Natural Gas Forecasting

Improving gas demand forecasting performance is a key factor in achieving efficient and profitable gas operations in a liberalised market. Poor forecasting performance can lead to high imbalance charges and operating costs.

GL's modelling techniques, have been captured in a suite of forecasting tools, including our Forecaster product supported by our team of highly qualified and experienced gas forecasting and modelling analysts. 

The value of accurate demand forecasting

Shippers/Suppliers need to provide nominations to gas transporters for the amount of gas they expect to provide to customers on a daily basis.  Improved demand forecasting reduces exposure to imbalance charges and the risk of having to buy or sell on the spot market during unfavourable conditions.  Improved medium term forecasting means better decisions can be made when buying longer term contracts and optimising the use of storage facilities.

For Transporters accurate forecasts are a key factor in ensuring security of supply. By understanding where gas demand is required on the network operating costs can be significantly reduced.  Improved utilisation of the transmission networks means that spare capacity can be sold to other pipeline users to maximise revenue.

Some of our clients achieve annual savings of over £1M pa with relatively small improvements in forecasting accuracy.  The savings will depend on existing forecasting capability and this is something GL can help you quantify and improve.

GL Forecasting Solutions

GL offer a range of forecasting solutions from small forecasting capability studies through to full turn-key forecasting systems.

Small Study
GL will examine a subset of your demand portfolio and carry out a short analysis and modelling assignment to demonstrate the potential accuracy achieved and provide a demonstration of Forecaster using your data.

Trial
This is very much a try before you buy approach.  Rather than go through the expense of a full implementation, why not implement our software and carry out a trial on a subset of your portfolio?  This can help support the business case for implementation or highlight areas for custom enhancements to meet your specific needs.

Standard Forecaster implementation
Our Standard Forecast solution comes with a pre-configured database setup with a set of generic models that will initially provide good quality forecasts in a variety of situations.  Clients can modify or create their own models or seek advice and guidance from GL.  A standard suite of training courses are available to ensure that clients get the most out of this cost effective solution.

Enhanced Forecaster implementation
This very much a turn key solution where GL can provide a variety of services including: custom configuration, data migration, interface development, analysis and configuration of models, installation, testing and end user training.

Enhanced Support
Our Enhanced Support service provides an agreed Service Level Agreement for key systems and covers the Forecaster set of utilities and any other custom developments clients have selected.  This is available for both Standard and Enhanced Forecaster implementations.

Forecaster key features

• A variety of Forecasting modelling methods including:
  Neural Networks
  Multiple linear regression
  Multiple Non-linear piecewise regression with cross terms
  ARIMA (Auto Regressive Integrated Moving Average)
  Adaptive Combination, several model results are combined into one forecast
  Profiler - uses historical data to generate hourly/daily demand shapes
  Bayes
• Scheduled forecasts and model training for unmanned operation
• Configurable forecast horizon for short, medium or long term forecasts
• The database includes a set of Oracle views so that clients can use their own reporting tools to interrogate the Forecaster system
• Hourly (profiles) or daily forecasts
• Flexible definition of model input variables and input data streams
• A powerful model configuration utility to specifiy model parameters, model specific  settings, forecast horizons and data intervals
• The ability to create individual and/or aggregate customer forecasts at customised portfolio or operational levels
• The ability to account for special events such as bank holidays and known shutdowns
• A variety of graphs and performance metrics showing forecast errors and performance against actual consumptions
• Facility to export information to Excel