Why this Post Now?
Process Safety is in position to undergo a significant, permanent, change. Big data, incidents like Macondo, and new technology are the enablers for this change. ACM Facility Safety is focused on developing a Predictability Model using big data analytics, and learnings from previous industry incidents (Macondo etc.), 1000's of PHA data sets, and audits.
Every day my confidence grows that this change in the world of Process Safety will happen soon. It is so encouraging, that one can start turning visions like “ be seen as safe enough with evidence (more than PSM plans, see the impact, feel the impact) by doing the right things" or “ be as good (safe) as you can be, these are the truths in Process Safety” by taking the valuable information inputted by teams of people. These people dedicated their knowledge and enthusiasm to tell the truth in PHA sessions about hazards and safeguards. Analytics can distill these truths down to simple metrics, critical elements and benchmarks. We have extracted these “truths” from PHA’s using ACM Analytics.
Undone Recommendations is a barrier, preventing people from being excited about what PHA's can really provide. There are 1000's of undone recommendations.
Can Analytics help here? It's a good place to start by using analytics to prioritize 1000's of undone Recommendations. It removes the black mark against PHA's for those people managing the Recommendation budgets. They know this situation well; “too many PHA's & then too many Recommendations, how do I handle that?” (An interview quote from an executive of a multinational oil company, while preparing for my first book).
This post suggests that the industry is struggling to prioritize undone Recommendations being accumulated every day by 10,000's, and that applied Analytics is the way to address this gap. The following is a brief background on Analytics and the application to prioritize undone Recommendations.
18 months ago we began the Risk Alive project and started challenging the “checkbox” normalcy of PHAs, HAZOPs, LOPAs, SIL studies. We learned and now understand that the data in the “checkbox” exercises really represents truths of hazardous scenarios and processes from employees, operations, and designers. It is now understood that when the “checkbox data” is shared, integrated and analysed, the resulting clarity can be powerful, insightful, efficient, and safer.
3 months ago I was inspired by a YouTube video given by the CEO of Technip at a GE conference in Italy. It was simple, inspiring, and truthful. Everything seemed to be relatable to Process Safety. The message was around the new reality inside the economic downturn, that to reduce cost required a culture-behavioural step starting with the front lines of the company. The Technip CEO demonstrated with the Technip success on the “total recordable injury number”, reduced by 50%, on hard hat safety. It took a culture adjustment starting with the front lines to make Technip successful. The front lines rejected normalcy, and challenged everything.
I was inspired thinking about this and thought that it seemed relatable to the challenges in Process Safety. Doing the right thing must mean more than following the “norm & perceived regulations”. If the front lines could challenge the “normalcy “ of PSM and Process Safety, would it change the way we analyze, understand, and learn from PHAs? Is there more purpose to PHAs than applying checks on check boxes and developing recommendations?
7 days ago I was reminded by the Final Macondo report why we need a change in Process Safety. The final report was issued 6 years after the incident. The lessons learned were similar to past incidents about human error, lack of understanding about which safeguards are critical, which causes/threats are significant, which recommendations and past incidents needed to be understood and considered. The perception reiterated in the report was that meeting existing regulations is not good enough, and setting the bar with HSE metrics on people safety, and inferring therefore Process Safety, is good as well, is far from the truth. The effort by BP during Macondo was heavily weighted to HSE and hard hat safety lagging indicators. It was the evidence of “safe enough”. It was okay for a while until the big guy, Mr. Process Safety, landed a knockout punch. A knockout punch that impacted all four corners of the ring, People, Environment, Assets and Reputation. (Item#27 in Macondo Investigation Executive Summary).
Every day 100’s of risk assessments are conducted. Everyday 1,000’s of recommendations are developed from those risk assessments. Every day new recommendations join the ranks of undone recommendations from past risk assessments (PHA's). Every day the total industry risk continues to climb from the continuous accumulating undone recommendations.
The accumulated risk climbs like interest on accumulated debt. The debt payments increase as the number of incidents increase and the severity of the incidents increase. Undone recommendations accumulating will surface again in the form of more frequent near misses and more incidents. Industry will pay the price. Jobs will be lost or moved to other countries. The message can be the same in another incident report 6 years from today, unless the front line workers make a culture shift. Those that think unsafe days don’t happen at their facility will be surprised to learn that, yes, incidents happen where the risks are mostly unknown - like interest payments on debt. Making the interest payments get you by, but the bank calling a debt due and payable would be painful and surprising, something like unsafe days and Process Safety.
Undone risk assessments and undone recommendations are just like inactive safeguards. The analytics of a safeguard are usually formed by making observations and asking questions. Here are a few questions below which relate to safeguards and recommendations. These questions are formulated from putting one’s self in different situations over time. Embedding the answers in analytics can protect us from not being aware, not understanding everything, not being the designer, or not being the engineer. These questions may exist before a PHA is conducted or years after a PHA is conducted.
What about a safeguard or recommendation? What’s to know anyway?
- Is it working? Is it done?
- What’s the ROI? How many risk reduction points per dollar invested?
- What’s the priority of this safeguard/recommendation and ranking compared to all other safeguards and recommendations?
- What is the cause/threat contribution by the safeguard or recommendation compared to others?
- What is the change in likelihood of an unsafe day at this facility?
- What is the risk reduction contribution impact to each of the four corners of the ring?
- What is the incident contribution by known incidents missing this safeguard or recommendation?
- What is the criticality ranking inside this facility of this safeguard or recommendation?
- What is the criticality ranking (profile) of this safeguard compared to the criticality ranking (profile) of other facilities?
- How are the brother/sister safeguards impacted by the safeguard or recommendation?
- What is the Profile of this safeguard compared to the profile in other facilities? Similar equipment? Similar processes? Similar chemical hazards?
You might be questioning if this makes sense, suggesting that analytics can help answer these questions. It might seem difficult at first but the idea of using analytics to solve issues or giving us a better perspective and understanding has been around for a long time. When Risk Alive was forming, a TED talk on facial recognition inspired us and led us to believe that this dream was attainable. The TED talk described an experiment that aimed at selecting the best TV commercial based on people’s facial reactions. Based on these reactions, the experimenters could identify the emotions and the most positive reactions of the viewers. All the results were analyzed using analytics. If the developers of facial recognition software could pre-determine the feelings, attitude and behaviour using a camera and then analytics, why not analytics on risk data sets for Process Safety. Like emotions, all possible outcomes are known and predictive. The same chemicals and equipment, with the same hazards, are being used over and over again. Why couldn't analytics help save Process Safety? Analyzing undone Recommendations may be a good place to start.
Follow this link to an attachment which explains a bit more. Apologize for the marketing slant on the attachment. If you can get past that, then there are some new things to think about.
“Be the best you can be” was a message my grandson shared with me when I asked him what do you know that’s really important? You have only lived 8 years compared to my 60. I believe analytics for Process Safety and helping to resolve challenges like too many recommendations qualify as really important.