Yueh Huang, Risk Systems developer and analyst, and Clive Merifield, Business Development Manager, at ZTP, the UK energy consultancy and software specialist.
COVID19 has materially changed the level of focus applied to energy consumption forecasting. Invariably the pandemic is driving the requirement for many organisations to re-forecast, and we are finding that this is having to be undertaken on a much more frequent basis due to the heightened levels of uncertainty organisations are currently experiencing.
The consumption forecast forms the basis of an organisation’s energy budget. Any change of consumption forecast will therefore directly impact an organisation’s financial forecast/budget, which will itself require reforecasting to ensure the correct financial resources are allocated. Budget forecasting and reforecasting are best undertaken on a platform which offers meter level granularity in order to improve forecast accuracy and to automate workload.
Significant price risk management and hedging implications also result from volume forecast changes. From a strategy perspective – an organisation may need to change its stance on hedging, perhaps moving away from trading its entire forecast volume months/years in advance, leaving a proportion open to maintain some flexibility – but this change will have implications which require consideration and modelling etc. Secondly, from a trading perspective, the revised forecast may require costly corrective trades to be placed and it is highly desirable to avoid this where possible
The supplier’s perspective must not be forgotten either. It is in the supplier’s interest for consumption forecasts to be accurate. Inaccuracy will lead to additional cost. This extra cost invariably finds its way onto the customer’s bill – in the simplest form via the volume tolerance clause.
For the reasons outlined previously (amongst many others) it is imperative that energy teams proactively monitor actual consumption against forecast consumption and keep a constant eye on contractual volume clauses. A basic framework of consumption forecasting can be established by normalising the energy usage with historical operating activities and weather conditions. Based on the types of business, the operating activities can be captured by different variables such as operating hours (office), footfall (retail) or production (manufacturing). Commonly, weather condition is proxied by Heating Degree Days (HDD), Cooling Degree Days (CDD) and/or Composite Weather Variable (CWV). Figure 1 shows a linear relationship between LDZ non-daily metered demand and CWV and indicates that CWV can explain approximate 95 percent variation of the demand.
After the normalised consumption is generated, it is vital to check if parameters used for forecasting need to be finetuned due to the change of current conditions, such as the substitution for high energy-consuming machines or upgrades of a HVAC system. In order to do this effectively, an organisation needs a platform which tracks actual consumption data against the consumption forecast. Ideally the platform will track actual consumption against forecast on a daily basis. Any discrepancy implies possible parameters adjustment. The platform should therefore provide automated key stakeholder alerts where consumption is atypical, or there is a risk of triggering a tolerance clause. Additionally, it is important to be able to quickly identify which sites and meters are over or under consuming to speed up cause identification.
COVID19 is causing organisations to re-forecast consumption on a more regular basis – creating consequences for existing energy budgets and hedging strategies. Energy teams should consider engaging stakeholders to develop a basic framework of consumption forecasting by normalising the energy usage and adapting any on-going changes to improve consumption forecast accuracy.
The proactive monitoring of actual consumption against the forecast is now imperative. In order to effectively monitor forecast accuracy, an organisation needs a digitalised platform which tracks actual consumption data at meter level against the consumption forecast, alerting key stakeholders automatically so that any variance can be investigated and resolved.