvertent opening of valves or reverse flow are similar as for the failure of inlet control valves.

Chemical reaction.  The potential for many runaway reaction scenarios is often dependent on the presence of contaminants, the overall composition, and reaching an onset temperature⁴.

Hydraulic expansion.  In cases with process heat transfer, the approach temperature of the process determines the potential for hydraulic expansion.  There are also cases where the potential for vaporization of the trapped fluid is determined by the approach temperature.

Required relief rates.  Despite the numerous lists of potential causes of overpressure, the calculations of required relief rates for overpressure protection systems tend to be one of only a handful of techniques⁵.  Each technique has common process parameters as input.  In these cases, we are interested in situations in which a change in the process parameter may cause the existing overpressure protection to become inadequate.  Usually, one will use the safe operating limits as input to ensure adequate protection; however, here we are assuming these limits have not yet been set (or more commonly, were not considered in the relief system design basis in the first place).

High pressure fluid input via a flow limiting element.  In some cases, the maximum process flow is used as the basis, resulting in a direct correlation to the upper flow rate limit.  In other cases, the upstream pressure is used to establish the driving force to push fluid into the protected equipment.  The composition and temperature can also affect the flow rates and the capabilities of the overpressure protection system.

Heat transfer.  In cases involving a heat transfer fluid, the flow rate and temperature dictate the heat transferred to the protected equipment.  For systems relying on vapor-liquid equilibrium, the pressure has a direct effect.

Runaway reactions.  Reaction kinetics are often dependent on composition and temperature.

Impact on sizing.  In the design of relief systems, the calculated relieving requirements are then used to size relief devices that have sufficient capacity for relief.  Often, standard sizes are selected that provide more capacity than required.  Also, only one scenario dictates the size of the relief device, and so the relief device may have significantly more capacity than required for other scenarios.  It would be impractical to evaluate every overpressure scenario to determine rigorously the limits of a process parameter that would result in undersizing of a relief device.  As a result, it is common for non-reactive systems to evaluate only the scenario that dictates the sizing using a scaling factor derived from the calculation technique.  Flow rates are often directly proportional, while pressures are often a square root.

One question that arises commonly is the effect of composition and temperature on the design basis for sizing.  Cases where we look for significant impact include those where low molecular weight non-condensables (hydrogen, methane) might be involved.  Regardless, in our experience, the sizing basis for non-reactive systems is not very sensitive to these parameters, presumably because of offsetting effects and minor changes in key properties (for example, the isentropic expansion coefficient for vapor).  To illustrate this, we have taken two common source models (heat input from external fire and mass input from isothermal pipe flow) and evaluated relief sizing with different simple hydrocarbons, holding other variables constant.

Figure showing results of sensitivity calculations (temperature and molecular weight) for an external fire scenario

Figure showing results of sensitivity calculations (temperature and molecular weight) for a pipe flow calculation

n-Pentane has the average molecular weight of the components selected for these simple examples (propane, n-butane, n-hexane, and n-octane being the remainder), and is used as the baseline for comparison with the results for the other fluids.  Fluid properties are obtained from REFPROP⁶.  The results for n-octane are telling: for the external fire case, a 16% increase in required relief area despite an increase in molecular weight of 60% and absolute temperature of 27% (an actual temperature difference of almost 200°F, 446°F for octane vs 251°F for pentane).  For the pipe flow case, an 11% increase in required relief area for a similar change in molecular weight and absolute temperature.  The results for butane and hexane are more representative of what we would expect of the precision around molecular weight and temperatures: ±20% for molecular weight and ±10% for temperatures.  The required relief areas are ±6% for the external fire case and ±3% for the pipe flow case.

Management of Change.  An important benefit of evaluating these limits for use in defining safe operating limits is in the management of change process.  Having these key values enumerated facilitates the evaluation of the potential effects of a change that affects a safe operating limit⁷, which can be challenging when there is a dependency of the parameter on other aspects such as the design of the overpressure protection system.

---

[1] 29 CFR 1910.119, Process safety management of highly hazardous chemicals. §1910.119(d)(2)(i)(D).
[2] American Society of Mechanical Engineers. “2010 ASME Boiler & Pressure Vessel Code, 2011a Addenda, Section VIII, Division 1 – Rules for Construction of Pressure Vessels”. Jul 2011; New York, ASME: §UG-98.
[3] American Petroleum Institute. “API Standard 521: Pressure-relieving and Depressuring Systems”. 6th Edition, 2014 Jan.
[4] Fisher HG. “An overview of emergency relief system design practice”. Plant/Operations Progress 10(1); Jan 1991: 1-12.
[5] AIChE Center for Chemical Process Safety.  “CCPS Guidelines for Pressure Relief and Effluent Handling Systems”.  2nd Edition,
2017.
[6] Lemmon EW, Huber ML, McLinden MO. NIST Standard Reference Database 23: Reference Fluid Thermodynamic and Transport Properties-REFPROP, Version 9.1, National Institute of Standards and Technology, Standard Reference Data Program, Gaithersburg, 2013.
[7] Aziz HA, Shariff AM, Rusli R. “Interrelations between process safety management elements”. Process Safety Progress 36(1); March 2017: 74-80.