Initiation, start/stop and termination

The flexibility described so far is confined to the respective phases. In addition, two decisions that provide flexibility are available in all phases. First, the operator is allowed to abandon the project at all stages (termination flexibility). Abandonment implies a cost, which is dependent upon the phase the project has reached and the number of production wells that must be plugged.

Second, the operator may always choose to just wait (i.e., temporarily stop the development) and not pick a certain action (initiation flexibility and start/stop flexibility). This can be favourable if the oil price fluctuates heavily, since by waiting the operator might get a higher price. There is no direct cost associated with the decision to wait, but if the platform has been constructed the fixed operating cost accrues.

Based on the preceding discussion of the options available to the operator as the project develops, the decision space can now be summarised by figure 2. As pointed out the decision making freedom of the operator depends on the phase of the project, and except for the start/stop flexibility and the termination flexibility the other flexibility types are phase dependent.

Production

During the production phase the flexibility arises from the operator’s option to decide the level of
production, the drilling of production wells and, if possible, any increase in production capacity of the production unit. Both the drilling of production wells and any increase of the platform’s production capacity are examples of capacity flexibility.

Reservoirs are by nature three-dimensional, and should be modelled by three-dimensional models. However, the computational demand of such models (cf. Lia et al (1995)) inevitably hampers their usefulness in a comprehensive framework. Thence, a much simpler approach is proposed, in which the maximum production level is described by a zero-dimensional tank model with perfect communication throughout the reservoir.

The well rate, and thence the production capacity of a well (see appendix A), follows a Markov
process, where transition takes place if the field is producing. The capacity of the wells for the next production period is revealed at the end of the present production period. Hence, the production capacity is assumed known when the production decision is taken. Without any production the production capacity of a well remains the same.

By modelling the reservoir volume and the well rate as stochastic variables the (originally
deterministic (cf. Wallace et al (1985b)) tank model provides a production profile with stochastic
escalation, plateau level and duration, and decline. This is considered adequate to describe the
uncertainty surrounding the reservoir at an early stage of the development. Note that the assumption of a stochastic well rate is not in accordance with the foundation for the tank model, which requires a homogenous and well behaved reservoir. The tank model should therefore be considered a convenient framework for development of a simple relationship between important reservoir parameters and the production profile, rather than a strict condition.

To curtail the model size the production decision is implemented as a binary choice, where the
platform either produces at maximum level or it does not produce at all. This is similar to enforcing a so-called “bang-bang” solution (cf. Dixit and Pindyck (1994)).

Production wells can be drilled at all stages in the production phase, and the possible number of wells is independent of platform concept. As for the exploration wells, the production wells are assumed drilled in clusters of predetermined size, and in a predetermined sequence. The total number of clusters that can be drilled at the field is restricted by an upper limit. The production wells have infinite lifetimes, i.e., each well can produce throughout the entire production period.

It is further assumed that the platform capacity can be increased at all stages. Naturally this requires that the platform concept is designed for optional capacity, and that available space has not been utilised by previous installations of additional capacity. The increase is only limited by the available space. Hence, it is possible to use all expansion area at one stage. Capacity expansions are made in steps of predetermined size. The cost of increasing the capacity is dependent upon the magnitude of the expansion and (generally) the concept.

During the production phase, the platform incurs fixed operating costs. These are dependent on the installed capacity and the platform concept, but independent of the production level. In addition there is a variable cost associated with the production of oil. This cost is given per produced barrel.

Engineering and construction

The engineering and construction phase does not contain any decisions, but carries out the decisions made in the conceptual study phase. Any pre-drilling of production wells decided in the conceptual study phase is done during this phase.

It is assumed that the time spent on engineering and construction is independent of the selected concept (cf. Wallace et al (1985a)), and any differences between various concepts are therefore limited to the construction and operating cost, the installed capacity and the flexibility to increase the capacity. Construction cost for the production unit (the semi-submersible) is related to the installed capacity and the possibility to expand the capacity at subsequent stages. The cost is positively related with both. That is, the higher the installed capacity is, the higher is the construction cost, and the more
capacity flexibility the platform offers, the more costly is the concept.

Conceptual study

A concept is here defined by the installed production capacity of the platform and the option to
increase this capacity during the production phase. The production capacity of the platform should in this context be conceived of as a combined measure of the platform’s production, processing and storing facilities. Hence, the production capacity specifies the total well stream that can be handled by the platform. In the conceptual study phase the operator thus decides the initial production capacity of the concept, as well as the possibility of increasing the capacity at later stages (capacity flexibility). A concept where the production capacity can be increased at subsequent stages, i.e., during the production phase, requires both additional space, e.g., on the platform deck, and extra carrying capacity. This induces additional costs, which can be considered costs of obtaining flexibility. Where convenient the production capacity will be referred to as simply the capacity of the platform.

The conceptual study phase also includes the decision of pre-drilling of production wells. This to
make sure that the production of oil can start immediately after the production unit is located at the field. Without pre-drilled wells the operator might experience a delay due to drilling of production wells after the production unit has been constructed. (It is not possible to convert exploration wells to production wells in the model.) It is further assumed that the first production well reveals perfect information about the well rate for the first production period.

Exploration

The typical cost of an exploration well is today USD 15 million, and in the last decade the exploration has amounted to 10 to 30 percent of the accrued investment costs. Including exploration in the model should therefore improve the analysis significantly.

The basis for the reservoir assessment is the operator’s a priori probability distributions of the(technically recoverable) reservoir volume and the well rate. These distributions are typically obtained through seismic surveys (the volume) and wildcat wells. However, it is possible for the operator to obtain more information about the reservoir volume before the development starts (i.e., the operator has what is here termed information flexibility). In the model this is achieved by drilling of additional exploration wells.

Drilling of additional exploration wells is restricted to periods before the conceptual choice is made. The wells are drilled in clusters of predetermined size, where a cluster may comprise one or several wells. Information received from the well(s) is assumed binary, and either indicates a low volume or a high volume. (Perfect information about the reservoir volume is only obtained through production.) The indication of a low volume corresponds to the well(s) not hitting oil, while indication of a high volume is obtained if the well(s) hits oil. The probability of receiving a given well information is conditional upon the true reservoir volume, and the information is used to update the operator’s probability distribution in a standard Bayesian manner.

The degree of uncertainty in developing an oil field

The development of an offshore oil field is a task characterised by its versatility and high degree of uncertainty. The selection of the development strategy is made early in the project’s lifetime, and at the time the decision is made the information concerning the field is often scarce. For instance is neither the future production nor sales prices known with certainty. The problem facing the decision maker is therefore a problem with imperfect information. This makes the decision making process a challenging one, and the methods applied should offer adequate support for evaluation under uncertainty. The criticality of good decision making at this stage is further stressed by the fact that the choice of a development strategy is of great consequence for the profitability of the project.

Several decades may pass between inception and completion of the project, and throughout this time many disciplines are involved. A model that aims to cover the complete development must therefore be based on rather crude approximations in order to get a solvable model. The requirement for a compact representation of the problem is of course not unique to an oil field development project, but the inherent complexity makes this demand a critical one. To facilitate the modelling the project is thus depicted by a simple phasing (figure 1). The four phases cover the complete development project from the time before the PDO (Plan for Development and Operation) is submitted to the government to the abandonment.

The operator bases the PDO on information about the reservoir properties retrieved by seismic
surveys and exploration well drilling. It is however still possible to obtain further information by
additional well drilling, as captured by the exploration phase. Having completed the appraisal of the reservoir the conceptual study is carried out. The choice of concept involves, in addition to the selection of a production unit, a choice of flexibility. As any possibility to alter the configuration of the production unit is restricted by its free space and carrying capacity, the concept design is essential to the development strategy. Following the choice of concept is the engineering and construction of the production unit. This may either be the construction of a new unit, or the modification of an existing unit. Finally the depletion of the field is captured in the production phase.

The main types of flexibility present in oil field development projects

The main types of flexibility present in oil field development projects.

- Initiation

- Termination

- Start/stop

- Information

- Capacity flexibility

Of particular interest is the capacity flexibility, i.e., the option to change the scale of the project.
This may be achieved by changing the production capacity of the production unit, and/or by
changing the production capacity of the reservoir. Both the degree of decision making freedom and the number of stochastic variables are increased, implying a much more computationally demanding model.

The value of flexibility is closely related to the uncertainty. An analysis of flexibility must therefore be made together with an assessment of the uncertainty surrounding the project. Hence, the selection of stochastic variables should support the flexibility of interest. Based on previous studies and discussions with Norwegian operators three variables are considered stochastic in the model;

- oil price

- reservoir volume

- well rate

These variables are of major importance to the production profile and the cash flow of the field.

The requirement for focus improved development strategies

The focus on improved development strategies has strengthened over the last 10-15 years, and can partially be traced back to the decrease in field size that has been experienced. While the size (5 year moving average) of discovered reserves in the beginning of the 80’s was between 80 and 100 million Sm3 oe for the North Sea, the corresponding figures since 1990 has been below 20 million Sm3 oe. The reduced size has made new fields become economically more marginal, and has put emphasis on the need for so-called flexible development strategies.

Introduction

The average size of discovered petroleum reserves on the Norwegian continental shelf has declined steadily over the last years. As a consequence, the fields have become economically more marginal, and new and flexible development strategies are required. This paper describes a stochastic dynamic programming model for project evaluation under uncertainty, where emphasis is put on flexibility and its value. Both market risk and reservoir uncertainty are handled by the model, as well as different flexibility types. The complexity of the problem is a challenge and calls for simple descriptions of the main variables in order to obtain a manageable model size. Real options demonstrate the significant value of flexibility that illustrate the shortcoming of today’s common evaluation methods. Particularly capacity flexibility should not be neglected in future development projects where uncertainty surrounding the reservoir properties is substantial.

Oil Field Development

Real options analysis can be used in developing oil fields and the option to change the scale of a project when technical uncertainties exist in the field development.

In order to apply real options analysis to oil field development, the company must look at:

- Project assumptions

- Project expectations

- Key uncertainties