Part 2: Hydrogen, Agendas and Anxiety
Series 1, previously published on LinkedIn, January 10, 2023 (with minor edits)
In this episode I discuss the misunderstandings about risk, how it is assessed, and the unnecessary fear of system complexity in relation to safety.
The Department of Business, Energy & Industrial Strategy (BEIS) had contracted engineering consortium ARUP+ to conduct a study into the safe use of hydrogen for heating in domestic situations. The document recording their findings, was Safety Assessment Conclusions Report incorporating Quantitative Risk Assessment (QRA) and it became highly contentious, primarily at the instigation of The Hydrogen Science Coalition, the online publication Hydrogen Insight and other antagonists. This eleven-part series was published January 10, 2023 on LinkedIn in response.
This was part of a much larger scope of work sponsored by BEIS and fell under the Hy4Heat programme within ‘Workpack 7’. I discuss the entire programme in a series of articles to be published later.
Original article with minor edits follows.
Hydrogen and Risk in the Home
The QRA report on the safety of domestic hydrogen (produced by ARUP+) commissioned by the UK government agency BEIS, is conservative, arguably to a fault. This has become a target for those who are far less measured and are willing to exploit the abundance of caution. In an article published last year (Whitby Hydrogen Village and the Weaponisation of Fear), I discussed the organised opposition to domestic hydrogen pilot schemes, and those initiatives are linked to this study in two ways:
They are both part of the same programme with the common objective of transitioning from natural gas to a safe hydrogen network.
They are both subject to a campaign that exploits public concerns with the common objective of mobilising a popular rejection of hydrogen.
Under various guises the controversy continues with anti-hydrogen antagonists running a crusade against any positive mention. One post was a response to the sort of ‘push back’ I was providing. There was a doubling down on the idea that hydrogen is ‘not intrinsically safe’. I did respond but for reasons perhaps best known to the original poster, my comments would not pitch up, but the sense of the reply I had intended to give was that there is no such thing as ‘intrinsically safe gas’.
The Intrinsic Safety Misnomer
Intrinsic safety relates to the source of ignition not the fuel. Even calling a gas ‘inherently safe’ is something of a misnomer, because even if it is non-flammable, non-toxic and unpressurised there are still safety risks. Any gas in high proportions can cause asphyxiation and even oxygen at concentrations approaching 100% is toxic. How is this relevant? Simply that the safety of all gas in any application, has to be engineered-in, because it is never ‘inherent’.
When we think of safety we think of risk. In the sense that hydrogen heating is yet to be piloted for domestic use in the UK, it is an unprecedented risk, which is the justification for an ultra-conservative approach. If natural gas were to be assessed for domestic use in the home today, there would certainly be more mitigations than were envisaged in the late 1960s, when it started to replace coal gas in the UK.
It took decades for flame supervision technology to be regulated, and even now, Flame Failure Devices (FFDs) are still not mandatory for appliances is an owner occupied homes. I return to this point under the heading, ‘The Natural Gas Baseline’, three paragraphs below. I discuss this in Part 11: What Would Kletz do.
Mitigations and Outcomes
If two scenarios are to be compared, the one with the highest overall risk is not necessarily the one with the largest number of hazards, because it matters how severe and likely each of them are. The same can be said for the number of mitigations because they are linked to the hazards. Mitigations make undesirable events less likely, so to associate the identification of mitigations with reduced safety, is to have it a little backwards.
You see, there is a mistaken view, oft repeated, that residual risk increases with the number of mitigations because there are more ‘single point failures’. Incorrect. More mitigations stop them from being single point failures, and for it to be any other way, additional mitigations would by default, make a system less safe, which is ludicrous. It is not a case of ‘more things to go wrong’ but ‘more things to go right’. There is nothing controversial or fanciful about this; this is the least you need to know to make a serious contribution to the conversation. It is also tied into the complexity myth that I will discuss after the next section.
The Natural Gas Baseline
The ARUP+ QRA is a comparative analysis, where the baseline for safety is that of natural gas. The risk associated with the use of natural gas in the home, is broadly understood and accepted by the public, albeit at an intuitive level.
It so happens, that the additional mitigations that ARUP+ identify for hydrogen would also be an improvement upon current practice with natural gas. FFDs (Flame Failure Devices) are not obligatory in owner occupied homes, yet around 40% of explosions and fires occurred as a result of them not being fitted to old appliances. FFDs are assumed for all hydrogen appliances.
However, the focus of the ARUP+ work is the risk of gas escapes between the Emergency Control Valve (ECV) and the appliances, with the following leak categories:
Large leaks are defined as having a breach area equivalent to a >7mm diameter hole typically as a result of physical damage (accidental or deliberate) and they can create large gas accumulations in a building. The Excess Flow Valves (EFVs) would close under such circumstances thereby greatly reducing the probability of a flammable atmosphere arising.
Medium leaks are defined as having the equivalent open area of a hole between 3 and 7mm diameter and are therefore most likely to result from physical damage. These can produce flammable pockets of gas in poorly ventilated spaces, therefore the mitigation for this is adequate ventilation.
Small leaks are defined as having a breach area no greater than 2mm equivalent diameter and according to ARUP+, are not capable of creating sufficient gas accumulation to ‘produce injuries’. The report notes that this class of leak is detectable by smell and represents 97% of all reported leaks.
To be consistent, those who oppose domestic hydrogen on safety grounds, should at least be interested in how EFVs might improve the safety of current practice with domestic natural gas applications. Of course, it seems likely that many who have such anxiety about hydrogen, previously had little interest or opinion about gas safety.
For example, some of the objections I saw in online comments were aimed at the specification of ventilation, something that is largely incorporated in existing building regulations for other reasons. It is at least questionable whether these people are really sure about what they are opposing.
Complexity Myths
It is a common misconception that complexity results in more risk, but when you consider the most robust systems, the required level of safety arises from deep complexity. The environment that an aircraft operates in is far more challenging than that of a vehicle on the ground, but despite this it is generally accepted that statistically, flight is the safest form of travel. Yet by applying the logic of my interlocutors, the fact an aircraft needs the additional measure of wings, would make it less safe than a car.
In complex systems, Reliability, Availability, Maintainability and Safety (RAMS) - which I discuss in Part 8: The RAMS of Hydrogen Excess Flow Valves, are properties that emerge from contingency in design, which add up to overall resilience. If you have any kind of system interlock, that engineered layer provides additional safety because of the increased complexity, not in spite of it. Recognising this may not be intuitive, but here are some things to consider, which might help make this more accessible. Hopefully at least one of them will be familiar enough to you to be a useful example.
Dynamic Positioning (DP): if we compare the number of system nodes and level of complexity between a DP1 and DP3 class vessel, would the far simpler DP1, be safer?
Safety Instrumented Function/Level (SIF/SIL): can SIF and SIL be achieved at any level by making a priority of reducing the number of components and interlocks?
Safety Instrumented System (SIS): if sensors are to be polled, would the voting logic 1-o-o-1, be better than 3-o-o-5 for the preservation of critical detection?
Layers Of Protection Analysis (LOPA): would eliminating protective layers improve safety?
Mechanical Isolation: would a single isolation barrier be safer than a double block and bleed?
Electrical Isolation: would a single locked out breaker safer than two locked out breakers in series?
Microelectronic Reliability: as electronic components become more complex do they become less reliable?
Without requiring you to turn your screen upside down, I can tell you that the answer to all the above is, 'no'.
In March 2023, I will publish my own commentary on the ARUP+ QRA, serialised over several weeks.
This entire series is published in parallel and the following parts tackle a series of topics related to domestic hydrogen safety. The vehicle for this is the deconstruction of an article that was critical of the ARUP+ QRA. On that journey I identify errors, misinformation and misrepresentation of some basic facts. References are provided in support of the analysis.