This case is not final orders of the Review Commission as it is pending Commission Review!
This decision can have HUGE impacts throughout the PSM/RMP community, as many covered processes are “interconnected” to a boiler and many may also be “co-located” such that a boiler explosion could impact the covered process. This case involved a boiler that used “refinery fuel gas” as its main fuel source; the RFG was a collection of unusable fuels from the refining processes (e.g. PSM covered processes) that were dumped into a header and sent to the boiler as a fuel. The refinery argued against all three of OSHA’s positions: 1) interconnectivity, 2) not used SOLELY as a fuel, and 3) co-location in an attempt to say the boiler was not part of a PSM-covered process and the court disagreed with every aspect of their argument(s). Both OSHA and ALJ both viewed the steam this boiler generated as a “critical utility” (my term) used in “safety systems” such as inerting, fire extinguishment, snuffing, etc. So in the end, the ALJ agreed with OSHA that the boiler was “connected” to a PSM-covered process AND it was “co-located”and was NOT exempted under the “used solely as a fuel”, making this boiler part of the PSM-covered process.
On September 28, 2012, a boiler exploded during a turnaround at a Refinery killing two employees. In response, OSHA initiated an inspection of the Refinery on September 29, 2012. On October 29, 2012, OSHA initiated a second, simultaneous inspection of the worksite in response to complaints about the conditions in the refinery warehouse. As a result of the inspections, OSHA issued two separate Citations and Notifications of Penalty (“Citations”) to Respondent. The Citation for Inspection No. 663538 alleges one other-than-serious, eleven serious, and five repeat violations of the Act, with a total proposed penalty of $234,500.00. The Citation for Inspection No. 778042 alleges one repeat, one other-than-serious, and eleven serious violations of the Act, with a total proposed penalty of $46,600.00.
NOTE: this article will focus on Inspection No. 663538 which involved a boiler explosion and the PSM-covered process.
Factual Background
The refinery is a 70,000 barrel-per-day (bpd) crude oil refinery, which produces gasoline, propane, propylene, butane, fuel oils, and solvents. The refinery is broken into separate zones, each of which performs a different function in the refining process. The citation items in Docket No. 13-0791 (Inspection No. 663538) focus on Zone 2. Zone 2 contains the Alkylation Unit, the Fluid Catalytic Cracking Unit (FCCU), and the Wickes boiler, which caused the explosion and prompted the inspections leading to this litigation. The citation items in Docket No. 13-0644 (Inspection No. 778042) focus on alleged violations in the warehouse, as well as general safety items identified throughout the refinery.
The explosion that killed two employees on September 28, 2012, originated at the Wickes boiler, which is part of the FCCU located in Zone 2. The Wickes, as described by many of Respondent’s employees, “was by far the workhorse of the plant for steam.” It is one of four boilers that provide steam to the 225-pound steam header, which, in turn, routes steam for use in various processes throughout the plant. Some of those processes include providing emergency steam to the riser, which clears it of HHCs; injecting steam into the FCCU process to drive high-end products out of the crude oil, also known as steam-stripping; purging low-lying gases in the firebox of the Alky Unit heater during start-up; powering turbines to pump product; and serving as a back-up to the electric pumps. The Wickes itself is fueled by two separate fuel streams within the refinery—the refinery fuel gas (RFG) system and a natural gas fuel line. The RFG system, which is the primary source of fuel for the Wickes, is a fuel recycling system of sorts. Various processes throughout the plant, such as the FCCU, refine crude oil to a saleable product. As a result of these processes, a certain amount of non-condensable flammable gas remains. Though this gas cannot be converted into a saleable product, the refinery still uses it to fuel various processes throughout the plant.
These “off-gases” that are produced throughout the plant are directed via pipeline to a fuel drum, where the refinery fuel gas is treated. The resulting gas mixture is then piped out of the drum into a 4.1-mile pipeline network that leads to different processes throughout the refinery, including the Wickes. Such is the process for normal operations; however, in some instances, such as during a turnaround, the Wickes can be powered by natural gas alone.
In order to start the Wickes boiler, Respondent had to go through a fairly detailed process, which involved no fewer than three employees. The first step requires the CT to purge HHCs from the boiler’s firebox for 30 minutes by blowing air into it. Once the firebox has been adequately purged, the pilot light has to be lit. After the pilot is lit, an operator is directed to open the fuel gas bypass valve, which introduces the RFG mixture into the firebox. Each operator that testified gave a slightly different description as to how this part of the process is carried out. For example, one stated that he was told to turn the bypass valve “one-quarter of a spoke” and to leave it open for 5–10 seconds, though he admitted there was not a set amount of time to keep the valve open. Another testified that he was trained to open the valve “slightly” or “just a little bit” and to close the valve if he did not achieve ignition “quickly” or “shortly”. During this process, another operator positions himself at the sight glass, which allows him to determine whether there has been a successful ignition.
With the exception of a shutdown because of a turnaround or annual boiler inspection by the State, the Wickes was operated constantly. As such, there were limited opportunities for operators and CTs to light the boiler. According to a shift supervisor, however, it was “not uncommon” for operators to experience a “hard start” when attempting to light the Wickes. A hard start is best characterized as a mini-explosion occurring within the firebox, typically a result of allowing too much fuel into the system. An operator stated that, instead of lighting smooth, a hard start causes the boiler to “woof” or “huff” as a result of a sudden pressure increase within the firebox. In some cases, this merely caused the boiler to spew dust and smoke; in others, the structure of the boiler actually bowed outward as a result of the explosion. In one instance, a worker, who was manning the sight glass, was actually struck by the boiler, which had bowed outward during a hard start.
The Turnaround
On September 28, 2012, Respondent was in the middle of a refinery turnaround. In order to facilitate repairs and maintenance, Respondent needed to produce steam to purge HHCs from various lines and equipment. This required starting up the Wickes boiler. Because the refinery was off-line and not producing fuel products, Respondent had to use natural gas to light the Wickes. According to a former console technician (CT), the Wickes had been taken off-line earlier that day to make a switch of the electrical supply circuits. Due to power supply problems, the crew implemented a temporary solution by running an extension cord to a small generator, which powered the controls and interlocks of the Wickes.
According to another CT, the previous CT reported that the temporary power supply was causing the vanes, which control air flow, to malfunction. Eventually the problem was fixed, and the CT began to increase the airflow to purge the Wickes firebox of any remaining HHCs.
At the conclusion of the purge, which lasted about five minutes, the CT reduced the airflow to 15,000 cubic feet per minute (cfm) to light the pilot. While the CT was working in the control room, two (2) Lead Operators and two (2) “A” Operators; and a “B” Operator were located at or around the boiler. The Lead Operators were located at the northwest corner of the Wickes and were overseeing the lighting attempt. One was positioned at the fuel bypass valve, and one was positioned at the sight glass to verify ignition. The other two operators did not have specific responsibilities related to the lighting process.
After the firebox had been purged, an operator opened the fuel gas bypass valve to introduce natural gas into the firebox. At some point in the process, one operator instructed another to close the valve because they had not achieved ignition. The operator did not comply with this instruction. When the operator confronted him, he informed the lead that he was taking instructions from the CT, who was standing nearby. After a brief interaction between the two, the lead moved to the north side of the boiler to check water levels. The CT continued to oversee the operator, who kept the bypass valve open.
As fuel was being introduced into the firebox, the other CT in the control room, looked at the CT’s console and noticed that the firebox was flooded with too much natural gas. He immediately radioed the operators to inform them that they should close the bypass valve. Shortly after the operator closed the valve, the boiler exploded. The operator, who was manning the sight glass, was pronounced dead at the scene, and the other operator, who was critically injured in the explosion, died twenty-eight days later.
Subsequent investigations by Respondent revealed shrapnel in the area surrounding the Wickes, and a ladder, which was attached to the west end of the boiler, that had been blown completely across the street. Additionally, investigators found that the valve was opened approximately one-and-a-half spokes and that fuel had been flowing into the firebox for approximately 5 minutes. Many of the operators and CTs involved in the lighting were disciplined, and one of the Lead Operators, was discharged.
Discussion
Prior to answering the question of whether any particular standard was violated, the Court must resolve two important issues.
First, the Court must determine whether the PSM standards cited by OSHA apply to the Wickes boiler. Respondent contends that, by including the boiler within the ambit of the PSM standard, OSHA has improperly expanded the scope of the standard beyond its intended purpose, which is to “prevent[] or minimiz[e] the consequences of catastrophic releases of toxic, reactive, flammable, or explosive chemicals”. 29 C.F.R. § 1910.119.
Consistent with that assertion, Respondent argues that although the boiler is physically connected to a PSM-covered process, the boiler itself is not subject to the PSM standards because it cannot contribute to, cause, or interfere in the mitigation of a catastrophic release of HHCs. See Delek Refining Co., Ltd., 25 BNA OSHC 1365 (No. 08-1386, 2015).
In response, OSHA has asserted multiple theories of coverage to suggest that the boiler would have just such an impact.
Second, the Court shall also address the issue of successor liability in the context of the repeat violations issued to Respondent. NOTE: this part of the case will be covered in a separate article/posting!
Ultimately, based on what follows, the Court finds that the PSM standard did apply to the Wickes boiler. OSHA’s application of the standard under this set of facts comports with its plain language and is consistent with its historical interpretation of the standard.
A. PSM Coverage
i.The Standard – 29 C.F.R. § 1910.119
The stated purpose of the PSM standard is to “prevent[] or minimiz[e] the consequences of catastrophic releases of toxic, reactive, flammable, or explosive chemicals.” 29 C.F.R. § 1910.119.
A catastrophic release, according to the standard, is “a major uncontrolled emission, fire, or explosion, involving one or more highly hazardous chemicals, that presents a serious danger to employees in the workplace.” Id. § 1910.119(b). The standard sets a threshold quantity of various hazardous chemicals – that threshold quantity (TQ) represents the point at which a particular chemical is considered capable of producing a catastrophic release. See id. § 1910.119(a)(1).
In this case, the operative question is whether the Wickes boiler is a part of a “process which involves a Category 1 flammable gas (as defined in 1910.1200(c)) or a flammable liquid with a flashpoint below 100°F (37.8°C) on site in one location, in a quantity of 10,000 pounds (4535.9 kg) or more” such that the standard applies. Such a determination would establish a prima facie case for coverage; however, the Court must also decide whether the exception for HHCs “used solely for workplace consumption of fuel” applies. See id. § 1910.119(a)(1)(ii)(A).
As the title implies, the focus of this standard are processes involving highly hazardous chemicals. Insofar as a process involves a threshold quantity of HHCs, it is covered, subject to certain exceptions. A process, according to the standard, is:
[A]ny activity involving a highly hazardous chemical including any use, storage, manufacturing, handling, or the on-site movement of such chemicals, or combination of these activities. For purposes of this definition, any group of vessels which are interconnected and separate vessels which are located such that a highly hazardous chemical could be involved in a potential release shall be considered a single process. Id. § 1910.119(b).
This definition indicates
(1) the basic understanding of a “process” and
(2) the potential boundaries for that process.
This was explained in more detail in the preamble to the standard, which states:
The term “process” when used in conjunction with the application statement of the standard establishes the intent of the standard. The intent of the standard is to cover a “process” where the use, storage, manufacturing, handling or the on-site movement of a highly hazardous chemical exceeds the threshold quantity at any time. The boundaries of a “process” would extend to quantities in storage, use, manufacturing, handling or on-site movement which are interconnected and would include separate vessels located such that there is a reasonable probability that an event such as an explosion would affect interconnected and nearby unconnected vessels which contain quantities of the chemical that when added together would exceed the threshold quantity and provide a potential for a catastrophic release. In order to clarify this intent, a new sentence has been added to clarify the fact that interconnected and nearby vessels containing a highly hazardous chemical would be considered part of a single process and the quantities of the chemical would be aggregated to determine if the threshold quantity of the chemical is exceeded.
Process Safety Management of Highly Hazardous Chemicals, 57 Fed. Reg. 6356, 6372 (Feb. 24, 1992).
ii. OSHA’s Theories of PSM-Coverage
OSHA (OSHA) asserts multiple bases upon which the Wickes boiler should be considered a PSM-covered process. Specifically, OSHA asserts that
(1) the Wickes is interconnected to a covered process through the refinery fuel gas system and steam header;
(2) the Wickes is located such that an HHC could be involved in a potential release involving other PSM-covered equipment;
(3) the exception for workplace fuel consumption does not apply; and
(4) Respondent treated the Wickes in its own internal documentation as a PSM-covered process.
In response, Respondent contends that
(1) OSHA’s interpretation of the standard improperly expands the scope of what is considered a “process”;
(2) the Wickes is not sufficiently close to PSM-covered equipment such that it could be involved in a potential release;
(3) the workplace fuel exception clearly applies; and
(4) the fact that it applied PSM-related practices to the Wickes is only reflective of “best practices” and not an admission of coverage.
1. Interconnection
The Wickes boiler, viewed in isolation, is NOT a PSM-covered process. There is no single point in time where it processes, uses, or holds a threshold quantity of HHC. Thus, the determination of whether it is covered necessarily depends on its connection or location relative to other covered processes. The dispute over interconnection stems from the second sentence of the definition of the term “process”, which states that “any group of vessels which are interconnected and separate vessels which are located such that a highly hazardous chemical could be involved in a potential release shall be considered a single process.” 29 C.F.R. § 1910.119(b) (emphasis added).
OSHA asserts that the definition establishes two separate bases upon which coverage can be established:
(1) interconnected vessels; and
(2) separate vessels located such that an HHC could be involved in a potential release.
Respondent contends, however, that the modifier “such that a highly hazardous chemical could be involved in a potential release” is applicable to both separate and interconnected vessels, thereby grafting an additional burden of proof for establishing PSM coverage under a theory of interconnection. The Court DISAGREES.
The Court DISAGREES.
This dispute stems from what is known as the Motiva Response, which was a formal interpretation issued by OSHA in response to Motiva Enterprises., LLC, 21 BNA OSHC 1696 (No. 02-2160, 2006). (Ex. C-3). See also Interpretation of OSHA’s Standard for Process Safety Management of Highly Hazardous Chemicals, 72 Fed. Reg. 31453 (June 7, 2007). In Motiva, the Commission grappled with what it believed to be an undefined term within the PSM standard’s TQ requirements for flammables; namely, what constituted “on site in one location”. Motiva, 21 BNA OSHC 1696. Due to the lack of clarity within the application paragraph, and less than convincing evidence, the Commission vacated the citation and placed the onus on OSHA to offer an “authoritative interpretation” that would be reviewed in future cases under “standard deference principles.” Id. at *4.
In response, OSHA issued a formal interpretive document in the Federal Register.
First, OSHA agreed that the language “on site in one location” in the application paragraph has considerable overlap with the definition of process. This was due, in part, to the fact that the definition of “process” was revised in the final rule to clarify that a single process includes both interconnected and co-located vessels, depending on proximity. Due to this change, OSHA noted that “the limitation placed on application of the standard to flammable liquids and gases denoted by the related phrase ‘on site in one location’ no longer carries the independent weight it had before OSHA clarified the intended meaning of ‘process.’” (Id.). However, its import was not entirely diminished, as “it continues to serve a separate purpose by operating to exclude coverage where the HHC threshold would only be met only if all amounts in interconnected or co-located vessels were aggregated but some of the amounts needed to meet the threshold quantity are outside the perimeter of the employer’s facility.”
Second, and more pertinent to this case, OSHA clarified the burden of proof relative to interconnected versus co-located processes by stating that the PSM standard “presumes that all aspects of a physically connected process can be expected to participate in a catastrophic release.” With respect to co-located processes, however, OSHA must prove that they are located such that a hazardous chemical could be involved in a potential release. Respondent takes issue with this formulation because it believes that such an interpretation is “in direct contradiction” with the plain terms of the standard.
Respondent’s primary argument in this regard is that “[t]he absence of punctuation between the term “interconnected” and “separate” establishes that the requirement that an HHC could potentially be involved in a release applies to both interconnected and co-located equipment.”
First, it is not clear what sort of punctuation Respondent is referring to.
Second, the basic structure of the sentence belies Respondent’s argument regarding plain meaning. The sentence describes two configurations on either side of the conjunction “and” and concludes that either configuration constitutes a “process” for the purposes of the PSM standard.
The first configuration is “any group of vessels which are interconnected”.
The second configuration is “separate vessels which are located such that a highly hazardous chemical could be involved in a potential release.”
In both cases, the noun is described through the use of a dependent clause, indicated by the term “which”. In other words, there is a basic, parallel structure on either side of the “and”, which can be diagrammed as follows:
“For the purposes of this definition, [A’s] which are [x] and [B’s] which are [y] shall be considered [C].” See 29 C.F.R. § 1910.119(b).
When analyzed in this way, the Court finds that OSHA’s interpretation, as expressed through the Motiva response, comports with the plain meaning of the definition.
Let us assume, however, that Respondent is correct to the extent that the definition of process is ambiguous. If a determination cannot be reached based on the text and structure of the regulation, courts then turn to “contemporaneous legislative histories of that text.” On such contemporaneous legislative history is the preamble to the final rule. See generally 57 Fed. Reg. at 6356; see also 72 Fed. Reg. 31453.
The preamble provides a clear distinction between interconnected and separate vessels:
“The boundaries of a ‘process’ would extend to quantities in storage, use, manufacturing, handling or on-site movement which are interconnected and would include separate vessels located such that there is a reasonable probability that an event such as an explosion would affect interconnected and nearby unconnected vessels which contain quantities of the chemical that when added together would exceed the threshold quantity and provide a potential for a catastrophic release.” 57 Fed. Reg. at 6372.
This discussion, which provides contour to the definition of process, makes clear that the term “process” extends to interconnected vessels and includes separate vessels, insofar as such vessels could reasonably be expected to participate in a catastrophic release. Given this explanation, the Court still finds that the standard presumes the potential for a catastrophic release when vessels are physically connected.
Finally, even if the preamble is somehow considered deficient in its clarification, the Court finds that the interpretation espoused by OSHA is both reasonable and consistent with its longstanding interpretation of the issue. There is nothing patently unusual or unreasonable about considering vessels that are physically connected by pipeline to be part of the same process, nor is it unreasonable to presume that vessels connected in such a way could be involved in a potential release of HHCs. This has been OSHA’s interpretation of the standard since its inception. (Ex. C-4). Accordingly, the Court finds that OSHA’s interpretation of the standard is reasonable and, therefore, entitled to deference.
According to the P&IDs involving the Wickes boiler, it is physically interconnected to otherwise-covered PSM processes in two ways.
First, the Wickes boiler is connected to both the Alky Unit and the FCCU through the RFG pipeline. It is undisputed that the Alky and the FCCU are PSM-covered processes by virtue of the quantity of flammables contained in each.
Second, the Wickes is connected to virtually all of the refinery’s processes through the 225-lb. steam header. As such, OSHA has, at the very least, established a prima facie case for PSM coverage, because interconnected processes are presumed to have the potential to participate in a catastrophic release.
However, such a presumption could be rebutted by a showing that the interconnected processes at issue could not participate in or contribute to a catastrophic release. Perhaps anticipating the potential failure of its argument regarding the presumption associated with interconnected processes, Respondent also argues that the Wickes should not be considered interconnected to a covered process under the terms of the standard.
First, Respondent suggests that the Wickes is not a “vessel” because it does not store or contain any measurable quantity of HHC.
Second, Respondent argues that the concept of interconnection, as espoused by OSHA, does not merely equate to a physical connection between equipment; rather:
[T]he concept of interconnectivity is merely intended to address a situation in which connected vessels within a single process that contain quantities of HHC, such as flammable gas storage tanks, will be deemed to satisfy the threshold requirement even though the amount of flammables in each individual vessel is less than 10,000 pounds.
This theory does not operate to extend coverage to any structure, regardless of its form or contents, that is physically connected to a PSM-covered process.
The Court DISAGREES.
The Commission dealt with a similar attempt to narrow the scope of the standard in Delek Refining Co., Ltd. In that case, OSHA alleged that the employer violated a portion of the PSM standard by failing to inspect and test its positive pressurization unit (PPU) in the control room of its own FCCU. The PPU was designed to pressurize the control room to prevent hazardous vapors, which are a byproduct of the FCC process, from entering the control room and poisoning the employees inside or causing an explosion hazard due to the presence of wiring, which could serve as an ignition source. Delek contended that the PPU was not “process” equipment, because it was not directly involved (physically connected) in the process of converting crude oil into usable fuel.
Although the specific subsection of the PSM standard at issue in that decision was different, the Commission still had to address the question of what constitutes the boundaries of a process. The Commission made it clear that the focus of the standard—the process—was not as narrow as suggested by Respondent. According to the Commission:
[T]he PSM standard does not require that every part of a ‘process’ come into contact with hazardous materials. 29 C.F.R. § 1910.119(b) (defining ‘process’ as ‘any activity involving a highly hazardous chemical”) (emphasis added). Here viewing the ‘activity’ involving the FCC unit in its entirety, the PPU is part of a ‘process’ covered by the PSM standard because it is an integral part of the ‘manufacturing, handling [and] onsite movement of [highly hazardous chemicals].”
Citing favorably to an OSHA Interpretation Letter from Richard Fairfax to Howard J. Feldman, the Commission noted that machinery not containing HHCs can nonetheless be a part of a process insofar as such machinery is used to control, prevent, or mitigate catastrophic releases.
The Court finds that Respondent places undue emphasis on individual terms such as “vessel” and “interconnected” at the expense of the focus of the standard as a whole—the process. As noted by the Commission, the definition of “process” is broad— it is any activity involving a HHC, including any use, storage, manufacturing, handling, or on-site movement. See 29 C.F.R. § 1910.119 (emphasis added). So broad, in fact, that the Commission held that the PPU in Delek’s FCC control room, although not a vessel, was part of the FCC process because it could affect or cause a release.
In this case, the connection between the Wickes and the FCCU is more concrete:
The off-gases produced by the FCCU are directed via pipeline to a fuel drum, which mixes the off-gases treats them and directs the resulting product to the Wickes.
The Wickes is clearly an activity that involves an HHC because it uses the treated off-gases from various processes around the refinery. It is, in fact, a downstream endpoint of the RFG process.
During normal operations, there are multiple processes that feed the RFG system, including the FCCU and the Alky Unit.
These processes, with the exception of a turnaround, are basically running all the time.
As the Court observed during the trial, the bypass valve that controls the flow of RFG can apparently be left open indefinitely without an alarm – it was not until the CT happened to look over the shoulder of another CT and noticed a large amount of fuel in the firebox that the order was given to shut it down. Further, the Wickes, which all witnesses testified is almost always running, requires a constant stream of fuel. Thus, even if the Court accepts Respondent’s assessment of the RFG pipeline’s capacity, which it determined to be 860 pounds of fuel, that assessment disregards the source of the fuel, such as the FCCU and the Alky, which feed the RFG system and are directly connected to the Wickes.
The Court cannot find any basis in the regulatory history or the language of the standard itself that would suggest such an arbitrary determination of what is interconnected. Accordingly, the Court finds that the Wickes was interconnected to a covered process, and, as such, should be considered a single process. See 29 C.F.R. § 1910.119(b).
Hydrocarbon fuels used solely for workplace consumption as a fuel
Respondent further contends that even if the foregoing is true, the Wickes should still be exempt from coverage. According to section 1910.119(a)(1)(ii)(A), the following are exempted from PSM coverage:
“Hydrocarbon fuels used solely for workplace consumption as a fuel (e.g., propane used for comfort heating, gasoline for vehicle refueling), if such fuels are not a part of a process containing another highly hazardous chemical covered by the standard.” Id. § 1910.119(a)(1)(ii)(A).
The intended scope of this rule was described in the preamble to the standard, wherein the American Petroleum Institute noted that OSHA’s intention in providing exemption (b)(1)(ii)(A) was to exclude the enormous number of small business locations across the nation which would not be covered by the proposed rule, except for their on-site storage of hydrocarbon fuels for low-risk applications such as heating, drying, and the like. Such activities are not the subject of this rule, and this exclusion is entirely appropriate. On the other hand, interpreting this exclusion to apply to hydrocarbon fuels used for process-related applications such as furnaces, process heaters, and the like at facilities covered by the rule was not intended. 57 Fed. Reg. 6356, 6367.
At the very outset, this exception had a very limited scope:
small businesses that used on-site hydrocarbon fuels “for low-risk applications such as heating, drying, and the like.” Id. (emphasis added).
The exception was not, however, intended to cover process-related applications such as process heaters and furnaces. It is significant that this was recognized by API, which promulgates consensus standards covering the petroleum industry. See, e.g., 72 Fed. Reg. 31453, 31454 (citing API 750 as the basis for the definition of “process”). Thus, the issue is, again, one of degree:
Is the Wickes, as compared to process heaters and furnaces, which are explicitly not covered under the exception, properly considered a part of a process involving another highly hazardous chemical covered by the standard?
Respondent contends that furnaces and heaters are more directly linked to a process than a boiler, because furnaces and heaters typically apply heat directly to a product, whereas a boiler merely supplies steam to a header, which directs that steam to various processes around the refinery.
The Court is not convinced by the furnace versus boiler distinction urged by Respondent, nor is it convinced that the workplace fuel exception applies. Though the preamble mentions furnaces and process heaters as specific process-related applications, the list is not exhaustive, but exemplary.
See 57 Fed. Reg. 6356, 6367 (exception does not cover “furnaces, process heaters, and the like”) (emphasis added).
To the extent that process heaters, furnaces, “and the like” are the examples of what is not covered by the exception, and considering Respondent’s argument that there is a qualitative difference between the manner in which a furnace is connected to a process, as opposed to a boiler, the Court will address the manner in which the Wickes is connected to other PSM-covered processes and determine whether that connection is sufficient to establish PSM coverage.
While the Wickes is the downstream endpoint for the Refinery Fuel Gas (RFG) system, it is also a starting point for many other process-related applications. The Wickes’ core function is to produce steam. That steam is used in multiple process-related applications throughout the refinery. For example:
(1) When the FCC emergency shut down (ESD) system is activated, steam is directed to the riser, where it knocks down gases to prevent further catalyzing of crude oil;
(2) Steam is used as a catalyst in certain reactions, driving high-end products from crude oil, also known as steam-stripping;
(3) In normal operations and emergencies, steam was used as a primary source to drive turbines that pumped product and as a back-up to electric pumps;
(4) In the Alky, steam is used to snuff out low-lying gases and purge fugitive HHCs from the heater prior to lighting it (in much the same way that air is used to snuff gases in the Wickes);
(5) Steam is used as a heat medium in an exchanger, which transfers heat to a process; and
(6) Steam hoses are used to put out small fires in a process pipe.
On the face of it, all of these applications are a process-related to some degree. Nonetheless, Respondent contends that the steam producing system is a mere utility and that it has specifically determined that “the boiler could not cause or interfere in mitigating the consequences of a catastrophic release.” Resp’t Br. at 31 (citing Ex. R-84).
Respondent, much like the employer in Delek, urges a narrow view of the concept of process-relatedness. In Delek, the employer cited an OSHA Interpretation Letter, which contained language stating that “‘[t]he boundaries of the covered process are based on the equipment which contain [highly hazardous chemicals].’”
Delek, 25 BNA OSHC 1365 at *7. The Commission disagreed with such a narrow reading, focusing on the following language:
OSHA does not agree that utility systems are categorically outside the scope and application of the PSM standard. It is OSHA’s long-standing position that utility systems are part of the PSM-covered process when employers use them to control/prevent and mitigate catastrophic releases… .
* * *
[T]he proper safe functioning of all aspects of a process, whether they contain [highly hazardous chemicals] or not, are important for the prevention and mitigation of catastrophic releases of [highly hazardous chemicals], due to their direct involvement in the overall functioning of the process. As a result, it is OSHA’s position that if an employer determines that a utility system or any aspect or part of a process which does not contain a [highly hazardous chemical] but can affect or cause a release… then, relevant elements of PSM could apply to these aspects. OSHA’s position is that any engineering control, including utility systems, which meets the above criteria must be… inspected/tested/maintained per OSHA PSM requirements. Id. at *8–9.
Here, Respondent cites the same letter, in addition to another interpretive document, for the essentially the same proposition. (Ex. R-83, R-84). Specifically, Respondent contends that, notwithstanding the numerous ways in which the Wickes is connected to various covered processes, it has analyzed those connections and specifically determined that the failure of the Wickes would not “cause an HHC release or interfere with the consequences of an HHC release…” (Ex. R-84).
Like the employer in Delek, Respondent places significant emphasis on the “if an employer determines” language to argue that the determination of the boundaries of a PSM-covered process “is the responsibility of the employer, not OSHA.” Resp’t Br. at 30. While there is no doubt that the PSM standard is a performance standard, which allows an employer some discretion as to how a particular hazard should be addressed, “there is no indication in the language of the PSM standard or its regulatory history that OSHA meant to give to employers, at their sole discretion, the option of excluding equipment from the standard’s coverage.” Delek, 25 BNA OSHC 1365 at *9. Thus, the interpretive letter states that if an employer makes a determination that a component failure in the utility system cannot affect, cause, or interfere in the mitigation of a potential release, the employer must be able to proactively demonstrate why the utility system is no longer a part of a covered process. (Ex. R-83). In other words, the determination must be reasonable. See, e.g., Siemens Energy & Automation, Inc., 20 BNA OSHC 2196 at *1 (No. 00-1052, 2005) (performance standard give a “certain degree” of discretion but meaning of standard interpreted in light of what is reasonable).
Respondent argues that it conducted the analysis of the Wickes as described above and concluded that the boiler could not cause or interfere in mitigating the consequences of a catastrophic release. Specifically, Respondent points to the testimony of the PSM Manager, who states that he considered the failure of the Wickes and other aspects of the 225-pound steam system as part of his analysis of a loss of heat to a covered process. The PSM Manager and the current VP/GM of Respondent’s KS refinery and former operations manager at the Refinery in question concluded that a failure of the Wickes would not have such an effect because the other boilers that sourced the steam system could produce sufficient steam to continue operations at the refinery and that any temporary effects would only impact product quality. Respondent’s expert testified similarly.
This determination, Respondent contends, was reinforced by the record evidence, including:
(1) the Wickes was taken offline once per year for an annual inspection;
(2) the refinery had redundancies in place such that only two of the four utility boilers were needed to contribute steam to the header.
Further, Respondent also argues that the snuffing steam system, as used in the Alky heater firebox, was only for small fires and that no evidence was presented to show that such a fire could cause a catastrophic release of HHC.
The Court has a different perspective on the record evidence, as well as the sufficiency of Respondent’s determinations regarding the impact of a loss of steam on PSM-covered processes.
First, the PHA/Hazop analysis performed by the PSM Manager was, according to his testimony, focused on the impact of too little or too much heat being supplied by the Wickes and how that could cause a loss of containment. In response to a question regarding whether he was confident that he considered a loss of steam in all PHAs for covered processes, he stated, “I’m confident in that based on the questions you have to ask yourself in a HAZOP of too much heat or too little heat. And steam provides heat to our processes.” The problem, however, is that the functions described above are not limited to supplying heat to a particular process. It is also used to snuff out fires, remove HHCs from the FCC riser in emergencies, and purge HHCs from furnace fireboxes during the lighting process. Based on the Court’s review of the PHAs, there is no indication that the impact on these safety functions was considered.
Second, in an attempt to downplay the significance of the Wickes, the current VP/GM of Respondent’s KS refinery and former operations manager at the Refinery in question noted that it is one of four boilers on location at the refinery and that there is a redundancy system built in to reduce the refinery’s reliance on any one boiler. While this may be the case, there was no independent evidence, by way of PHAs or SOPs, to indicate that the system was designed this way. Further, Respondent’s employees testified that the Wickes was the workhorse of and a main contributor to the plant’s steam system. Respondent lent credence to that characterization by choosing the Wickes as the boiler of choice for the turnaround. Respondent recognized that problems with the Wickes and connected steam system could lead to process upsets. While those upsets likely had the most direct impact on product quality, there was also testimony that such upsets may also impact the use of certain safety measures associated with the steam system.
That the safety measures associated with a covered process could be affected by a boiler system upset is alone sufficient to warrant finding a connection sufficient to establish the inapplicability of the exception. See Delek, 25 BNA OSHC 1365 at *8 (citing favorably to OSHA Interpretation Letter stating “proper safe functioning of all aspects of a process, whether they contain [HHC] or not, are important for the prevention and mitigation of catastrophic releases”). Just because a redundancy system is built in does not remove a particular boiler from the ambit of the standard. See id. (“OSHA’s position is that any engineering control, including utility systems, which meets the above criteria must be… inspected/tested/maintained per OSHA PSM requirements.”).
The key is the connection to the process, and whether a failure in that connection could have an impact on a potential catastrophic release of HHCs. As testified to by the current VP/GM of Respondent’s KS refinery and former operations manager at the Refinery in question, certain process upsets, if left alone for a long enough, can cause a catastrophic release. For example, what if the emergency shutdown system in the FCC Riser cannot be activated because the purported steam redundancy system failed? Under such a set of circumstances, surely it would be reasonable to conclude that a failure at the Wickes would have an impact on the system’s ability to control, prevent, and/or mitigate a catastrophic release.
As noted above, the Commission in Delek determined that the PPU in the control room was governed by the PSM standard. The PPU did not have a direct connection to the process; rather, it was a control to prevent the spread of harmful gasses that were a result of the FCC process, which could, in turn, prevent the control room from managing the refining process. Delek, 25 BNA OSHC 1365 at *8.
The connection of the Wickes to various processes throughout the plant was not nearly so attenuated. The Wickes provided steam, which was used directly in the various PSM-covered processes throughout the plant in both a production – and safety – related capacity. In its safety-related capacity, the steam provided by the Wickes served to control, prevent, and/or mitigate catastrophic releases through its use as a snuffing and purging agent. While such uses may not be a complete or sufficient control in and of themselves, the Court finds that such a connection is sufficient to bring the Wickes under the umbrella of the PSM standard.
At a very basic level, the Wickes connected to PSM-covered processes on the front and back end:
It is fueled by off-gases from the FCCU and Alky, and, in turn, it supplies steam to those same processes. The explosion in this case provides a clear example of how physical connections between processes can lead to a catastrophic release. There was no independent, automatic control that could stop the flow of fuel to the Wickes during the lighting process; the explosion that resulted from flooding the firebox was only mitigated by the fact that a CT happened to notice the overflow of fuel. Independent of that, there was nothing to impede the flow of fuel to the system (although it was natural gas, the same event could have occurred with RFG). Further, to suggest, as Respondent has, that this was a worst case scenario disregards the fact that, but for the CT intervening, gas would have continued to flow to the firebox even after the explosion. In fact, the unit supervisor testified that gas was released into the atmosphere as a result of the explosion, noting a smell of gas in the air.
In light of the foregoing, the Court finds that the Wickes boiler is a critical aspect of multiple PSM-covered processes, is not subject to the workplace fuels exception, and, therefore, was properly cited under the PSM standard under a theory of interconnection.
2. Proximity to a Covered Process
An additional basis for coverage urged by OSHA is that the Wickes, independent of its connections to covered processes, was “located such that a highly hazardous chemical could be involved in a potential release”. 29 C.F.R. § 1910.119(b).
As a result of the explosion, there was significant damage to surrounding equipment, including piping and valves; and the ladder and platform, which were attached to the Wickes, were blown across the street and hit the operator shelter. OSHA contends that, in addition to the damage described above, parts of the FCCU process lines, including the Intercat loader and process pipe racks, could have been impacted by flying shrapnel. Based on its location relative to other aspects of the FCCU process, as reflected in the FCCU Equipment Location Plot Plan, OSHA’s expert, concluded that the Wickes’ location was such that it should be considered part of the FCCU process. See 29 C.F.R. § 1910.119(b).
Respondent contends that the Wickes is not close enough to any covered process such that a highly hazardous chemical could be involved in potential release and, therefore, should not be considered a single process with any adjacent PSM-covered processes, such as the FCCU. See id. Respondent places significant emphasis on the way this particular explosion occurred to support its argument that the Wickes was not sufficiently close to a covered process to be considered a part of that process and therefore covered under the PSM standard. In particular, Respondent points out that the closest aspects of a process that contains any HHC is the FCCU reactor column, which is approximately 100 feet away. Noting that there was no damage to equipment beyond a 10–15 foot radius, and that no release of HHC occurred, Respondent contends that this “worst-case scenario” shows that the Wickes could not participate in a catastrophic release.
The Court DISAGREES.
As noted by OSHA, the Wickes was centrally located in the FCCU Equipment Location Plot Plan. Thus, before any discussion of distance, the Court finds that the Wickes is at least situated such that it could impact co-located, covered processes, i.e., not in some remote location. As to distance, it is true that many of the covered processes are not located within the apparent radius of the blast zone (10–15 feet) as determined by Respondent; however, that assessment disregards one very large piece of shrapnel that traveled much further: the ladder and platform, which were previously attached to the Wickes. As a result of the explosion, the ladder and platform attached to the east side of the Wickes were propelled across the street and hit the operator shelter. Respondent’s expert estimated the distance from the boiler to the shelter was about 40 feet. In addition to the ladder and platform, the photographs also show a significant amount of refractory that had been blasted across the street at the operator shelter. Had the ladder and platform simply been blown in a different direction as a result of the explosion, perhaps toward the FCCU, it is reasonable to assume a catastrophic release would have occurred.
The fact that a catastrophic release from an adjacent PSM-covered process did not actually occur under these circumstances does not, in any way, establish that such an eventuality could not occur. See 29 C.F.R. § 1910.119(b) (deeming as a single process separate vessels “which are located such that a highly hazardous chemical could be involved in a potential release”) (emphasis added). The fact that a larger explosion did not occur is likely attributable to two factors:
(1) the CT noticing the excessive flow of fuel to the firebox and directing the operators to shut it down; and
(2) the Wickes was being fueled by natural gas and was not using the RFG pipeline at the time of the explosion.
The Court is mindful of the fact that the explosion occurred shortly after the order to shut the bypass valve; however, the valve connecting the RFG and natural gas lines to the Wickes were within the blast radius, as exemplified in the photographs taken of the west end of the boiler after the explosion. If the boiler was running on RFG at the time, damage to the fuel lines or simply an inability to turn off the valve after the explosion could lead to a catastrophic release. Although Respondent has argued that the RFG system only contains approximately 1500 pounds of fuel gas at any given time, as noted before, that assessment does not take into consideration the source of that fuel—processes such as the FCCU and Alky. Under normal operations, the Wickes is constantly consuming fuel and the FCCU and Alky are constantly producing it. This constant loop of off-gas production and consumption leads the Court to conclude that Respondent’s attempt to place artificial boundaries on the RFG process such that a covered process would not be affected is misguided and disregards the concrete connection that exists between the Wickes and the FCCU, for example.
Perhaps the strongest justification for deeming the Wickes to be part of a single process, and thus PSM-covered, is the potential impact on the control room. As noted above, the ladder and platform assembly, along with a significant amount of refractory, were blown across the street and into the operator shelter, which housed two CTs. In Delek, the Commission found that the control room (operator shelter) and the controls associated therewith were part of the overall FCC unit process:
Delek’s refining process includes operating the FCC unit as a whole, and this is done from the FCC unit’s control room, which is kept in safe working order by the PPU. Without the PPU providing positive pressure, hydrocarbon vapors could leak into the control room and—because of the wiring there—cause the type of catastrophic explosion that the PSM standard was intended to prevent. And short of such an explosion, the toxic vapors could harm the employees inside the control room, compromising the management of the refining process. We find, therefore, that the PPU is an integral part of the overall FCC unit “process.”
25 BNA OSHC 1365 at *9.
The key point in the passage above is that an incident, such as an explosion at the Wickes, which compromises the management of a PSM-covered process could cause the type of catastrophic event that the standard was designed to prevent. As such, the Commission held that even the positive pressurization unit (PPU), whose connection to a PSM-covered process is even more attenuated than the control room itself, was governed by the PSM standards. Id. at *8 (“The PPU’s regulation of the control room’s positive-pressure atmosphere makes the PPU integral to that “control”—and thus a “control” itself—because, as discussed above, entry of hazardous hydrocarbon vapors into the room could prevent the control room from managing the refining process.”).
Throughout its brief, Respondent was intently focused on whether the putatively impacted process contains a threshold quantity of HHCs. The Commission made it clear that the scope of the standard’s coverage is not so narrow. Instead, the Commission takes a holistic approach to the issue:
“[T]he PSM standard does not require that every part of a ‘process’ come into contact with hazardous chemicals…. [V]iewing the ‘activity’ involving the FCC unit in its entirety, the PPU is part of a ‘process’ covered by the PSM standard because it is an integral part of the ‘manufacturing, handling, [and] on-site movement of [highly hazardous chemicals].” Id. at *7.
The Court sees no difference between the potential impact on the control room in Delek and the circumstances presented here, wherein the control room was actually in the line of fire of the explosion. Respondent was presented with direct evidence that this could be the case in 2008 when Respondent performed a blast study for the FCCU as part of the PHA revalidation of the Wickes. At that time, the refinery determined that the operator shelter adjacent to the Wickes should be pressurized and hardened to meet overpressure requirements. In other words, an integral aspect of a PSM-covered process could be impacted by an explosion at the Wickes. Nevertheless, Respondent maintained its narrow view and concluded that additional measures were unnecessary to protect process vessels and equipment in the FCCU.
Based on the foregoing, the Court finds that the Wickes was located such that an event, like the explosion that occurred in this case, could affect or cause a catastrophic release. Accordingly, the Court finds that the Wickes boiler is subject to the PSM standards under either the interconnection or proximity theory of coverage.
3. Respondent Treated Wickes as PSM-Covered
As further support for its argument that the Wickes was a PSM-covered process, OSHA contends that Respondent essentially treated the Wickes as such. Respondent argues that, to the extent it treated the Wickes as PSM-covered, it only did so as a matter of best practices and that taking additional precautions should not subject it to liability. The Court notes that while Respondent’s treatment of the Wickes, in and of itself, is not sufficient to establish PSM coverage, it undercuts Respondent’s claims that it conclusively determined that the PSM standard did not apply.
OSHA identified the following as examples of the Wickes being treated as part of a PSM-covered process:
(1) In 2008, the Wickes experienced a “hard start”, and the incident report characterized the event as a “PSM Incident”;
(2) the plot plan and various P&IDs for the FCCU include the Wickes; and
(3) Respondent performed Process Hazard Analyses (PHA) and implemented Management of Change (MOC) procedures on the Wickes.
The PSM Manager contends that he inadvertently checked the “PSM Incident” checkbox while inputting the findings of an hourly employee that assisted in the incident investigation and that such documentation does not reflect his or the refinery’s opinion as to PSM coverage. Further, Respondent claims that the PSM Manager determined the Wickes was not PSM-covered when the PHA revalidation for the Wickes was performed.
Contrary to Respondent’s arguments, the Court cannot find any documentary evidence that Respondent made a conclusive determination that the Wickes was not PSM-covered. (“If an employer makes this determination, then, the employer must be able to proactively demonstrate why the utility system is no longer part of the covered process.”).
The problem for Respondent is that the documentation that would normally be used to establish coverage does not reflect the sort of proactive demonstration of non-coverage; rather, as the Court indicated above, the evaluations performed by, or at the request of, Respondent either lack any affirmative determination of non-coverage or should have put Respondent on notice of potential coverage. See Section IV.A.ii.2, supra (discussing blast study and potential impact on adjacent operator shelter). Instead, the PSM Manager testified that Respondent “must have ruled out” that an explosion at the Wickes would impact adjacent processes; however, even he admitted that his conclusion was “pure speculation”.
While it is true that the PSM standard is performance-based, and thus places the onus on the employer to determine how to comply, Respondent has not provided a reasonable basis for its determination. As noted above, the PHA/Hazop analysis performed was focused on the impact of too little or too much heat being supplied by the Wickes and how that could cause a loss of containment. This analysis did not take into account numerous other ways in which a failure of the Wickes could impact other processes to which it was connected, such as snuffing steam in the Alky heater’s firebox and emergency steam to the FCC riser. This narrow view comports with Respondent’s arguments throughout and fails to account for the Wickes’ significant connections to covered processes throughout the refinery.
CLICK HERE for the case file.