Delivering Better Decisions, Faster Planning And Trading Of Energy

Delivering Better Decisions, Faster Planning And Trading Of Energy
Spread the love

Digitalisation is about automating workflows and reducing the degree of assumptions in decision-making, enabling more precise executions at all levels to provide higher assurance of delivering the desired outcome. Article by Andrew McIntee, Marketing Manager, KBC (A Yokogawa Company).

Across organisations, decision-making should be about seeking to improve performance. The challenge is that decisions comprise of different layers of factors, all on different timescales, each with different levels of uncertainty and varying degrees of impact.

When any part of a decision-making process is not fully known, assumptions must be made. With assumptions, come increased levels of uncertainty of achieving the desired outcome.

Digitalisation is about automating workflows and reducing the degree of assumptions in decision-making, enabling more precise executions at all levels to provide higher assurance of delivering the desired outcome.

In the decision-making process, four key areas are being improved:

  • Higher sufficiency and fidelity of data
  • Deeper understanding of patterns and inter-relationships
  • More effective dissemination of key information
  • More action-oriented foresight

For scheduling engineers aiming to deliver the most economic dispatch of energy, have to contend with factors such as electricity price contracts (eg: real-time pricing, time of use pricing, critical peak pricing), variability of natural gas prices, the capability of process plants to become electricity providers to the grid and a wide variety of energy sources (fossil fuels, renewables, etc). Accommodating these with multiple time-sensitive constraints over various time horizons imposes additional challenges.

Visual MESA Multi-Period Optimizer (VM-MPO) brings together data analytics, first principles digital twins of energy systems and multi-period constraints in a purpose-built mixed integer optimization to continually ensure that the right decisions are made about where to deploy energy at lowest economic cost. VM-MPO enhances the energy and chemical industry’s leading real-time optimization technology, Visual MESATM Energy Real-Time Optimizer (VM-ERTO) by adding an upper decision layer where the time-sensitive variables are optimally defined.

It allows scheduling engineers to incorporate more precise forecasts into their decision-making processes, enabling more aggressive actions to achieve the optimum, thereby yielding incremental benefits.

Delivering Better Decisions, Faster Planning And Trading Of Energy
Figure 1: The impact of digitalisation on the decision-making process.

Working At Your Best Within Your Constraints

Working within your hard constraints, such as pump capacity and temperature limits, are well understood and easy to incorporate. However, there are additional time-sensitive, or multi-period constraints, for example:

  • Planned out-of-service periods
  • Caps on emissions
  • Fuel tank farms management (inventory)
  • Minimum/ Maximum Start / Stop time of equipment
  • Thermal energy storage management (inventory)

Businesses must always strive to maintain their “licence to operate”. To achieve this, VM-MPO accommodates inclusion of operational, safety, legal, environmental and contractual constraints, such as CO2 emissions trading quotas, changes of loads, yield shifts and product qualities.

Using the constraints, the VM-MPO model is built as a digital twin, representing the system’s behaviour under different operating conditions. First principles-based analytics tools respect the laws of nature, handle non-linearities and complex relationships. VM-MPO is then structured such that the interdependency of the variables through time can be identified and the optimization can solve at speed

Delivering Better Decisions, Faster Planning And Trading Of Energy
Figure 2: Layers of forecasting analytical rigour.

Going Beyond The Horizon

Over time there has been an evolution in the level of accuracy of forecasting analytics in optimisation problems:

  • From assuming last month’s performance will be repeated;
  • To simple spreadsheet-based calculations;
  • To statistical data analytics predictions with cause and effect; to
  • To a sophisticated ensemble which fuses statistical data analytics, high fidelity first principles digital twins of energy systems and multi-period constraints in a purpose-built, mixed integer optimisation.

Figure 2 illustrates the layers of analytics technology advancement with the key difference being the depth of action-oriented foresight associated with each.

 

Assuring Your Ability To Execute Decisions

Value is captured when the right decision is made and there is effective follow-up on execution of resulting actions. However, in many cases action execution is what limits value capture.

VM-MPO provides tailored outputs and actions to operators – in open-loop mode – with clear descriptions of actions to be taken, time frames to execute the actions and quantification of benefits of action execution.

There is always the option to remove reliance on operators and organizational silos altogether by shifting to closed-loop optimisation. This approach has been proven to be successful with sites that have improved computational capabilities, a culture that treats real-time data as an asset and values the continual technical usability of the solutions.

 

Delivering Better Decisions, Faster Planning And Trading Of Energy
Figure 3: Optimal TES operation.

Case Study

Houston-based Thermal Energy Corporation (TECO) has provided cooling and heating to institutions in the Texas Medical Center, which covers more than 19 million square feet of customer buildings at 18 institutions.

In the first stage of the implementation, VM-ERTO was used to provide the optimal load of chillers and boilers. This saved US$1.1 million per year of total operating costs through implementing the optimal recommendations.

In the second stage, the optimal operation of the Thermal Energy Storage (TES) was included. The main objective was to identify how much chilled water should be stored or used at each time of the day in order to minimise the electric power that is consumed.

Given that the main energy source of this system was electric power, predicting its day-ahead price was very important. To achieve this, VM-MPO accessed day-ahead market electricity price forecasts for different hubs as well as real-time price.

Figure 3 shows the optimal operation. The dark grey bars represent the charge/discharge of the TES, whereas the light grey area represents the TES accumulation of chilled water. Adding this extra layer in the decision-making process would increase the benefits by approximately a 10% increase on top of the VM-ERTO savings.

 

Summary

VM-MPO brings together data, first principles and multi-period constraints in a purpose-built multi-period, mixed integer optimisation to continually ensure that the right decisions are made about which generation assets to start up, shut down and where to deploy energy at lowest economic cost over the time horizon.

VM-MPO enhances real-time optimisation, by adding an upper decision layer where the time-sensitive variables are optimally defined to be able to solve for multi-period constraints. VM-MPO forms an important part of an integrated optimal scheduling and real-time optimization offering for energy systems. It enables better decisions, faster for planning, scheduling and trading of energy.

 

FOLLOW IAA ON FACEBOOK, LINKEDIN AND TWITTER!

CHECK OUT OUR LATEST ISSUE!

WANT MORE INDUSTRY INSIGHTS? SUBSCRIBE TO IAA NOW!

 

 

Of Power, Smart City And Sustainability
Megatrends In Energy Create A Path To Innovative Solutions

Scroll to Top