Skip to main content

Shell saves 2MUSD through energy optimization

Electricity represents a major cost at Nyhamna operations. Identifying optimization opportunities while ensuring stable and safe operations were part of this challenge. Find out how Norske Shell achieved energy optimization and savings with our dynamic digital twin – Kognitwin Energy.

Challenge

With a projected annual cost of 900 million NOK - 1 billion NOK, electricity is the single highest cost at Nyhamna, and 85-90% of the electricity consumption origins from natural gas compression.

The Nyhamna plant exists of two main parallel processing units (production trains).  Within each of these trains, and in most of the utility systems, there is numerous equipment in parallel operations.

Based on the operational conditions and external factors, the number of operating equipment and their set points need to be adjusted in order to be optimized in terms of energy.

Twinenergyop1.png

For this use case, we have mainly focused on the following equipment:

▪ Sea water pumps (4 in total)

▪ Sea water/main cooling medium heat exchangers (10 in total)

▪ Cooling medium pumps (3 in total)

▪ Export compressors (4 in total)

The main focus for the control room operators at any plant is to ensure safe and stable operations of the facility. Our industry experience shows that it is often difficult to identify optimization opportunities with a manual approach. Especially those opportunities that are limited in time, or outside the already identified optimization scope.

We like to look at this from the perspective of chess. While the global top players can find amazing opportunities after maybe 40 min of thinking, the best chess computers are able to find an even better move in seconds. If you bring the chess computer to the player, their joint capacity is exceeding their isolated capacities.

In an energy facility case, as Nyhamna, Kognitwin Energy plays the ‘chess computer’ role, adding digitalization and advanced intelligence to the table.

Some of the risks and challenges that limit efficient energy optimization:

  • For short time opportunities, that could save a lot of power, energy companies are likely to miss with today’s tools. No automatic proposal for optimization opportunities.
  • No dynamic tool to check the effects of potential adjustments (equipment/set points) in the facility. It is not worthwhile to try out optimizations if you can’t ensure safe and stable operations.
  • Large amounts of recycling in both sea water and cooling medium, which is not efficient, and it would be better to either shut down equipment or utilize the capacity in a better way.
  • Too low flow in heat exchangers → Marine growth

 

Solution

To support Shell on optimizing Nyhamna in terms of energy, the dynamic K-spice process simulator runs alongside the physical plant. This model is “always on” and enable real-time comparison between measured sensor data in the plant and simulated data from the high-fidelity model.

Dynamic simulations and operator training simulators are at the core of Kongsberg domain for the last 30 years, and it is a fundamental capability in the Kognitwin Energy deliveries.

The “always on” simulator runs in a cloud environment and is accessible for the plant’s operations and process engineers, being effortless to start a new parallel simulation, either initialized with the current state or historical conditions. With this solution, Shell has full flexibility to test out optimization scenarios in a safe way before implementing them in the physical plant.

From low hanging fruits to endless flexibility and opportunities for the operators. Kognitwin Energy doesn’t stop there. Combined with Hybrid ML, machine learning algorithms automatically add extra features and search for optimization prospects.

The solution that is now running in the cloud environment and by the click of a button, 4 parallel model containers are started, each containing the full dynamic process model for Nyhamna. The models are initialized in the following way:

  • Similar to the always on model, runs on current conditions.
  • A look a head model, takes the current conditions and runs it much faster than real-time.
  • Conservative optimization opportunity from the ML models
  • More aggressive optimization proposal from the ML models.

This is called a what-if analysis, and the results from all the simulator runs are collected in a graphical user interface in the web. This interface will tell users which of scenarios have the best performance in terms of power consumption. All the models are currently constrained to maximize the production rate.

Twinenergyop2.png

Kognitwin Energy brings information to operators and recommendations, like:

  • Number of export compressor trains to operate
  • Optimal number of utility equipment in use (sea water pumps, cooling medium pumps, etc.)
  • Various system setpoints (temperature, pressure, etc.)
  • Opportunities for uneven loading between parallel compressors
  • Potential operational constraints/integrity issues, e.g. high velocities

 

Impact

As a result, Norske Shell achieved in the Nyhamna facility:

  • 1 MW saving represents 4-5 million NOK in OPEX mitigation, with this tool it’s already identified savings of 4 MW.
  • Dynamic testing of all planned changes in the plant directly from a web interface.
  • Short time opportunities saving significant power.

Tailored user interface empowering operators to know exactly which adjustments to make.