(44c) Optimal Supply Chain Network Design for the Upstream Sector of the Oil and Gas Industry

This work proposes a novel approach for the optimization of generalized supply chain network design problems (G-SCND) applied to the upstream sector of the oil and gas industry. A supply chain network is typically represented by complex graphs including flows of goods, materials and information linking the different nodes making part of it. These nodes stand for suppliers, warehouses, distribution centers, cross-docking facilities and/or demand points. Nowadays, market globalization along with the shortening of product life-cycles and the need for high standards of responsiveness challenge modern supply chains to be agile and flexible enough to face a changing environment.^[1] The optimal design of supply chain networks has become a strategic field in recent years due the implication of the logistic expenses in the overall cost of organizations. With some differences according to the sector, logistic costs average 7% to 9% of the sales of a company, reaching up to 30% for certain chemistry industries. In global terms, the IMF roughly estimates that the logistic costs are about 12% of the global GDP.^[2] As a result, the optimization of the supply chain network design may have an enormous impact on the operational costs. The search for efficient, flexible and robust network designs gives rise to a very challenging and interesting problem. Researchers have addressed this topic with different tools, including heuristics, optimization and simulation models, though several issues still remain open.^[3]

The design of a SCN is a strategic and long-term problem to be addressed by one or a group of organizations (extended supply chain network concept). It involves decisions on the location and type of facilities to be built, which products to storage, the inventory policies to adopt and how the goods will be transported to reach the clients. The aim is to fulfill customer demands while minimizing the net present value of capital expenditures and operational costs. Furthermore, the network responsiveness should be planned accounting for the service level required by the customers. In the particular case of the O&G upstream sector, the development and exploitation of oil and gas reserves in a low-price environment force the operators to efficiently provide a large variety and volume of materials towards the surface facilities in order to carry out tasks related to location building, energy supplying, infrastructure set-up, drillings, completions, workovers and well maintenance. Moreover, the growth of unconventional assets has deepened the need for efficient supply chains due the large quantities of materials (including proppants and water) required, when compared to conventional operations.

One of the main weaknesses identified in the literature is the absence of generalized models addressing current issues related to facility sizing and location, and supply chain network design problems. The vast majority of the works have focused on Fixed-Supply Chain Network Designs (F-SCND) approaches, which pre-determine the number of stages or “echelons” in the SCN together with the type of facility to be installed in each echelon (typically warehouses, distribution centers and factories). In fact, roughly 80% of the works propose networks with only one or two echelons, and an important number of them account for the distribution of a single product.^[4] Little relevance and attention have been paid by researchers to the inner logistic flows, and the possibility of direct supplies from not-end echelons to final clients. F-SCND approaches are based on rigid models that usually lead to non-optimal solutions.^[4] To face these weaknesses, Generalized Supply Chain Network Design (G-SCND) models have been recently proposed to represent more flexible networks, involving the location of various types of facilities in several nodes, multiproduct flows and demand fulfillment from any node in the network, among others features. The potential of G-SCND approaches relies on the ability to overcome operational issues and trade-offs along the SCN structure, capturing the intertwined nature of decisions and leading to more efficient results.^[5]

This work presents a novel G-SCND approach applied to the O&G upstream sector. It takes into account all the distinctive characteristics of the O&G industry in order to obtain a comprehensive conceptual model. The clients in this model correspond to the oil and gas production areas, whose wells demand different types of materials and services for their development, operation and maintenance. The proposed model differentiates between capital investments and operational flows. The first one comprises all the needs (materials and services) required to develop a new well, and directly depends on the companies’ projections over their assets. Operational flows, on the other hand, reflect the need for maintenance of the active well population, whose size increases through time as new wells are completed and brought to production. Additionally, the proposed model does not limit the number of echelons or layers. Instead, through a novel formulation, the optimal number of echelons is determined by the model, depending on the type of product to be supplied. Economies of scale functions are used to model capital expenditures in new infrastructure and operational costs. This typical non-linear relationship between the size of a facility and the corresponding cost is such that every additional unit size added to a facility is more economic than the previous one. A similar behavior is observed for the unit handling and transportation costs. To capture this, hierarchical types of facilities (with different capacities) are proposed in the model, each one being able to be located in a set of potential nodes. The main difference between them is given in terms of the minimum and maximum flows admitted. Fixed costs, unitary handling costs and transportation cost are also dependent on the type of facility involved. Finally, the concept of opportunity cost is introduced in order to represent the responsiveness of the SCN through the measurement of the time required to serve the wells. In the oil and gas industry, the equipment used for drilling, fracturing and workover tasks is, in most cases, rented, representing a major fixed cost for these companies. Maximizing the utilization of this equipment is crucial and mainly depends on the reduction of the travelling times and the rapid assistance from closer nodes. Also, when a problem occurs in a well, the production is interrupted representing a significant loss for the company. Thus, minimizing the time to repair wells is important in order to reactivate the production as soon as possible.

A mixed integer linear programming (MILP) formulation is developed to mathematically represent the conceptual model here proposed. The objective is to establish the most convenient location, hierarchical structure and size of facilities that minimize the net present value (NPV) of all the capital investments and operational costs over the planning horizon, including: (1) transportation costs, (2) acquisition costs, (3) financial stock costs, (4) handling costs, and (5) opportunity costs. The model makes primary, inner and secondary logistic decisions related to materials acquisition from suppliers, the inner transportation between facilities, and deliveries to final customers for demand fulfillment. A set of real-world case studies with different demand patterns, geographical distributions and product families is addressed to assess the potentials of the proposed approach. Very efficient SCN designs are obtained with different types of facilities installed at several nodes. The division between investment and operational requirements allows to further perform a sensitivity analysis with different projections for the development of new O&G reserves, which permits to assess the robustness of the SCN in today’s volatile price environment.

In short, a novel G-SCND approach is proposed for the upstream sector of the oil and gas industry. The main concepts introduced are the elimination of pre-determined echelons’ structures and the capture of economies of scale through the modeling of a hierarchical set of facilities that are able to be installed in any potential location. The formulation comprises all the components commonly found in the operation of modern O&G upstream supply chains, yielding an effective and comprehensive SCN representation.

1-Shah, N., Process industry supply chains: Advances and challenges. Computers and Chemical Engineering, 29 (2005) 1225–1235

2- International Monetary Fund, World Economic and Financial Survey (January 2017).

3- Tsiakis P., Design of Multi-echelon Supply Chain Networks under Demand Uncertainty. Ind. Eng. Chem. Res. 40 (2001), 3585-3604.

4- Melo M.T. et al, Facility location and supply chain management – A review European Journal of Operational Research 196 (2009) 401–412.

5-Kalaitzidou M.A., Optimal Design of Multiechelon Supply Chain Networks with Generalized Production and Warehousing Nodes, Industrial & Engineering Chemical Research, 53, (2014), 13125−13138.