...many operating companies are mobilizing to increase their facility's charge rates. If you are one of these companies looking for improvements, correctly evaluating your flare system's maximum capacity could "make or break" your project's feasibility studies.
Your First Hurdle:
- 1. Determining the current flare system design basis.
- 2. Verifying if the loads and piping configurations reflect current operations.
- 3. Assessing which loads require changes with the planned modifications.
- 4. Reanalyzing the flare system with the new loads.
What are the Solutions?
Currently, API Standard (STD) 521 recommends three solutions for (as they refer to it) the "refinement of disposal system design load."
- 1. Dynamic System Load Modelling,
- 2. Load Reduction Credits, or
- 3. A combination of both
(However, even the API standard cautions the reader that all three of these methods require detailed and complex.)
So to help those operators find the right method for their situation, I sat down with John Burgess, P.E., one of Smith & Burgess' Principal Engineers, in order to see how he has guided our clients in the past.
John, I have recently noticed a surge of questions regarding §5.3.4 of the current API STD 521, 6th Edition. Let us start with a basic question, "What RAGAGEP does the user of these so called 'refinement methods of disposal system design load' need to be aware of?"
API STD 521, 6th Edition, §4.2.6 mentions:
|"Although favorable response of conventional instrumentation should not be assumed when sizing individual process equipment pressure relief, in the design of some components of a relieving system, such as the collection header, flare, and flare tip, favorable response of some instrument systems can be assumed."|
Additionally, §188.8.131.52 mentions:
|"Consideration may also be given to the capability for and response time available for operator intervention as a means of reducing system loads."|
Are there any cautions that the user should be aware of when taking credit for the favorable response of some "instrumentation systems" and "operator intervention?"
For "instrumentation systems," API STD 521, 6th edition, §4.2.6 states:
|"The decision to base the design of such systems on excluded or reduced specific loads due to the favorable response of instrument systems should consider the number and reliability of applicable instrument systems."|
|"When doing so, the user shall consider what other demands can be placed on the operator during the upset."|
Over the years, many methods have been developed to account for "instrument response in disposal system sizing and analysis." These techniques range from the extremely simplistic yet conservative; to the extremely complex and potentially less conservative, but much more realistic.Edgard Kurnia:
OK, so I am guessing that these credits for "instrumentation systems" and "operator intervention" can be taken during the "refinement of disposal system design load" (API STD 521 §5.3.4). This section mentions two main methods: i) Dynamic System Load Modelling and ii) Load Reduction Credits. Can you elaborate more on the differences of these two methods?
Let us start with the "refinement of disposal system design load." The simplest means of load definition is to identify all the applicable overpressure scenarios and simultaneously relieve them to the flare header. This will certainly provide a conservative estimate of the peak load to a disposal system. However, very few flare systems are designed (nor should be designed) for that large of a flare load. And due to the nature of the various systems, several factors make this peak load unrealistic:
- 1) Basic process controls will mitigate many overpressure scenarios
- 2) Operators will intervene by shutting down systems
- 3) The time from normal operating pressure to relief pressure will differ from system to system depending on:
- Heat input
- Pressure differential
- Other factors
- 4) Secondary trips and controls
- 5) High integrity shutdowns and controls
Dynamic System Load Modelling only addresses the 3rd factor above. Historically, this method can result in 30% - 40% reduction in flare load from a single source; however, it may be able to further reduce the load with a more refined and detailed analysis. The drawback is that this method is also computationally intensive (see API STD 521, 6th edition, section 4.3.3), which may result in high cost due to complexity and software cost.
For "Load Reduction Credits," one solution is called "Flare Quantitative Risk Analysis (QRA)." This method allows for the inclusion of all the factors mentioned above at the expense of greatly increasing the number of potential hydraulic combinations that may need to be evaluated. The onset of parallel computing capabilities and virtual environments has improved the associated time and expense. However, the costs to identify and characterize the independent safeguards and alternate load levels continues to remain high by today's standards.
The largest difference between these two techniques is associated ongoing lifecycle costs. Updating a dynamic simulation is as costly as the initial evaluation; while the QRA (presuming, there are minimal changes in the safeguard configuration) is much less complex to re-evaluate.
Seeing that QRA has a lower ongoing lifecycle cost, for which kind of flare system concerns would you recommend?
The QRA methodology can be used to address any concerns associated with the hydraulics of the flare header size and configuration – backpressure, separation, radiation, flame out dispersion, etc. This also ranks the likelihood of the calculated value, such as back pressure at the outlet of relief devices or estimated vessel accumulation pressure with probability. Then, there are established methods to correlate this accumulated pressure with consequence (such as API Recommended Practice 581 §7.2.5).
Finally, what kind of documentation should the user prepare when using the "refinement of disposal system design load" method to address any regulatory compliance audit?
The documentation required in either method is a detailed report identifying the assumptions, input data, and results of the engineering analysis.
All right everybody, that will be the end of this short conversation between I and John Burgess. I hope this will help you to make the decision that you have been yearning to and move forward with your project. For those that feel what we talk about is not in the level of details that you are expecting, have questions, or who want to share their related experience related to this matter, feel free to reach out to Edgard or John.