Intelligence in C/ETRM – Automation Examples

Post by Pioneer Solutions on July 3, 2018

Artificial Intelligence (AI), Machine Learning (ML), Robotic Process Automation (RPA), and of course digitization have been and still are hot buzzwords. AI has been around for some time, while the term Machine Learning is a more recent term, often opportunistically used to rebrand AI. Machine Learning can be considered a current application of AI based around the idea to let machines access data and let them learn for themselves. Rule-based systems use rules prepared by humans, and some may see it as old school AI. With the increase of digital information being available, both Rule-based systems and ML have a lot to offer. With their promise of automating mundane tasks as well as offering creative insight, the energy trading market has been recognizing the benefits. Next generation ETRM systems have adopted AI, thereby embedding intelligence in various functional applications as routine technology as illustrated in the following Automation Examples.


The development of neural networks has been key to teaching computers to think and understand the world in the way humans do. Essentially it works on a system of probability – based on data fed to it, it can make statements, decisions or predictions with a degree of certainty. By sensing or being told whether its decisions are right or wrong, it modifies the approach it takes in the future.

This neural network method is used in short-term load and price forecasting, where it will use historical price and demand data, forecasted weather data, and calendar attributes such as on-peak, off-peak hours, weekdays, weekend days, holidays etc. Under the neural network method, historical input data are divided into train and test sets. By taking the train data set, the method performs optimizations in an iterative fashion and provides the optimized model parameters. Those model parameters are used in the test data for the model validation. Once the models are validated the optimized model parameters with new input data are fed in the model for the load and/or price forecasting.


AI often revolves around the use of algorithms. An algorithm is a set of unambiguous instructions that a computer can execute. Algorithms can perform calculation, data processing and automated reasoning tasks. Where the algorithm is trying to solve an actual problem, the formulas are used in the process.

In trading, algorithms are used to simulate or optimize operations. Modern ETRM systems support the ability to create What-if scenarios, to see the impact of price and volume changes on portfolios and P&L. Mechanisms can be set up in support of market strategies based on set levels, averages, deltas, or targets for volume and price. A ‘live’ ETRM system facilitates such operations by reporting position changes and receiving market information in real-time and having direct interfaces with the various Trading Exchanges.

Flow Optimization

In mathematical optimization, Dantzig’s simplex algorithm is a popular algorithm for linear programming.

Such linear programming simplex method can be used for Gas Flow optimization, to allow users to schedule gas from a receipt point to a delivery point. Gas can be scheduled manually, or the system can propose an optimized flow based on constraints, such as location, pipeline, contracts, delivery path, supply and demand side. Flow Optimization can automate the complex equity gas nominations and the optimization of gas flows to meet demands. Based on the results that the optimizer provides, remaining Path MDQ and End positions on Supply and Demand side will change. The Flow Optimization/Scheduling process creates nomination records that can be submitted to the Pipeline.


Often used in artificial intelligence applications, the term ‘rule-based system’ is applied to systems involving human-crafted or curated rule sets. This comes in handy, when wanting to apply Robotic Process Automation (RPA).  RPA is a software-based approach, where a virtual worker replicates the user actions of a human to further optimize for operational accuracy, cost and speed.

This is very suitable in especially the routine and rules-driven tasks in the back office but does require a sufficiently automated rule-based back office process. When a C/ETRM system provides an integrated solution, the back-office will benefit from straight-through-processing and built-in workflow management thereby streamlining the contract-to-bill process. A C/ETRM system that has a notion of Billing Determinants and Charges Types, will be able to automate 95% of all contracts, whether EFET / ISDA master agreement, complex PPA, Intercompany, or Generation and Renewables contracts. The combination of a highly configurable C/ETRM module that can capture and automate the rule-based invoicing, and RPA will free up human resource capital and allow companies to achieve and maintain operational excellence.

We could offer more examples of automation where intelligent rule-based functions are used, such as in a Dispatch Cost Evaluator or Storage Optimization, but you get the idea of how modern C/ETRM systems have adopted AI. Research in Artificial Intelligence started in the 1950s, experienced a revival in the 1980s, and now in the 21st century experienced a resurgence following advances in computer power and ‘big data’. As a result, Artificial Intelligence techniques have become an essential part of business software applications, including next generation ETRM systems.


Leave a Reply

Your email address will not be published. Required fields are marked *

+ 49 = 52