Prankul Middha analyses the benefits of using methods such as computational fluid dynamics to assess risk to life, property and the environment
The popularity of natural gas as an energy source is expected to grow substantially because it can help achieve two important goals: providing reliable energy supplies and reducing adverse impacts on global climate and the environment. Natural gas is the most environmentally friendly fossil fuel due to the lowest CO2 emissions per unit of energy, around 30% less than petroleum and 45% less than coal.
There are huge reserves of natural gas, but a large component is located in remote regions far from where it can be used. For delivery to end users, the gas is typically converted into liquefied natural gas (LNG), a process called ‘liquefaction’, by cooling it to temperatures around –160°C, and reducing the volume to approximately 1/600th of its original volume. This reduction allows it to be shipped aboard specially designed LNG vessels. After arriving at its destination, LNG is often warmed to return it to its gaseous state – known as regasification – and delivered to customers through pipelines or land transport.
Many of the largest natural gas fields are located beneath the ocean floor and far from land. Thus, the gas needs to be liquefied offshore. For this reason, several floating LNG liquefaction facilities (FLNGs) are currently being developed.
Where pumping gas to shore can be prohibitively expensive, FLNGs make the development of those fields economically viable. Additionally, since all processing is conducted at the gas field, there is no need for any onshore infrastructure, which significantly reduces the project’s environmental footprint.
LNG is rapidly gaining an important role as direct use fuel – that is, without need for regasification – with operational cost and pollution reduction benefits in road, rail, air and especially marine transport. LNG bunkering facilities are being expanded at ports globally, especially in highly regulated areas such as the European Union.
Natural gas is combustible and a gas leak can lead to fire and/or gas explosion upon ignition. To ensure safe and reliable operation, particular measures are taken in the design, construction and operation of LNG facilities. LNG cannot burn or explode in the liquid state; it must first vapourise, then mix with air in the proper proportions – the flammable range is roughly 5-15% – and then be ignited. This can happen in case of an LNG leak, from loss of containment, for example. Therefore, hazard analysis of LNG facilities is a very important step to ensure the safety.
Due to the cold temperatures associated with LNG, it is also important to consider cryogenic hazards such as fracturing of hull of a floating LNG facility. Historically, the LNG industry has a very good safety track record, but it is not immune from accidents. For example, a January 2004 explosion at the LNG liquefaction facility in Skikda, Algeria, resulted in 27 deaths and losses worth almost US$1 billion.
The design of any LNG facility, whether floating or onshore, requires careful assessment of the risks associated with it to protect life, property and the environment. Frequently, and especially at concept or early design stage, analytical models are used to assess the hazards and risks. These models offer several advantages: ease of set-up; speed of computation; and in some cases, coupling with risk calculation packages. However, the quick setup and calculation come at the price of accuracy.
These models are often based on semi-empirical correlations and therefore should be used with caution when the modelled conditions deviate from the range of validation. For example, these correlations do not take into account the presence of obstacles – such as piping, equipment, structures – along the path of the release or of the dispersing vapour cloud, as well as the flame following any subsequent ignition.
On the other hand, detailed models based on computational fluid dynamics (CFD) can evaluate the details associated with gas cloud development – including environmental conditions, wind speed and direction, leak location, leak rate and direction, and the presence of obstacles.
A two-level approach can be considered for carrying out these studies. The initial approach is to carry out a ‘realistic worst-case’ consequence study.
If the design explosion load is of interest, a number of dispersion calculations generate ‘realistic worst-case’ gas clouds, which are subsequently ignited to obtain a pressure load. This may be applicable for onshore facility siting where the analysis needs to confirm that the potential hazard does not extend beyond the boundaries of the facility.
If potential consequences of the initial approach are not acceptable, which would generally be the case for congested FLNG facilities, it is necessary to perform a more comprehensive study, including ventilation, dispersion and explosion calculations, to reduce unnecessary conservatism of the chosen worst-case scenarios.
As risk is a combination of the probability of an event and its consequences, these are combined with weather statistics, leak frequencies and ignition probabilities, to evaluate the expected explosion risk. This enables a facility to set tolerance criteria regarding the probability of specific events occurring.
Consequence modelling based on probability is also used to evaluate alternative layout possibilities and possible risk reducing measures at an installation. It a powerful tool that can not only evaluate the risk level, but can also give invaluable information required for evaluation of further risk reduction.
Risk analysis can have a significant influence on the design and cost of LNG facilities. Therefore, the use of simple modelling tools – with their faster computing times but more onerous simplifying assumptions – may not be as cost-effective compared with the more complex but more accurate CFD models. These include FLACS, a specialised CFD tool developed to address process safety applications such as gas dispersion and vapour cloud explosions in complex geometries. Fortunately, the two models are not completely exclusive of each other and it is fully possible to use engineering judgement to develop a ‘hybrid’ approach where the simple modelling tools are used for screening of a large number of scenarios and the CFD models can then be used to simulate the scenarios that are deemed to be influenced by the 3D environment.
Prankul Middha is general manager of safety and risk management specialist GexCon.