If you have been around ammonia refrigeration very long you have most likely heard a horror story or two related to “draining an oil pot”.  This task is also very well know by OSHA and EPA inspectors and is often used as a point of review for operating and maintenance procedures during PSM/RMP inspections.  So to best lay this all out in some logical progression lets first look at the task in a larger picture. (UPDATED on 4/11/12 with ANSI/IIAR 2-2008, Section 14.2 references)

Draining an oil pot WITHOUT A DOUBT involves OPENING THE PROCESS.  There is NO WAY AROUND THIS SIMPLE FACT, but many in the industry still choose to debate this.  Let’s be frank here, in order to get oil out of an oil pot, one has to OPEN the oil pot to do so.  This is CLEARLY what OSHA and EPA had in mind when they mentioned “Process Opening” in the Operating Procedures section of the standard/rule when they mention the “Safe Work Practices”.  So when we open the oil pot there are several risks and hazards at play that MUST BE assessed and mitigated using engineering controls, administrative controls, and Personal Protective Equipment (PPE).

Let’s first look at our PRIMARY MEANS to control a hazard… Engineering Controls.  Now keep in mind, every oil pot may be different, even within the same facility oil pots can be very different!  So one risk/hazard assessment MAY NOT apply to all oil draining tasks.  But for the sake of clarity, let’s say our oil pot is inside a well ventilated engine room.  This means our ventilation system is DESIGNED, INSTALLED and MAINTAINED per a “Recognized and Generally Accepted Good Engineering Practice” (RAGAGEP) such as ASHRAE 15 (specific code for ammonia refrigeration machinery rooms).  Let me be very clear here – this ventilation system is a KEY ENGINEERING CONTROL for this task!  Without it, I would certainly escalate this task to one requiring a Supplied Air Respirator (SAR) AND require a full evacuation of the machinery room to all personnel NOT donned in this level of respiratory protection (more on this later).  But with a properly designed and functioning ventilation system this routine task can be managed in a very different manner. 

The second order of engineering controls we should be using for this task is the “Control of Hazardous Energy” or what we like to call Lockout/Tagout (LOTO).  In most processes, certainly not all, there will be an isolation valve that can be CLOSED AND LOCKED per site LOTO policy that will ISOLATE the oil pot from the refrigeration process.  We should also seriously consider using a DOUBLE BLOCK arrangement if two isolation valves are available.  And if there is a bleed between these two isolation valves, this bleed needs to be locked OPEN so we can now be protected by a Double Block and Bleed arrangement.  But there are very few processes designed with this type arrangement on their oil pots; however, new plants being designed and constructed today should SERIOUSLY CONSIDER adding this arrangement for the safety of their workers doing this routine task.  So lets say we have only one isolation valve to separate the oil pot from the process.  No “ifs”, “ands”, or “buts” allowed here… LOCK IT OUT!  This is the ONLY thing that is separating us from the full system charge which can be hundreds of thousands of pounds of the HHC (e.g. ammonia)! 

The next and finally engineering control is the “dead-man” valve.  This is the last valve in the sequence and the one that we will use to actual OPEN the process and to CONTROL the opening.  It, without a doubt, must be a “dead-man” valve.  If we do not have a dead-man valve I am back to making this task require a SAR and Full Evac of the machinery room!  For those that may not be familiar with a “dead-man” valve, it is simply a spring-loaded CLOSED valve, which requires a worker to HOLD OPEN the valve for the oil to drain from the oil pot.  If the worker lets go of the valve handle, the valve will automatically close via a spring inside the valve.  In no manner should this valve be manipulated in order to keep it open – sadly a common error made by a lot of workers.  By this I mean we MUST NOT use tape, string, wire, bucket, etc. to hold this valve open!!!  The valve is there as a last line of defense in the event something goes wrong and the workers have to evacuate the area immediately.  The valve is designed to close automatically in order to CLOSE the process opening we were using to drain the oil.  So in summary we have THREE INDEPENDENT LAYERS of ENGINEERING CONTROLS in place for this tasks and this is actually a VERY GOOD safety system (if all three are in place and FUNCTIONAL), but we are not yet done.

Our second layer of protection for our routine task is our… Administrative Controls.  These will be our procedures, training, and staffing.  As I said earlier, this task is WITHOUT A DOUBT a process opening.  We could argue if it is a ROUTINE or NON-ROUTINE, but it is a process opening and a procedure is REQUIRED by 1910.119(f)(4).  Now we have the option of calling this procedure an Operating Procedure, Maintenance Procedure, or Safe Work Practice; but we MUST have a written procedure for this task.  PLEASE NOTE, although not required, many facilities will have a checklist to accompany the procedure to ensure that workers do not miss a critical step.  In my world of petro-chem PSM we did the non-routine process openings via a safe work permit and the routine openings via the operating procedures.  Let me be very clear here… the requirements were the EXACT SAME regardless of where the procedure resided and regardless if it was called an operating procedure, maintenance procedure, or safe work practice.  By this I mean, just because it is called an operating procedure does not lessen the safety requirements vs. if we called this a safe work practice.  This procedure MUST include the methods of energy isolation (e.g. LOTO) and I personally had a “machine specific” LOTO procedure for EACH OIL POT in my plant(s).   The PPE required for the task, and let’s not fool ourselves into thinking that all of our prep work eliminated all the hazards;  I will talk about PPE as our LAST LINE of DEFENSE later on, but PPE is an ABSOLUTE requirement for this job and will be HIGHLY DEPENDENT on how well we can use engineering and administrative controls to protect the workers.  The procedure should also specifically call out that ALL PROCESS OPENINGS require a worker and a “stand-by” worker donned in ONE LEVEL of PPE ABOVE the worker doing the task and part of the LOTO for the task.  This is a common PSM requirement when opening a process that contains a Highly Hazardous Chemical (or as EPA called them Extremely Hazardous Substance (EHS).  The one level of PPE above is needed for assistance should the worker doing the task be overcome by an unexpected uncontrolled release of the HHC/EHS (more on this later on, as there are some specific rules to play by).  For example, in our scenario, the worker doing the task would be in a Full Face Air-Purifying respirator and therefore the stand-by person would be donned in a Self-Contained Breathing Apparatus (SCBA) that is in the pressure-demand mode (e.g. positive pressure).  The procedure needs to instruct the workers to barricade a safe distance from the process opening to ensure that personnel who are not protected by the energy isolation (e.g. LOTO) and PPE are kept well away from the potential release point.  As I said earlier, this may mean that we need to evacuate and the entire machinery room.  But in most cases, a simple barrier using barricade tape (per the site’s policy) or even traffic cones work (per the site’s policy) that can clearly demarcate the danger zone will suffice.  The “stand-by” person will be positioned OUTSIDE the danger zone and is ONLY allowed to entry this area in the event the worker doing the tasks becomes incapacitated (more on this later on – but this is a CRITICAL point and has ER implications related to both 1910.120 and .134).  I would also like to point out that the distance for this DANGER ZONE may vary from oil pot to oil pot based on the level of engineering controls (e.g. process isolation and ventilation) being utilized for the task.  Determining factors that we should apply to establishing this safe distance are:

  1. the size of the oil pot,
  2. time the process will remain open,
  3. pressure in the vessel after isolation from the process,
  4. temperature of the HHC/EHS,
  5. hazards associated with the HHC/

Since we are talking about ammonia oil pots I will make a very generally statement and suggest that we have at least a 15’ buffer from the discharge point, with the mind-set that there may be processes that would require the entire machinery room be evacuated and the stand-by person located just outside the room and in radio contact with the worker performing the task.

Other administrative controls that come into play with this task are our respiratory protection program and our certified PPE Hazard Assessment; both require WRITTEN programs/documentation and training.  Do not be surprised when an OSHA CSHO requests the “Certified PPE Hazard Assessment” for draining oil pot(s), as this specific task is one that should be assessed and the PPE requirements certified (see 1910.132(d)(2) for more on this requirement).   Those that are using a Safe Work Permit for process openings and these permits are issued by supervisors (e.g. a member of management), these can go a long way to meet this certified assessment requirement, as long as the permit covers the hazards and the PPE needed for the hazards and their degree of hazard and are approved by a supervisor or manager.  I have seen OSHA also accept these permits as PPE Hazard Assessments when they have been approved by hourly personnel who have been through a rigorous “permit writing” training that included PPE Assessment training.

Our last line of defense is… Personal Protective Equipment (PPE).  In NO WAY should we ever use PPE in place of available and feasible engineering controls and administrative controls.  Instead our PPE should be based on the engineering and administrative controls available; however, when dealing with HHC/EHS we never rely SOLELY on engineering and/or administrative controls either.  When dealing with any HHC/EHS we will often times rely ON ALL THREE LAYERS of protection, even when we have multiple controls within each layer of protection.  As in this case, we have three independent layers of engineering controls in place and yet I am going to require my personnel to still done critical PPE, such as an air purifying respirator (APR) and gloves. 

Please allow me to digress here a bit and discuss the use of APRs.  I require my personnel to wear APR’s and NOT SAR’s because I have Industrial Hygiene sampling data on “draining oil pots” that clearly demonstrates that there is no risk of overexposure even without a respirator.  But as I said earlier, each oil pot arrangement may be different so we need exposure data on all those that may be different to determine if respiratory protection is necessary or what the acceptable level of respiratory protection is needed.  Without this data, I would suggest that some facilities would have a situation where Supplied Air Respirators (SAR; i.e. SCBA) would be required based on exposure data!  I would also STRONGLY suggest that this task be done with a calibrated direct reading air monitor.  In fact, if your meter is one of the fancier ones in use today, it will connect to your computer and you can download the data obtained during the task and begin to build your exposure data library.  In the scenario I have described in this article, we are using the APR’s merely as a safety precaution, as I have data that shows this task does not over expose the workers (and I use ACGIH’s TLV of 25 ppm rather than OSHA’s PEL of 50 ppm).  The APR is being used more or less in this situation as an “emergency escape respirator”, so in the event of a failure in our energy isolation device (e.g. isolation valve) the worker will have the ability to escape from harm’s way.  Let me be very CLEAR here… if we have such a failure in our energy isolation, the worker and the standby are NOT equipped to handle this event and MUST EVACUATE to a safe distance and this would WITHOUT A DOUBT be an EMERGENCY SITUATION and anyone responding into the warm or hot zones would be an “emergency responder” and would fall under the requirements of 1910.120(q).  This is based on the simple fact that our release has gone from a “controlled release” to an “uncontrolled release” and this “uncontrolled release” needs to be assessed by a trained emergency responder to determine the level of response necessary to SAFELY manage this “uncontrolled release”.  Remember, we now have an OPEN process with the potential for the entire charge to be dumped.  I would hope that the evacuating personnel are capable of assessing the seriousness of the risk posed by this event and on their way out of the machinery room begin emergency shutdown procedures per the site’s emergency shutdown procedures.

Now let’s get back on track discussing the PPE that should be used for draining an oil pot.  Other than the full-face APR discussed above, the worker needs proper gloves rated for ammonia exposure.  These are NOT leather gloves!!!  There are several materials of construction that are acceptable for ammonia gloves and in most cases the regular length glove will be acceptable; however, my facilities always required a longer gauntlet style glove for added protect to the wrist and arms.  It is not uncommon for some facilities to require some type of torso protection for all line/process openings, but in my opinion, this is not a need in our scenario because of all the layers of protection we have in place.  I would require the regular work uniform include long plants, long sleeves, and safety shoes.  If workers are permitted to wear short-sleeve shirts I would certainly require gauntlet style gloves as we want ZERO exposure to liquid ammonia on bare skin.  I would also like to remind everyone that we are basing this PPE level on top of the engineering controls and administrative controls already in place.

With that part of our discussion out of the way, let’s look at the actual task of draining oil from the oil pot and how this task falls into OSHA Compliance demands.  What I am about to explain will hold true for BOTH PSM/RMP covered processes and those that are NOT covered by PSM/RMP, but other OSHA standards and both OSHA and EPA’s General Duty Clause.  Under our LOTO of the task, and donned in our PPE, the worker will slowly open the dead-man valve, cracking it open only slightly to determine the level of risk associated with the opening; bearing in mind that the worker is RIGHT at the point of discharge, UNLESS (red font indicates an UPDATE on 4/11/12) we comply with ANSI/IIAR 2-2008, Section 14.2 which states, in full:

Provisions shall be made for removing oil from piping and equipment where oil is likely to collect.

14.2.1 Detailed operating procedures suitable for each drain point shall be provided for oil draining operations. Safety and personal protective equipment shall be specified.

14.2.2 Oil draining shall be conducted only by trained technicians and shall not be left unattended while in process.
EXCEPTION: Permanently piped automatic return systems.

14.2.3 Oil removal shall be accomplished by one or more of the following:

a. A rigid piped oil return system.

b. A vessel equipped with an oil drain valve in series with either a self-closing or manual quickclosing emergency stop valve connected to the oil drain point, a vent line, a vent line isolation valve, and an approved pressure relief device.

c. Piping which provides capability for isolation and refrigerant removal to another portion of the system.

d. An oil drain valve in series with a selfclosing or manual quick-closing emergency stop valve. When draining to atmosphere, rigid piping routing the oil 2 to 4 ft [0.6 to 1.2 m] away from and within sight of the valves shall be provided. Use of temporarily attached rigid piping and emergency stop valves is permitted.

e. Any other suitably engineered system.

connect a sort hose to the drain line so as to move the discharge point away from the worker.  I have been told that levels of 1,000 ppm have been measured from the surface of the drained oil!  He/She will then continue to open the valve slowly, all the while assessing the situation as the open the valve more, NEVER fully opening the valve. 

PLEASE NOTE:  As I mentioned earlier, we are opening the process to do this task; however, for many facilities opening the dead man valve will NOT be the actual first opening of the process for this task.  Many facilities have a plug in the end of the drain line, which I STRONGLY RECOMMEND for every open ended line on a process containing any HHC/EHS (regardless of quantity).  The first actual OPENING should be considered when the worker removes the plug from the end of the line.  So this means that ALL of the prescribed safety measures MUST BE in place at the time the plug is removed.  I make mention of this, as some facilities may not be viewing the removal of the plug as an “opening of the process”, but in fact there have been some serious injuries and deaths from this simple activity.

It is fully expected that some ammonia vapors will be released during this task, thus we call this a “controlled release” (vs. the uncontrolled release scenario mentioned early).  I remind everyone of the need to have the exposure sampling data available to demonstrate what this exposure potential is and that our PPE MUST match up to this level of exposure!  I would also remind everyone of the suggestion to use a calibrated direct reading instrument to monitor ammonia levels during the task.  If all goes well and our layers of protection all work, we work in an ammonia atmosphere without any issues.  Workers are protected and OSHA standards are complied with; however, as I stated earlier it is when something goes amiss that takes us to the next level (e.g. emergency response).

Let’s say something does go wrong; I want to offer a couple scenarios as each one may look similar to the next but our response to each could be considerably different.  Let’s look at the worst case first: 

1) worker doing the task collapses for an unknown reason in the danger zone (e.g. barricaded area).  We have no idea as to why the worker collapsed but based on the task taking place and the situation at hand, we HAVE TO ASSUME that is it is associated with the task.  If the workers were using a calibrated direct reading instrument and the stand-by person can SAFELY determine that the atmosphere is SAFE (e.g. below 300 ppm of NH3 and at least 19.5% O2) then he/she can enter the danger area and render aid – AFTER FIRST SOUNDING the alarm for medical assistance (e.g. sounding the site alarm, calling 911, etc. as per the site emergency action plan).  Although we have a worker down, we have VERIFIED we have a SAFE ATMOSPHERE in the machinery room and therefore those responding to render medical assistance and none of the actions of the stand-by worker will be considered an “emergency response” as defined by 1910.120.  In this scenario the three driving OSHA standards will be 1910.38 (Emergency Action Plan) , 1910.151 (Medical  Services and First Aid), and possibly 1910.1030 (Bloodborne Pathogens).  But please keep in mind that if the stand-by person could NOT verify a safe atmosphere (e.g. he/she did not have a calibrated direct reading instrument) this scenario could NOT be done in this manner AND WOULD NEED TO BE AN EMERGENCY RESPONSE!

2) worker doing the task collapses in the danger zone while air monitors (both fixed sensors and the portable one the worker is wearing) are alarming.  In this scenario it is not clear what is happening, as we have high levels of ammonia (e.g. alarm on the direct reading instrument and the machinery room alarms are sounding).  This is where the “stand-by person” becomes a critical path and their one-level of PPE ABOVE the downed worker becomes more evident.  This worker will have on an SCBA that is in the pressure-demand mode (e.g. positive pressured) that is rated for IDLH atmospheres.  I would also like to take a minute and point out that we ARE BOUND by 1910.134(g)(3) Procedures for IDLH Atmospheres in this scenario!  If our respiratory protection program does NOT contain Procedures for IDLH Atmospheres, facility personnel will NOT be permitted to follow this scenario; meaning that AS SOON AS THE WORKER GOES down the stand-by person can ONLY call for outside help – THEY MUST NOT ENTER THE DANGER ZONE.  Those responding to this emergency situation (“from outside the immediate release area”) will be “emergency responders” that fall under 1910.120(q).  But let’s say our respiratory program DOES include Procedures for IDLH Atmospheres that meet or exceed the OSHA requirements in 1910.134(g)(3).  In this situation, the stand-by person notifies their supervisor, or designee identified in the site Procedures for IDLH Atmospheres, that assistance is needed.  THIS COMMUNICATION IS CRITICAL and it MUST be done BEFORE the stand-by person enters into the IDLH atmosphere.  I would also like to point out that we are NOT entering a permit-required confined space in this scenario, as those rescue requirements are CONSIDERABLY more detailed than these requirements.  The sole function of this stand-by person is to remove the downed worker to a safe location.  As you can imagine there are a ton of variations that come into play with this, but let’s for the sake of keeping this to an article rather than a book say that the rescue is occurring at ground level and the rescuer is strong enough and physically capable, using the rescue equipment available, to remove the downed worker from the IDLH atmosphere.  As they are entering the IDLH atmosphere they are actually under 1910.134(g)(3); however, if the stand-by person is unable to perform their duties and there is a need for personnel responding from outside the immediate release area needed to enter this IDLH atmosphere, their actions will be deemed an “emergency response” and must be directed by an Incident Commander, with both the IC and the responders falling under 1910.120.  (PLEASE NOTE: if a response involved a downed worker inside a permit-required confined space we will have even more standards and safety precautions that we must follow!!!!)

3) worker doing the work makes an error and causes an uncontrolled release during the tasks.  In this scenario let’s say the energy isolation somehow fails and we get an “uncontrolled release” of ammonia through the oil pot.  Keeping in mind that the PPE we have prescribed to be worn for this task is NOT designed, nor was it ever intended to be used in an area with an “uncontrolled release”.  We based our PPE on our exposure data and the fact that we have layers of engineering and administrative controls; however, we have now entered into a situation where we have NO exposure data and we know our engineering controls have failed us in some way.  In other words, we are NOT prepared for this “uncontrolled release” and we need to evacuate the area.  This is where the road splits…

If our “uncontrolled release” was just a “hiccup” and the release has stopped and our ventilation was capable of controlling the atmosphere we could return to the machinery room once we have determined the situation has returned to normal exposure levels.  BUT BE CAREFUL doing this, UNLESS WE KNOW 100% what caused the hiccup and we know it is under control; however, if we have no idea what caused the hiccup and there is a potential for it to occur again (and it can happen again if we do not know what caused it in the first place) then we will need to regroup and reassess our ALL THREE layers of protection to ensure they are adequate before returning to the machinery room, but this return would NOT be considered an emergency response as we have removed the “uncontrolled release” and the hazardous atmosphere from the room.

If our “uncontrolled release” was a continuous release AND was large enough that we had data indicating that the release was enough ammonia that it was creating a hazardous atmosphere within the machinery room (e.g. the ammonia sensors in the room are in alarm), anyone entering the room would have to do so under an Incident Command and be a trained responder falling under 1910.120(q).

In our three scenarios stated above, we have to assess the potential risks to determine if the release will be treated as an “incidental release”.  In simple terms an “incidental release” is one that is of such quantity, complexity, and degree of hazard that the worker can SAFELY clean-up the hazardous material using their basic HAZCOM, PPE and training.  This worker MUST ALSO have received training so that they can clearly distinguish between an “incidental release” and one they need help with due to their lack of needed PPE or training.  In other words, if they are equipped and trained to clean up the material using their everyday PPE and the hazards of the material and the quantity of the material is such that they do NOT need specialized PPE or specialized assistance from personnel who have more specialized training (e.g HAZMAT Technician) then this release is considered an incidental release and we handle it with our basic safety programs.

Although OSHA has stated in their Compliance Directive Generally, the release of anhydrous ammonia, for example, from a refrigeration unit would necessitate an emergency response under HAZWOPER, they do so under the premise that we MUST ALWAYS assume that a response to an “uncontrolled release” of any HHC/EHS is an “emergency” UNTIL WE HAVE OBTAINED the necessary data that tells us otherwise.  OSHA is trying to prevent workers and employers from becoming complacent with their HHC/EHS, which is easy to do for many facilities especially in the refrigeration industry where many facilities believe… “it’s just ammonia”!  Along these lines, OSHA has provided us with eight (8) basic criteria that would put our response in the “emergency mode” and require us to follow 1910,120(q).  These eight criteria are:

1. The response comes from outside the immediate release area;

2. The release requires evacuation of employees in the area;

3. The release poses, or has the potential to pose, conditions that are immediately dangerous to life and health (IDLH);

4. The release poses a serious threat of fire or explosion (exceeds or has the potential to exceed the lower explosive limit or lower flammable limit);

5. The release requires immediate attention because of imminent danger;

6. The release may cause high levels of exposure to toxic substances;

7. There is uncertainty that the employee in the work area can handle the severity of the hazard with the PPE and equipment that has been provided and the exposure limit could easily be exceeded; and

8. The situation is unclear, or data are lacking on important factors.

Let’s briefly discuss each one of these criteria, providing OSHA language along the way.

1. The response comes from outside the immediate release area.  "Emergency response" is defined as a response effort by employees from outside the immediate release area or by other designated responders (i.e., mutual-aid groups, local fire departments, etc.) to an occurrence which results, or is likely to result, in an uncontrolled release of a hazardous substance. Responses to incidental releases of hazardous substances where the substance can be absorbed, neutralized, or otherwise controlled at the time of release by employees in the immediate release area, or by maintenance personnel are not considered to be emergency responses within the scope of 1910.120(q). Responses to releases of hazardous substances where there is NO POTENTIAL safety or health hazard (i.e., fire, explosion, or chemical exposure) are NOT considered to be emergency responses. The standard covers responses "by other designated responders." The use of the "or" means that responders are a separate group, different from employees within the immediate release area, directed to respond to the emergency by the employer. Employees working in the immediate release area (not just outsiders) are covered if the employer designates them as emergency responders. The standard, 29 CFR 1910.120(q), uses the term "responders" generally to refer to employees who respond to emergencies.  The Superfund Amendments and Reauthorization Act (SARA), the statute that mandated HAZWOPER, directs broad coverage of all employees responding to emergencies with no limitation on their location. SARA states, ". . . standards shall set forth responding requirements for training of workers who are responsible for responding to hazardous emergency situations who may be exposed to toxic substances." (See SARA 126(d)(4).  For an emergency to be covered by the standard, conditions causing a dangerous situation which involve hazardous substances are sufficient; there need not be both an emergency and a response by outside responders before the employer prepares for an emergency.  For example: A release of chlorine gas above the IDLH, obscuring visibility and moving through a facility, is an EMERGENCY situation even if the initial responders are from the immediate release area. Employees who would respond to this release, whether they work in the immediate area or come from outside, would be “emergency responders” and MUST act in accordance with 1910.120(q).  Employees must not respond to releases in the immediate release area that would otherwise require outside assistance from a trained hazardous materials personnel merely because the definition of an emergency response states that an emergency response is "… a response effort by employees from outside the immediate release area."  Conversely, incidental releases of hazardous substances that are routinely cleaned up by those from outside the immediate release area need not be considered emergency responses solely because the employee responsible for cleaning it up comes from outside the immediate release area.  For example: Paint thinner is spilled in an art studio and the janitor is called from outside the immediate release area to mop it up. The janitor does not have to respond in accordance with 1910.120(q), although the janitor would be expected to understand the hazards associated with paint thinner through hazard communication training and PPE training.

2. The release requires evacuation of employees in the area.  As was discussed earlier, this means that if the workers doing the oil pot draining had to evacuate the area because their “controlled release” turned into an “uncontrolled release” (for whatever reason) and this “uncontrolled release” overwhelmed their engineering controls, administrative controls, and PPE, they would be forced to evacuate the area.  Thus anyone returning to this danger area to either attempt to stop the release or to even just “investigate” the uncontrolled release would have to do so under 1910.120(q).

3. The release poses, or has the potential to pose, conditions that are immediately dangerous to life and health (IDLH).  This one is a bit tricky because our HHC/EHS (e.g. ammonia) has a relatively low IDLH level.  But if you were able to follow me in the earlier part of this article where I talked about the exposure data, PPE Assessment, and the Procedures for IDLH Atmospheres we can handle this one.  First, we have to keep in mind that there is a HUGE DIFFERENCE between our “controlled release” of ammonia during our task and the “uncontrolled release” scenarios I have mentioned.  OSHA allows us to work in IDLH atmospheres, as long as we meet (or EXCEED) 1910.134(g)(3).  The KEY TRIGGER here is the CHANGE from a “controlled release” we have assessed and prepared for - to an “uncontrolled release” that we have not assessed and may not be equipped and/or trained to work in.  We MUST respect IDLH atmospheres and in doing so we take safety to a higher level; but we cannot fear IDLH atmospheres when we have an HHC/EHS in our workplace.  We have to prepare for the degree of hazards the material poses and be able to identify when we are NOT equipped, nor trained, to work in this type of environment.  It may be easier to understand if I say it like this… there are four categories that impact us when working around hazardous materials and hazardous atmospheres:

1) working in a controlled release environment where we know the exposures and they are BELOW IDLH,

2) working in a controlled release environment where we know the exposures and they are ABOVE IDLH,

3) working in a controlled release environment where we do NOT know the exposures,

4) working in an uncontrolled release environment where we know there is a known or potential IDLH atmosphere.

I have been in situations where I have taken my breathing zone direct read meter from my collar (reading 20 ppm of NH3) and placed it at the leaking valve stem and have it read 300 ppm (IDLH for NH3).  Now there are people that could tell me they have seen a valve stem fail and cause a major release and certainly create an IDLH atmosphere, and so they would say that working on a leaking valve stem is an “emergency response”.  I would respectfully disagree, but with this clarification… working on a leaking valve on a derailed railcar in the middle of a bean field where I have nothing but my PPE and maybe some administrative controls (e.g. incident action plan, and response plan) I would agree 100% that this is WITHOUT A DOUBT an “emergency response”.  But in our machinery room where we have our mechanical ventilation that was designed, installed and maintained in proper working order, the situation is like night and day from the railcar in the bean field scenario.  I will admit that I may (and have from time to time) have procedures in place that when working on or investigating an order of NH3 that the worker(s) were required to place the ventilation into “emergency mode”, but this was done SOLELY as a precautionary measure and NOT because of the size of the leak.  I would also have the workers be equipped with a calibrated direct reading instrument to ensure they were certain of their exposures.  If they found an IDLH atmosphere, then they had to either follow the procedures for Working in an IDLH Atmosphere or if they were not equipped or STAFFED to do so then it was:

  • EVACUATE the area and either obtain the required staffing and equipment;
  • evacuate the area and isolate the leak from outside the machinery room and allow the ventilation to clear the room, then enter into a maintenance mode that would include LOTO and PPE to make the needed repairs; and lastly
  • if we could not isolate the leak and/or the ventilation was overwhelmed and we were creating hazards that could impact people not even associated with the work, then we MUST sound the alarm for additional help which would put the “additional help” in the “emergency responders” category and 1910.120(q) would apply.

4. The release poses a serious threat of fire or explosion (exceeds or has the potential to exceed the lower explosive limit or lower flammable limit).  This one does not really apply to our workers draining the oil pot, as it is really meant to protect those workers who may find themselves in a flammable atmosphere that is not at a level to be called an IDLH atmosphere.  So when I say that this does not apply to our workers draining an oil pot, I mean that YES ammonia is explosive (LEL of 15-16%) and YES we must be concerned with this as responders, but our two workers are NOT responders and as soon as the IDLH level (e.g. 300 ppm for NH3) is achieved they evacuate.  Even if we were to work in the IDLH atmosphere using our procedures that meet or exceed 1910.134(g)(3), we would have to reach 150,000-160,000 ppm before we entered into the flammable range.  This means that we would be in Level A suits.  BY ALL MEANS, those workers who are trained to work in IDLH atmospheres MUST BE trained to recognize this and understand that their PPE is NOT rated for a flash fire and therefore they are NOT equipped to work in that flammable environment.  I could write an entire book on the hazards of working in an IDLH atmosphere, but for the sake of time I will offer this advice… in your Working in IDLH Atmosphere procedures we should put a MAXIMUM CONCENTRATION that the workers are permitted to work in.  With ammonia, I set my limits at 4% or 40,000 ppm (1/4th (or 25%) of the LEL).

5. The release requires immediate attention because of imminent danger.  This one drives most people crazy as they rightfully associate ammonia with “imminent danger”.  But lets put this into perspective… have you ever smelled ammonia?  When you did, were you in “imminent danger”?  Ammonia has one big advantage over some of the other HHC’s and EHS’s in that it has a very low odor threshold.  So we have a leak and we identify the leak by the odor in the machinery room; at this point we are not in “imminent danger”.  We also use our other senses to help us identify “imminent danger”, such as if we see an ammonia cloud, we need to move quickly away from it as the concentration will be 40,000 ppm or higher!  I am in NO WAY condoning this, but I have been in machinery rooms that have a constantly level of 3-5 ppm each day I was in the room (its one way to keep your auditors out of your machinery room J!)  I did not feel I was in “imminent danger” from this concentration (but by the end of the week I felt I was in danger from what we found in the auditJ) so the leak that was contributing to this 3-5 ppm ammonia in the room was a release that did NOT require immediate attention because of “imminent danger”.  Let me demonstrate how this may be different with another HHC/EHS.  Take chlorine for example, it has an IDLH of 10 ppm (vs. 300 ppm for NH3).  If I were to walk into a room with 3-5 ppm of chlorine in the atmosphere, this would be an “imminent danger” situation and the “uncontrolled release” would be deemed one that would require an evacuation or a response to stop it at the point of release, hence it would be an “emergency response”.

6. The release may cause high levels of exposure to toxic substances.  This criterion is very vague and I am assuming that OSHA is using it to ensure employers do not send employees into situations where they would be overexposed to hazardous substances from “uncontrolled release” that have not exceeded an IDLH level, but do pose significant hazards to the workers.

7. There is uncertainty that the employee in the work area can handle the severity of the hazard with the PPE and equipment that has been provided and the exposure limit could easily be exceeded.  This one is along the lines I have discussed throughout the article.  Worker MUST BE trained on the limitations of their PPE and how to identify when those limitations have been reached.  There are a lot of times where the worker’s focus on respiratory protection “blinds” them into thinking that respiratory hazards are the only hazards, forgetting about absorption, flammability, corrosive, etc.

8. The situation is unclear, or data are lacking on important factors. If we take just one thing away from reading this article, let it be this… An unknown atmosphere (i.e. we do not know BOTH the chemical and the concentration) MUST be treated as an IDLH atmosphere, until we have obtained enough data to assess the atmosphere to not be IDLH.  And the ONLY respirator that we can wear in an IDLH atmosphere is a Supplied Air Respirator (SAR) which can be either an air-line with escape air or an SCBA (both styles MUST be in the pressure-demand mode).

Often times workers may know the chemical involved, but not the concentration and this MUST be treated as an IDLH atmosphere.  Along with this, here is another CRITICAL PATH for using air-purifying respirators.  We MUST know four CRITICAL pieces of data BEFORE we can safely use an Air-Purifying Respirator:

1) The chemical

2) The concentration of the chemical

3) The IDLH of the chemical and that the atmosphere is NOT OVER the IDLH

4) The atmosphere has AT LEAST 19.5% Oxygen

If the user does NOT know ALL four pieces of this CRITICAL DATA, they MUST treat the atmosphere as IDLH and use an SCBA in the pressure-demand mode (e.g. positive pressure) until they have ALL of this data.

So to wrap this up, we have learned the three hazard control methods: engineering controls, administrative controls, and personal protective equipment and how they can be layered to raise the level of safety for workers doing hazardous tasks.  We have learned that anytime we open a process with a HHC/EHS that our risks go up substantially and therefore we need to establish a hierarchy of controls to meet this level of risk.  We also laid out the difference between working around a “controlled release” vs. an “uncontrolled release” and how our level of safety differs, as well as how an uncontrolled release should be managed as an emergency response until we have enough data to tell us all is safe.  But we also learned how a leak of a HHC/EHS does not necessarily have to be treated as an emergency response, as some will fall into the category of an “incidental release”.  But most of all, I hope you have come to learn that no two situations will ever be the same and it is CRITICAL that workers who work with and around an HHC/EHS  be trained to understand the limitations of their training and their PPE so that they do not place themselves in harm’s way.  It is ALWAYS better to err on the side of safety, than to rush in and be injured by the release, making yourself a victim and a hindrance to the response.

MANY THANKS to my friends and collegues for their time and expertise in reviewing this article.  Over 30 Safety/IH professionals in private industry and OSHA/EPA personnel involved with PSM/RMP (including NH3 refrigeration) offered insights and challenges to the ideas covered in this article. 

View 's profile on LinkedIn

  LinkedIn Group Button



seminar logo white

I am teaching another 1-Day Course for ASSE

Intro to Process Safety Management.

February 13, 2017 

The Rio, Las Vegas, NV


 kmcllaw logo

an unpaid endorsement



an unpaid endorsement


kemkey logo

an unpaid endorsement

safteng man copy






Safteng.net website reputation