Demand Response Outlook and Behavioural Economics - an Online Surveillance Approach

One of the main challenges of a quality demand response forecast is the lack of data. How to predict the consumer’s reaction to a never-felt stimulus? Classical models rely on a solid historical basis; without a past to learn from, it is difficult to predict the future.

A classical approach for innovation takes variations around the “S-curve" and requires a parametrization which is hard to obtain, as it (again) depends on the unknown.

Moreover, different innovations, experienced by different markets and cultures, exhibit distinct outputs. Figure 2 shows the various S-curves associated to the solar energy penetration in different European countries. Different innovations, experienced by different markets and cultures, exhibit distinct outputs. In other words, each market must be modeled according to specific characteristics and local consumer’s features.

Solar energy penetration - source: BP Statistical Review of World Energy (2017)

Objective

This paper proposes an approach to the short- to mid-term outlook of demand innovations, based on behavioral economics and consumer’s trends on-line, real-time monitoring. As history is still being made and consumers are still being exposed to novelties, there is no time to perform polls – and, if there was time, results would change before final result was processed. In this context, it is crucial to develop a continuous, self-adjustable monitoring procedure, able to capture and predict consumer’s evolution.

Methodology

This work uses an internet-based monitor. We look for consumer’s dynamics through crafted “trending topics” search as a “thermometer” for market evolution. As there is a delay between “talk” and “action”, it is possible to predict with a safe anticipation (six- to twelve months, in our experience) the load evolution and construct at least the next advancements of the S-curve.

Results

The paper discusses some possible algorithms for internet monitoring. However, initial tests are based on a simple data bank: the Google Trends, surveilled for one of our most important challenges: forecast the photovoltaic energy penetration - still in its earlier stages, not yet embraced by consumers, despite significant incentives (financing, transmission charges discounts)