The Strategic Advantage of MemComputing in Oil and Gas UPSTREAM

The oil and gas industry consists of complex supply chains that can be separated into three sectors; upstream, midstream, and downstream. In this blog, we focus on the value MemComputing delivers across several applications in the upstream sector.

Upstream

The upstream sector is focused on activities related to the exploration and production of crude oil and natural gas. Companies operating in this space perform extensive geological surveys, seismic analysis, and testing to identify areas where oil reserves are likely to be found – both onshore and offshore – and then use drilling rigs to bring them to the surface.

Exploration and production in oil and gas is extremely capital intensive due to the expensive equipment, highly skilled labor, and required resources. In order to maximize production and minimize capital and operating expenses, companies are turning to advanced computing tools, like the MEMCPU™ Platform, to determine the most efficient and profitable strategy to extract these resources.

MemComputing is working with several of the world’s “supermajor” oil and gas companies on applications in the upstream sector. Here are a few areas of interest we are exploring with our partners: 

Exploration and Development:

  • Project portfolio selection and management
  • Drilling optimization
  • Staffing

 

Production

  • Well placement optimization
  • Annual delivery planning
  • Production optimization and scheduling

Let’s dive a bit deeper and review some of these challenges in more detail.

Exploration and Development

Exploration is the process of searching for rock formations associated with oil or natural gas deposits. Once companies have located an economically viable field, well development occurs. 

Collecting, processing, and analyzing huge volumes of seismic, well, and other geological data is a very complex and time-consuming process. 

This is understandable as companies need to be confident that the reserves they are pursuing will be profitable and worth the risk of investment. Therefore, they deploy various state-of-the-art computational tools and techniques to identify, characterize, and understand the value of potential reserves. 

Reservoir Modeling 

Reservoir modeling is a critical step in this process. Using an assortment of geological data, engineers and data scientists create three-dimensional computer models of a given reservoir. These models are then used for reservoir simulation, where scientists feed them real-time field data to accurately predict the reservoir’s behavior. These simulations enable companies to make informed field development decisions, such as optimizing well placement and estimating reserves. 

     

Challenges with reservoir modeling:

  • Accurately & efficiently representing the reservoir’s structural framework and geometry
  • Achieve a representative model at an appropriate scale and resolution
  • Uncertainty – multiple realizations of different scenarios of reservoirs are created

Once companies identify good, profitable reserves, they begin plans to start drilling. This brings on an array of new operational and strategic challenges where MemComputing can improve decision making.

Production

Production is the process of extracting hydrocarbons from the targeted reserves. Most oil and gas companies employ specialized drilling firms for this and must pay for their labor crews and rig expenses. Typically, these firms can only commit to a schedule that is six months out. 

If a company schedules a crew to drill, cap, or begin production and isn’t ready for them, they still must pay even if they aren’t working. It is therefore critical to forecast as precisely as possible where and when these crews will be needed. 

However, before drilling, they must decide where to place their wells. 

Well Placement 

The goal of well placement is to identify the most efficient and profitable locations to place wells in order to maximize the recovery of oil and gas while minimizing the capital and operational costs for a given reservoir.

Sounds practical enough, right?

As it turns out, this problem is extremely difficult to solve optimally, and data scientists and engineers have studied different techniques for decades. To give an idea of its complexities, here are some challenges:

  • Many variables to consider:
    • Geological (reservoir architecture, fluid contacts, etc.)
    • Production (number of wells, type, production rate)
    • Economic (fluid pricing/drilling costs)
  • Geological/economic uncertainties
  • Conventional/unconventional reservoirs
  • Maximize Time Value of Money
  • Adhere to environmental regulations

Identifying well placement is an iterative process that can take weeks. Engineers provide input data to a complex reservoir simulation program and review its output. Based on this output, they refine the input data and feed it back to determine the most profitable location to place their well. 

Each time, they evaluate how quickly and how well the output converges on an answer. That is, they run it over and over, especially when the answers vary widely from run to run. Eventually, they start narrowing in on a solution and subsequent runs make fewer and fewer adjustments. 

This is an imperfect science and is time consuming. Narrowing in on a solution can take weeks to compute using expensive High Performance Computing clusters. However, the biggest known cost is the time value of money. These reservoirs hold billions of dollars of fossil fuel. The sooner it can be retrieved, the sooner revenues can be realized. 

MemComputing cannot replace the reservoir simulation software; however, our solution can provide the input into the simulator and review the output in an automated fashion. Further, our solution evaluates all outputs in relation to all inputs and drives improved inputs for each subsequent run. This results in faster and more accurate convergence on the final well placement recommendation. The result is more accurate and obtained in a fraction of the traditional time. 

But who’s operating these rigs, and how can we maximize their labor? 

Crew Scheduling 

In order to ensure well productivity and minimize downtime, there must also be dynamic crew scheduling, another intractable optimization problem that greatly challenges oil and gas companies today. 

Challenges: 

  • Maximize portfolio value
  • Minimize crew idle time
  • Thousands of dollars per day per crew
  • Multi-year schedule

MemComputing offers a novel approach to solving scheduling problems at scale while satisfying all variables and operational constraints. Should a disruption occur, such as bad weather or equipment failure, the MEMCPU Platform can rapidly re-calculate the optimal schedule to minimize downtime and labor costs. 

Offshore Production

For offshore production, many more logistical challenges arise, and operational expenses can skyrocket. MemComputing has provided solutions helping oil and gas companies optimize their operations to drive new levels of efficiency, cutting costs by tens of millions of dollars annually, and reducing carbon output.

Here are a couple of use cases demonstrating this value with our clients. 

Helicopter Fleet Schedule Optimization 

MemComputing was employed by a large integrated oil and gas company to schedule a fleet of helicopters for offshore deliveries to oil rigs and support vessels. 

Due to the complex nature of this problem, the company had historically relied on heuristic and manual scheduling systems, which resulted in sub-optimal schedules and unavoidable operational expenses. 

The MEMCPU Platform instead provided a near-optimal schedule that was much more efficient than best-in-class optimization solvers and solved the problem orders of magnitude faster. Not only was the optimal scheduling solution found in seconds, but the resulting savings in time, fuel, costs, and reduced maintenance was proven to deliver millions in annual savings depending on the fleet size. 

Read the entire case study here, and for a visual representation of our solution, check out this brief video. 

Maritime Delivery to Oil Rigs

MemComputing has also released a case study demonstrating its ability to optimize the shipping schedule of necessary production materials (goods, wet and dry bulk, fuel) to offshore oil rigs from the port. 

Again, due to the large number of variables, constraints, and the complexity of this problem, it’s intractable to solve even using today’s supercomputers. However, our MEMCPU Platform handles this problem at scale. 

It provides an optimized schedule that significantly reduces the number of ships and transits required to deliver these products, showing a potential savings of millions of dollars monthly while also making a significant dent in its carbon footprint. 

Read the full case study here.

Additional Applications

In addition to the applications listed above, MemComputing is applicable to many other upstream problems such as:

  • Drone routing for inspection/delivery
  • Predictive maintenance
  • On-platform work planning 
  • Operator route optimization

 

Looking Ahead

As the amount of data collected by oil and gas companies continues to grow, data science teams are increasingly relying on new technologies like MemComputing to unlock new strategic insights to improve the bottom line.

 

With MemComputing, companies across the energy chain can transform enormous datasets into actionable and reliable solutions while dramatically reducing operational costs, increasing profitability, and maintaining environmentally friendly operations.  

MemComputing’s team is ready to advance your competitive edge with a tailored mathematical optimization solution. Sign up for a free account today to start solving.