GAO: CSB Not Meeting Its Statutory Mandates

mailto
June 22nd, 2009| by Dr. Saraf |

According to a recent report (Aug. 2008) from the Government Accountability Office (GAO), the Chemical Safety Board (CSB) is not in compliance with its statutory mandates.  The CSB is an independent federal agency charged with investigating industrial chemical accidents and has been in operation since 1998. GAO report states that CSB is not investigating all chemical releases that have a fatality, serious injury, substantial property damage, or the potential for a fatality, serious injury, or substantial property damage.

The following table summarizes CSB performance in 2006-2007.

  2006 2007
Total accidents recorded in CSB database 822 920
Accidents with Fatalities 38 35
Accidents CSB Investigated 6 5
Accidents with Fatalities CSB Investigated 5 1

CSB is limited by budget and resources as to how many incidents it can investigate. GAO has recommended that CSB request the necessary resources from Congress to meet CSB’s statutory mandate or seek an amendment to its statutory mandate. Alternatively, GAO makes the following recommendations to help CSB cover more incidents.

  • consider using the work of government agencies, companies, and contractors, to a greater extent to maximize the board’s limited resources;
  • improve the quality of its accident-screening database by better controlling data entry and periodically sampling accident data to evaluate their consistency and completeness;
  • publish a regulation requiring facilities to report all chemical accidents, as required by law, to better inform the agency of important details about accidents that it may not receive from current sources;

To download a pdf copy of the GAO report on CSB click on the link below:
Chemical Safety Board: Improvements in Management and Oversight Are Needed.

Is your Organization Making the Right Decisions?

mailto
June 14th, 2009| by Dr. Saraf |

We make hundreds of decisions everyday in our lives – for example should I eat fries with my burger?

Daily decision making

(Photo Courtesy of Amber on Flickr)

Eating a few fries this one time will not increase risks of a heart attack. Plus I like the fries from this joint. I’m ready to order!!

That was a relatively easy decision; however, there are certain decisions that are more complex and may have significant implications to your organization’s future including organizational safety.

For example, if your organization is deciding on future investments or acquisitions. One would think that the basis for such complex decisions would be financial metrics – return on investment (ROI), life-cycle analysis, Net present value (NPV). Clearly these economic figures incorporate cost-benefit analysis. However, in a real world situation one cannot mitigate all the risks in an investment and therefore some of the most complex decisions are influenced by various other factors – internal politics, stakeholder interests, lack of complete knowledge. How can we bring rationale to such decisions?

Group Decisions

Group decisions in particular are prone to be influenced by emotions rather than logic. It is therefore necessary that everyone in the group understands the basis of decision and agrees to conclusions being reached. Group decision situations can be very effectively dealt with using Multi-attribute decision analysis (MADA).

Let me give you a simple example. Say you are in the market looking for a car, first you have to decide upon criteria for buying decision. You come up with three main attributes you would want in your new car – price, mileage and safety. But which of these three is more important to you? Would you be willing to sacrifice safety for price? So you want to prioritize attributes. Then you go from one dealer to other and compare different models. Come decision time what you are doing is mentally creating a scorecards for cars you have looked at and figure out the winner based on price, mileage and safety.

The same thought process can be duplicated for group decisions. What attributes to select? How to weigh the attributes? All these decisions are taken by the group and thus the end result represents “Wisdom of the Crowds”.

Of course to do this methodically I recommend using Analytical Hierarchical Processing (AHP) to select attributes, decide their weights and then final outcome. If you need further information on this topic please contact me.

When Should you Consider Group Decision Methods?

I recommend one should look into group decisions when considering investment of $5-million or more when you are trying to answer one of the following questions:

  • Which projects to fund and to what extent?
  • How to adjust operations and maintenance resources?
  • How to adjust funding levels as resource constraints change?
  • What acquisitions best fit with organizations mission/goals?
  • How to create an investment risk portfolio based on available knowledge?

Group decisions using MADA are an effective way of creating transparent decisions by efficiently capturing ideas of all stakeholders and providing a common decision platform.

Developing Trust While Communicating Risks

mailto
June 7th, 2009| by Dr. Saraf |

Communicating risks to public from chemical or nuclear facilities is a major challenge. We notice this over and over again with EPA community-right-to-know, nuclear power plants and lately with the LNG facilities.

I want to focus on a critical aspect of this risk communication – winning the trust of the community.

There are two major cognitive social psychology principles one must understand while planning a risk communication program:

1. If an individual has strong initial impressions of a hazard, then subsequent evidence will be influenced by these beliefs. New evidence will appear reliable if it is consistent with initial impressions. And similarly, new evidence will be dismissed if it is contrary to initial impression.

For example, if you are talking about risks from LNG terminals and the neighbors are concerned about explosion hazards or terror threats, you cannot simply dismiss them. I have noticed in some of these meetings citizens raising question about hazards and some PR firm answers LNG is a clean energy source. Within a minute this evidence will be dismissed!You have to acknowledging the hazards/risks consistent with initial impressions and then present your strategy for minimizing risks.

Remember people receive information from media, talking to other people, news, blogs, radio and all these come into account in risk perception. Without complete knowledge of the system, such as LNG terminal, decisions of a person are based on beliefs concerning likelihood of uncertain events. And we know these likelihoods are minimal – now all you have to do is show what steps are being taken in order to reduce the likelihood.

2. If an individual lacks strong prior opinions about a hazard, their risk perception will be influence by HOW the information is presented.

As an example, if a patient is offered a treatment option with 70% chance of survival (as opposed to 30% chance of death) he is more than likely to accept the treatment. Framing the statistics in terms of living rather than dying changes risks perceived by the patient.

So what does all this mean in context of developing a risk communication or crisis communication plan:

  1. Realize you will have two distinct categories of people – people who have developed prior opinions and people who have no prior idea
  2. Plan your presentation for both these categories.
  3. For audience without a prior belief system educate them with past statistics.
  4. Audience who have a prior opinion you have to acknowledge the hazards and focus on risk mitigation measures

You can keep the above guidelines and now you have to decide

  • What information to communicate
  • How best to put across the information
  • Who should communicate

Maintaining Process Safety During Tough Economy

mailto
May 31st, 2009| by Dr. Saraf |

We appear to be in a global recession – companies all over are struggling to deal with impact of economic slowdown. In an attempt to maintain profits, companies are drastically reducing their spending and also trimming down their workforce. It is anticipated that one-fourth of the employers in the US will undergo workforce reduction in 2009. As a result of pessimism and uncertainty about future, there is an increased level of anxiety amongst the employees. There are a few issues that a chemical manufacturing firm must address to ensure safe operation during the downturn:

  1. Morale of the workers:  The glooming uncertainty about jobs and increased level of anxiety leads to a disinterest amongst employees. It makes workers less productive and starts affecting their day-to-day decisions including the ones regarding safety.
  2. Procedural risk mitigation: There are many potentially dangerous scenarios where procedures are put in place as risk mitigation measures. In a state of anxiety and worry, a worker is more likely to commit an resulting in serious consequences.
  3. Loss of knowledgeable personnel: Trimming of workforce may result in loss of valuable tacit knowledge in the form of experienced employees.
  4. Scheduled maintenance: As a cost cutting measure, an operating company may decide to delay planned maintenance. John Bresland of CSB talks about this issues in the video below:

Recommendations for Managers

  • Revisit your high risk scenarios that require procedural risk mitigation measures. It may be worthwhile to investigate automation instead of human intervention.
  • Figure out ways to cheer employees and keep them in an upbeat mood. If you have any suggestions on boosting employee moral please write them in the comments section below this post.
  • Here is a free e-book (Author: Wendy Mack) - How to Prevent Panic and Promote Productivity in Turbulent Times.

Transforming Anxiety to Energy

Safety Forecast for the 21st Century

mailto
May 24th, 2009| by Dr. Saraf |

The 20th century was a time of great technological change that forever transformed how we live and work – changes that necessitated the birth and development of the field of Process Safety Management. The early years saw the evolution of mechanization into assembly lines and true industrialization.  Lack of access to South American nitrate during World War I, led to the creation of the synthetic chemical industry.  World War II fostered increased industrial growth and sophistication.  By the 1960s, we were building computers and beginning our race to the moon.  Industries grew becoming increasingly sophisticated and reliant on automated systems.  The 1970s brought the creation of the US EPA and OSHA.  The 1980s witnessed one of the greatest tragedies in the last century – an estimated 4,000 people died in the 1984 Bhopal accident. Since then, the process safety community has evolved in its approaches and methodologies to better manage risks.

But what have we, as process safety professionals, learned from the experiences of the 20th century? How can we use that learning to make the process industry safer in the 21st Century?

20th Century Accident Data
Based on data from Lee’s Loss Prevention Handbook, 551 incidents occurred during the 81 years from 1911-1995.

accidents

  • Out of these 551 incidents, 270 (49%) resulted in a zero fatalities
  • 374 (68%) of the accidents resulted in less than four fatalities
  • 18 incidents (3% of 551) resulted in 100 or more fatalities

A more elegant way to analyze the accident data is to construct a fN curve, where f is the cumulative frequency of incidents leading to N or more fatalities.

In this fN curve,

  • The first point represents the 281 incidents that resulted in one or more fatalities.
  • The last point is the Bhopal accident, which is estimated to have resulted in 4,000 fatalities

accidents

Based on the 20th century fN curve, we observe

  • In the 1911-1995 time frame there were 100 incidents that resulted in 10 or more fatalities – an average of 1.2 incidents per year in which 10 or more people were killed.
  • There were 18 incidents (3%) in which more than 100 people died – an average of 1 incident every six years.
  • There were 3 incidents (0.54%) in which more than 1,000 died – an average of 1 incident every 28 years.

Prognosis for the 21st Century

How will the fN curve look for the 21st century? Without a crystal ball, a simplistic prediction is that the 21st century curve will be identical to the 20th century. So should we expect 551 process incidents again by 2099?

As technology progresses, clearly it will impact the fN curve.Based on my experience, I’m going to propose a Saraf fN curve predicting fatalities in the process industry for the 21st century.

death vs accidents

Here are my predictions regarding the safety performance of the process industry in our  current century:

  • There will be a drop in incidents that result in 1-10 fatalities because of increased hard hat safety requirements and awareness among workers due to better access to hazard communication and training.
  • In the last century, 18 incidents resulted in 100 or more fatalities. This number will decrease in the 21st century as risk mitigation measures for high consequence incidents become more reliable.
  • I believe the frequency of incidents that result in 10-20 fatalities will remain comparable. Such incidents are typically a result of human error, siting issues, proximity to community, poor plant design, and maintenance/inspection programs. Based on my understanding, I do not envision elimination of these key safety issues within a matter of a few decades. Zero incident plants are not a reality.

Final Word

Every decade in the 21st century, the process industry will witness an incident that will result in at least 10 fatalities.

OSHA Working on Dust Explosion Regulation

mailto
May 17th, 2009| by Dr. Saraf |

OSHA is currently working on regulation for preventing dust explosions.

The first reported dust explosion was in 1890 in a sugar mill in Chicago. According to the CSB investigation, dust explosion has resulted in 130 deaths and 780 injuries since 1980. According to Combustible Dust Institute research of news reports, there were 200+ dust fires and explosions in the United States in 2008 alone. Following the 2006 recommendations from the Chemical Safety Board (CSB) to regulate “dust”, OSHA started investigating if the available dust explosion standards are adequate. After the 2008 sugar-dust explosion that killed 14 workers, OSHA has decided to regulate dust explosion.

A dispersed cloud of dust in a confined space when ignited can lead to an explosion. A major problem is workers are not aware of the hazard or they do not always know the dusts that can explode. Because dust explosions are manifestation of dust properties and confined space where the dust is released.  Dust comes as fine particles, chips, flakes made up of metals, wood, grains, foams, plastics, rubber, pharmaceutical drugs, coal, flour, or sugar.

There are existing NFPA standards aimed at dust explosions but the compliance is not mandated. It will interesting to see how OSHA approaches dust regulations.

External Damage: The Number 1 Cause of Natural Gas Pipeline Incidents

mailto
May 10th, 2009| by Dr. Saraf |

I recently came across a report from European Gas Pipeline Incident Data Group (EGIG) titled “Safety Performance Determines The Acceptability of Cross Country Gas Transmission Systems”. The paper presents incident data contributed by six European gas transmission operators over a 30-year period of 1970-2001.

An incident within this failure database implies unintentional release of gas and a release is classified in one of the three categories:

  1. Pinhole crack - diameter of leak is 2-cm or less
  2. Holes -leak diameter greater than 2-cm but less than pipe diameter
  3. Ruptures - diameter of leak is pipeline diameter

According to the report, the main causes for pipeline releases are:

  1. External interference: Activities such as digging, pipling, ground works, excavation, ploughing, slabbing, casing and sleeves may damage the pipeline.
  2. Corrosion: It can be external or internal
  3.  Material failure: Due to construction of material, defect specification
  4. Ground movement: dike break, erosion, flood, landslide, mining

Relative percentage of incidents based on incident cause is presented in table below - 50% of the pipeline releases appear to be caused by external interference. Furthermore, when a pipeline is damaged by external interference there is a 75% chance that it will lead to a hole or a guillotine rupture.

Initiating Events Leading to Pipeline Incidents

Initiating Event %
External interference 50
Construction defect 17
Corossion 15
Ground movement 7
Hot-tap error 5
Others 6

Of the corrosion incidents reported, 79% are caused by external corrosion. Furthermore, 74% of the external corrosion incidents were due to pitting.

Fires and Explosions in Biodiesel and Ethanol Facilities

mailto
May 3rd, 2009| by Dr. Saraf |

In an earlier post, I talked about safety in biofuel plants. Today, I would like to share couple of Google Maps (Source: John Astad of Combustible Dust Policy Institute) that graphically display incidents in biodiesel and ethanol plants.

Biodiesel Fires and Explosions Map (click to view)

According to the map, there were  8 fires and 6 explosions in biodiesel facilities in the U.S in 3-year period (2006-2009). This translates to approx. 5 incidents per year.

Ethanol Fires and Explosions Map (click to view)

Based on the map, there were 22 fires and 11 explosions in biodiesel facilities in the U.S. in 10-year period (1998-2009). This translates to approx. 3 incidents per year.

LNG Regasification Facilities: Consider High Pressure Releases

mailto
April 26th, 2009| by Dr. Saraf |

LNG import terminals in the U.S. follow NFPA-59A and 49CFR193 standards for facility siting.  As a part of the siting studies, LNG regasification facilities report thermal radiation exclusion zones and flammable vapor exclusion zones:

  • Flammable vapor hazard zones are based on a design spill from a single accidental source, usually a guillotine rupture.
  • Thermal radiation hazards are based on tank fire.

These events used for determining exclusion zones, pipe rupture and tank fire, have a low probability.  Surprisingly, the current standards have no requirements to model consequences from high-pressure LNG releases which are more likely.  Furthermore, the current codes do not reference Vapor Cloud Explosions (VCE) which directly affects plant layout and siting.

There is high-pressure processing equipment in LNG import and export terminals (e.g. vaporizers in LNG import terminals) that can result in aerosolized release of LNG and can potentially lead to an explosion.  Therefore, in addition to the consequence modeling performed for code compliance, facilities must model high pressure releases and VCEs for:

  • Optimizing plant layout
  • Deciding risk mitigation measures, and
  • Terminal/building siting

I discuss issues related to high-pressure releases in a recently presented paper at AIChE conference in Florida.
Sanjeev Saraf, Subodh Medhekar, and En-Hua Yang, “Have You Considered High-Pressure LNG Releases? They Are More Likely Than Catastrophic Failures”.

AIChE members can view the recorded presentation at the following URL:
http://aiche.confex.com/aiche/s09/webprogram/Paper145131.html

If you need further information, please feel free to contact me.

OSHA NEP to Focus on the Chemical Industry in 2009

mailto
April 19th, 2009| by Dr. Saraf |

In 2007, after investigating the explosion and fire at BP’s Texas City facility, the US Chemical Safety and Hazard Investigation Board (CSB) recommended OSHA to launch a PSM National Emphasis Program (NEP). The objective of the NEP was to reduce the hazards associated with petroleum refineries. As a part of the NEP initiated in June 2007, OSHA inspected 81 refineries.

So far OSHA has issued more than 500 citations as a result of the NEP audits with approximately 2.7 -million dollars in fines. During the NEP inspections, the top PSM elements identified for violation appear to be:

  • Operating procedures
  • Process safety information (PSI), and
  • Process Hazard Analysis (PHAs)

In 2009 OSHA plans to initiate a similar effort focuses on the chemical industry.