Wednesday, October 12, 2022

Review - CSB Publishes PES Refinery Investigation Results

Yesterday the Chemical Safety Board published their final report on the Philadelphia Energy Solutions (PES) accident that occurred in June 2019 and resulted in explosions, fires and a sizeable hydrofluoric acid release. The CSB reports that there were six relatively minor injuries on-site and no reported off-site effects. The report identifies the cause of the initial process leak and proposes recommended actions for the EPA, American Petroleum Institute, and ASTM International.

There are three additional final reports that the CSB plans to issue by the end-of-the year:

• Kuraray EVAL, Pasadena, TX,

• TPC Group, Port Neches, TX, and

• Husky Energy Refinery, Superior, WI


The CSB made the following recommendations in this Report (pgs 80-1):

• EPA - 2019-04-I-PA-R1 - Develop a program that prioritizes and emphasizes inspections of refinery HF alkylation units,

• EPA - 2019-04-I-PA-R2 - Revise 40 C.F.R. Part 68 (EPA Risk Management Plan) to require new and existing petroleum refineries with HF alkylation units to conduct a safer technology and alternatives analysis (STAA) and to evaluate the practicability of any inherently safer technology (IST) identified

• EPA - 2019-04-I-PA-R3 - Per the requirements in EPA Rule Procedures for Prioritization of Chemicals for Risk Evaluation Under the Toxic Substances Control Act, initiate prioritization to evaluate whether hydrofluoric acid is a High-Priority Substance for risk evaluation,

• API - 2019-04-I-PA-R4 - Update API RP 751 Safe Operation of Hydrofluoric Acid Alkylation Units to require protection of critical safeguards and associated control system components, as well as the installation of remotely-operated emergency isolation valves on the inlet(s) and outlet(s) of all hydrofluoric acid containing vessels, and hydrocarbon containing vessels meeting defined threshold quantities, and

• ASTM - 2019-04-I-PA-R5 - Revise ASTM A234 to incorporate supplementary requirements for piping used in HF service, as defined in HF supplementary requirements S9.1 through S9.7 in ASTM A106 version 19a.


While I am a chemist and not a chemical engineer, it seems to me that the CSB failed to call out a seemingly obvious chemical engineering mistake made in the HF water spray system. I was taught in a manufacturing environment that it was important to consider the failure mode of systems early during the design process. At its most basic, consider a powered valve. If power to the valve fails, then one of two states can result the valve can ‘fail closed’ or it can ‘fail open’ depending on how the valve is configured. If safety requires flow through that valve in the event of an emergency, then it should fail open upon the loss of power, otherwise the valve should be configured to ‘fail closed’. In this case, it is apparent that when power failed to the valves controlling the HF Spray system, they failed closed and had to be opened manually. On the face of it, this was poorly considered in the design process.

Now there may have been (almost certainly were) other considerations involved, but the safety system with the highest criticality should be the controlling factor in safety design. Relying on a control system to continue working properly in a refinery during a major incident is just asking for problems.

One other comment, this incident included a major release of HF (5,289-lbs by PES estimate – page 6) and only a small part of it (1,968-lbs) was sequestered by the late functioning water spray system. And no one was injured by HF. Folks that are opposed to mandatory IST implementation might be tempted to point at this incident as an example of why IST is unnecessary, no one was hurt why spend that money? Those people, however, need to remember that this incident included a large, unplanned, hazardous-waste incineration event and that was almost certainly responsible for the neutralization of the released HF. No refinery operator wants to have to rely on a large conflagration to control an HF release.


For more details about the report, see my article at CFSN Detailed Analysis - - subscription required.

1 comment:

Rosearray said...

Patrick, I posted your commentary of 10/12/22 on the PES Refinery Investigation Report on AIChE Engage Discussion Central, and today got the following response:

In reply to Patrick Coyle's comments:

Valves may also be designed to fail in their last position, not just open or closed
The preference for making the deluge valves fail open may seem obvious when viewed through the lens of one particular scenario. But one must also consider the consequences in other circumstances, such as during normal operation in cold weather. If the deluge valve failed open for any reason, the entire area could be coated in a thick layer of ice, and that may also create conditions adverse to process and personal safety.
So the choice may not be as "obvious" as Mr. Coyle believes, but he is certainly correct in his statement that other considerations are involved, and in the need to design for reliable activation of emergency systems when conditions warrant.

Don Lorenzo, PE

/* Use this with templates/template-twocol.html */