Evolution of petroleum system modelling: Some insights from 30 years in industry and academia

Abstract

The Norwegian Petroleum Directorate (NPD) recently released the Resource Report Exploration for 2018, of course focusing on Norway. With the exploration acreage extending from the Central Graben in the south to the Barents Sea in the north, the report covers data on generally mature exploration acreage (with the exception of the Barents Sea, which contains the bulk of remaining underexplored acreage). Nevertheless, in the NPDs assessment of exploration success they identified lack of petroleum charge as one of the main reason for dry wells, especially in mature areas (main reason for up to 61% of dry wells). In an earlier study by the NPD published in the year 2000, lack of petroleum charge was responsible for only 29% of the dry wells drilled. For the petroleum system analyst this is bad news, as a general assumption that our capacity to predict petroleum charge increases with maturity of the play, i.e. the number of wells drilled, appears to be false. When we are additionally confronted with the fact that the theoretical understanding of petroleum generation, expulsion and migration as well as the numerical modelling of these processes has reached a very high level of sophistication we should ask ourselves what is going wrong. To address this issue we should first have a look at how Petroleum System Modelling (PSM) evolved to its present state. The history of PSM, i.e. the modeling of petroleum generation and expulsion in the context of a basins geologic history, started as early as 1971 with a paper published by B. Tissot and R. Pelet from the Institut Français du Pétrol (IFP). There they discussed how a chemical kinetic model of petroleum generation could be linked to a basin burial history to predict and map the onset and extent of petroleum generation. Research in PSM in the following years was dominated by the IFP and, starting in the late 70’s and early 80’s, by D.H. Welte and co-workers at the Research Institute Jülich (KFA) in Germany. By the end of the 1980’s the first commercial service companies developing software were established, the best known of which are the IFP/Beicip Temis, the IES (now SLB) Petromod and the Platte River BasinMod software packages. Since then additional software solutions have appeared (Trinity, MPath, Migris, BMT), which were spun off either from research centers (SINTEF for Migris) or from the petroleum or service industry (Trinity/Zetaware from ARCO, MPATH/Permedia from an industry consortium, Tectonor for BMT). What is common to many commercial software packages is that they started in academic research institutes and maintained a close connection to the respective research entities throughout their history. The strong links between PSM companies and academic research were essential for the rapid development of the software, but can be seen to have also resulted in the implementation of too many elements of mainly academic interest, which make some of the software packages cumbersome or overloaded with options that many users cannot relate to. Nevertheless, the tight connection between software developers and academia was extremely beneficial to PSM in general. Examples include the development and application of kinetic models of petroleum generation-expulsion-alteration, implementation of diagenetic reactions, or fluid flow modelling techniques. From an industry perspective, PSM was rapidly adopted as a method to de-risk petroleum charge in frontier as well as mature exploration areas, especially by the larger energy companies. The interest in PSM resulted in significant Industry funding of academic and service-company defined research projects, which was pivotal in the development of the software solutions. Industry provided important feedback to both academia and developers regarding what was needed in the application of PSM. Examples of this type of collaboration are found in the late 80’s and early to mid 90’s, a time when the UK and Norwegian industry stepped up drilling in the high pressure and temperature areas of the Central Graben and Haltenbanken. Discovery of a gas condensate in Kristin field triggered the interest in the stability of oil at high temperatures, kicking off a series of research projects on the subject. In academia PSM was used to test kinetic hypotheses, mainly using 2D modelling. This branch of research continued in the attempt to integrate primary and secondary cracking models to predict GOR evolution as a function of maturity. Here the first two-component and later multi component kinetic models were defined, in close collaboration with the software developers. By the year 2000 the importance of understanding and including fluid PVT properties in petroleum system analysis became obvious, and research efforts were set in motion to better predict fluid properties using PSM. Research results on the controls and effects of petroleum biodegradation in the early 2000’s were also rapidly integrated into PSMs, although this task proved to be more difficult. At the same time significant research efforts were underway to better understand the unconventional petroleum systems. In this case, however, North American academic and industry research groups were in the forefront, with Europeans entering the scene with a few years delay. Implementation of e.g. organic porosity models was relatively rapid, and PSM entered the unconventional plays. Since the start in the late 70’s petroleum system modelling has evolved from the pure modelling of the burial history of a well in 1D, to full 3D modelling of petroleum generation, migration accumulation and alteration. PSM is now applied to both conventional and unconventional petroleum systems. It has become an established element of petroleum exploration, which also resulted in the recruitment of PSM expertise in the petroleum industry and has also evidently increased exploration success (i.e. the NPD study of 2000). The way PSM was used in the industry has also changed systematically. The move from 1 and 2D to 3D modelling occurred in parallel to the increase in availability of 3D seismics, which today in mature areas is essentially the standard. Large merged 3D datasets now allow the construction of basin scale 3D models, and most companies have regional models, which are updated continuously as new well or seismic data becomes available. The main use of PSM still focusses on reconstruction of the petroleum generation and expulsion history, the definition of drainage areas and expelled fluid volumetrics. Highly detailed multi-component kinetic models of petroleum generation, expulsion, migration and degradation are the exception, as these are so complex to set up and calibrate that the effort is rarely warranted. The focus in Lundin is on stochastic modelling to describe the range of uncertainty, especially for fluid volumetrics. PSM has today reached the stage of a mature technology, accepted in the industry to the stage that small and nimble PSM developers have largely been taken over by global service companies. The downside of this evolution is that PSM now competes internally in service companies with other technologies and is evaluated based on cost-revenue balances. We have seen that this is a tough place to be for niche technologies. Additionally, in the last decade, the fruitful and dynamic collaboration between academia and software developers as well as industry has seen a dramatic change. The combined effects of frequent drastic drops in crude oil prices (1998, 2002, 2008 and most recently 2014-2016) and increased governmental and public awareness of climate change have resulted in the reduction of funds for fossil energy research from both industry and national funding sources. This is especially the case in Europe, where for example in Germany governmental funding for fossil energy research is officially frowned upon. EU funding for fossil energy research has also essentially disappeared. This has led to a significant reduction of fossil fuel research with several well-established groups shifting focus from fossil fuels to environmental or climate research. While the interest in and application of PSM is still strong in the industry, the access to academic resources has declined. With academia stepping back, the development of PSM solutions now has moved more strongly into the hands of large service companies and the industry. We can expect to see effects of this change already today or in the future, where the decrease in supply of highly trained PSM experts from academia may affect the use of PSM in the industry. Coming back to the question raised in the first paragraph, why is it apparently so difficult to correctly predict petroleum charge on the Norwegian Continental Shelf (NCS)? There is, of course, no single answer to this question, but several aspects of the use or abuse of petroleum system modelling require discussion. The NCS has become a mature petroleum province, and the player picture has changed correspondingly. Where 20 years ago the area was dominated by the majors, since around 2007 a large number of small and medium sized companies have entered the shelf, attracted by changes in the Norwegian licensing and reimbursement systems. Here it is important to note that despite its wide recognition and application, PSM still is viewed as a minor contributor to exploration success, especially by small exploration companies, which lack the employee numbers to warrant hiring a dedicated PSM specialist (or simply think they can do without, the “source is not a problem” attitude). Such companies tend to rely on PSM consultancy, often provided by the software vendors. Here PSM is commonly performed by specialists in the use of a given software package, but commonly lacking the in-depth knowledge of the geology, geochemistry, fluid properties, etc. of the specific area under study. If we assume that petroleum system analysis is used by all players on the NCS then we need to look at how this job is done. Clearly, we can generally not complain about a lack of input data quality and resolution nor about model sophistication. However, it is clear that current basin modelling software have evolved to become extremely complex pieces of software. Maybe too complex. The correct use of the software implies that the user has to have a fundamental understanding of all the processes he/she wants to model. In an extreme case this can include not only the geology of sedimentary systems, but also tectonics, heat transfer, pressure evolution, pore-scale fluid transport processes, inorganic and organic geochemistry as well as reservoir engineering (e.g. PVT properties). As this mixture of expertise is, due to the reasons discussed above, now rarely taught at any university, the users have to either learn on the job or trust in the software, which is not always a good idea. Alternatively, we should think about the possibility that we, as PSM specialists, are over-selling PSM results. The quality of model results in the form of figures and 3D visualizations is so high that they can mask a poorly defined model, especially for the non-specialist. Impressive 3D displays of migration routes and predicted accumulations rarely represent the geologic complexity of secondary migration, but are excellent tools for boosting a prospect. Recently artificial intelligence and machine learning have been proposed as solutions to the complexities of PSM (also for seismic interpretation, or data analysis in general). My personal opinion is that this would make PSM, which is by many considered a black box, even blacker. We should additionally not forget that PSM is only a part of petroleum system analysis. The complete integration of PSM with all available geochemical, geologic and geophysical data as well as the continuous interaction with the entire exploration team is what makes petroleum system analysis a vital part of successful petroleum exploration. With the “changing of the guard” currently underway in the petroleum system analysis arena, it is clear that the industry needs well- and broadly trained PSM specialists, working with and integrating between disciplines to do the job correctly. In view of the decimation of active fossil fuel research groups and university teaching on PSM, it appears that industry and service companies have to either take this task upon themselves or actively foster the re-establishment of dedicated PSM research and teaching in academia.

AAPG Datapages/Search and Discovery Article #90349 © 2019 AAPG Hedberg Conference, The Evolution of Petroleum Systems Analysis: Changing of the Guard from Late Mature Experts to Peak Generating Staff, Houston, Texas, March 4-6, 2019