The amount of gas flared has not substantially declined over the past ten years [since the first edition of this guidance was produced in 2011], a reflection of the challenges that exist. These include lack of infrastructure, distance to markets, availability of skilled labour, capital investment constraints, ownership arrangements, and the absence of an efficient and effective regulatory framework or a functional authority to enforce regulations.
Routine flaring of associated gas represents the most significant source of GHG emissions from flaring at upstream production operations. Controlling this source of flaring is challenging and often requires major capital investments in new equipment and/or infrastructure to manage, process or export the gas, as well as viable markets to monetise the gas (see Figure 1, below).
Non-routine flaring is typically intermittent and of short duration, and can be either planned or unplanned. Good governance practice would include a documented justification for any non-routine flaring event, to enable analyses of root causes and identify mitigation options for such flaring. However, regardless of the cause or classification, the ultimate goal is to pursue gas capture solutions and operating regimes that eliminate the need for any type of flaring.
Non-routine flaring occurs at the well head as a result of drilling, completion and flowback activities. Flaring occurs at field locations, including gas gathering or processing facilities, due to planned maintenance and unplanned upsets or malfunctions. When components exceed design parameters such as allowable pressures, emergency relief devices automatically route the gas to a flare to maintain safety, usually for a period of minutes or hours. A compressor engine failure on a low-pressure gathering pipeline system can create unanticipated back-pressure in the system; this in turn can cause relief valves at upstream production facilities to send gas to flare automatically.
Equipment failures and upsets at downstream gas plants can create non-routine flaring at distant upstream production locations because take-away capacity is reduced or completely shut in.
The identification of potential flare source locations can be determined through a review of as-built piping and instrumentation drawings or on-site inspections. All process piping discharges into the flare header should be followed back to determine their origin. For example, process drains discharging into the flare knock-out vessel could also be a potential for flaring in the event of a failure of one or more low liquid level alarms in the process vessels; drain valve leakage; or operator error with respect to manually operated liquid drain valves.
More detailed methods of tracing flaring to its sources include analysing alarms and other process anomalies (e.g. low/high flow, high pressure, etc.) from data historians, for distributed control systems or central control room systems. Maintenance management systems can be queried for work orders associated with flaring events to identify failed instrumentation or process equipment.
Sampling and analysis of flare gas composition may be able to help identify the sources of gas. Process hazard reviews or revalidations can identify whether as-built systems match the original design layouts.
In reviewing processes and facilities, the design aspects that most affect the volume and duration of non-routine gas discharges to flare include piping design, equipment sizing, equipment choice/specification and instrumentation/control. Addressing these aspects can improve the frequency, duration and volume of non-routine flaring.
Root cause analysis (RCA)
A ‘bad actor’ programme is designed to identify the operational factors and/or equipment that lead to recurrent upsets that result in non-routine flaring. It is based on the Pareto principle, i.e. 80% of the issues come from 20% of the causes.
The strategy to identify ‘bad actors’ involves an RCA of production-related upsets based on the impact of each flaring event. RCA is the application of approaches, tools and techniques to uncover the causes of such events. The primary aim is to identify the factors that resulted in the flaring event, the nature of the occurrence, its magnitude (e.g. volume flared and lost production), the location, and the timing of the consequences. This information will allow a determination of the behaviours, actions, inactions or conditions that need to be changed to prevent the recurrence of similar outcomes. The results of the RCA are typically documented to provide a record of the incident investigation, corrective actions to be implemented and details of lessons learned.
With respect to non-routine flaring, the proximal reason why a particular flare event occurred may be obvious, for example a safety valve sending high pressure gas to a flare. However, the determination of the root cause of an event usually involves an investigation or engineering analysis of the circumstances leading up to the actual flare event. For example, the cause of the overpressure could be due to a disruption in the normal flows of material that causes a blockage or high-pressure zone in some part of the system.
The investigation or engineering analysis of an event that results in flaring should consider the overall context, which is to enhance profitability while striving to maximise safety and on-stream production time through more reliable equipment and process systems. Consequently, an RCA should yield useful information when it is focused on equipment or systems that experience one or more unscheduled shutdowns or failures during a specified time period that lead to a flaring event.
Targets for in-depth investigation can be identified through queries of maintenance management systems that look at classes of equipment for repair costs and the number of outages by equipment type. Flaring waiver requests can also evaluated. Graphical depictions (Pareto charts) can also be developed showing the frequency of flaring, the cost of repairs and the extent of lost production. From this analysis the important ‘bad actors’ can be identified, and a more detailed action strategy can be developed to address them.
The RCA should look for primary and contributing causal factors by examining:
- the inherent design of the equipment or system
- a defect in the material of construction or a system component
- faulty equipment build or system logic
- errors in the way that the equipment or system was installed
- inappropriate equipment or system for the operating conditions
- inadequate maintenance procedures
- improper operation.
Common elements in most programs aimed at identifying ‘bad actors’ consist of the following:
- Defining the scope of the analysis, either to consider the whole operation (entire field or gas plant) or limited portions (certain well sites or production facilities) over a selected time period. A broader scope requires more work, but is likely to identify unanticipated trends.
- Obtaining flare event logs and work order histories from the computerized maintenance management system for all in-scope assets
- Performing data cleansing/validation to eliminate duplicate or false records, completing partial data records and obtaining maintenance and associated downtime/lost-production costs
- Preparing Pareto charts in which operations or equipment that cause larger flaring volumes and lost-production costs rank higher than others
- Selecting the top 20% from the dataset for further review: these are the ‘bad actors’.
Rotating equipment and spares strategy - vital to thecause
Turbines and gas compressors are the most important examples of rotating equipment employed in the management of associated (flare) gas. The operational reliability of compressors can determine the success of an associated gas capture and utilisation project.
Significant considerations involved in the choice of compression equipment include cost (that is, capital, installation, operating and maintenance costs), operational flexibility, reliability and emissions.
The unavailability of compression due to maintenance or an unexpected upset can cause significant loss in revenue in the gas utilisation project. The installation of spare or standby units is an important consideration, despite the additional capital and installation costs. Although upsets or emergencies cannot be predicted, the scheduling of maintenance shutdowns can, and planned outages should be performed when lower capacities are required.
Spare units can be arranged such that each compressor station has one spare, but this can be costly where there are several processing plants in multiple fields. This prompts other considerations, including how to use the standby compressor. For example, a decision to operate with a dedicated spare, rather than operate both pieces of equipment at partial loads (inefficient), or to alternate units running at full load (frequent high-stress start-ups), ensures that a serviceable unit is immediately available when the other fails. It also means that it is unlikely that both units will reach the end of their lives at the same time.
Another approach is the standardisation of compressor makes, types and models across the entire operation of a company, such that one swappable standby spare is kept for several compressor stations. When the standby spare is used to replace a defective compressor at a particular station, the replaced compressor is repaired and then becomes the new standby spare for use if another compressor failure occurs at the same station or at another plant. This approach has the potential to significantly drive down the cost of standby sparing for compressors.
A lifecycle cost comparison of strategies typically requires a statistical simulation based on historical data, which includes:
- failure consequence and occurrence rate to inform a criticality assessment
- failure mode and effects analysis to identify dominant failure modes and possible risk mitigation tasks
- the frequency and characteristics of each failure mode
- maintenance data
- information on spares, including purchase price, storage cost, lead time, depreciation, and categorisation of logistical availability.
In determining whether to adopt a standby or a shared-spare philosophy, consideration should combine lifecycle costs with risk tolerance. For a standby philosophy, the initial number of compressors will be twice the number required for design rates, but the redundancy may be warranted where uptime is the driving factor. For a shared spare, the initial number of compressors will be lower, equal to the number required for design rates plus a certain number of extra units. When the risk tolerance for outages is higher, the number of spares will be lower. The final number will depend on balancing the desire to minimise gas flaring with the cost of providing (and maintaining) shared spare compression capacity.
This flaring management guidance, developed by IPIECA, IOGP and GGFR in partnership, outlines new developments in flaring management and reduction, and examines industry experiences with eliminating flaring, new technologies, business models, operational improvements and regulatory policy. It is available via www.is.gd/unabaz