Modelling Irrational Behaviour of Residential End Users using Non-Stationary Gaussian Processes

Nam Trong Dinh,Sahand Karimi-Arpanahi,Rui Yuan,S. Ali Pourmousavi,Mingyu Guo,Jon A. R. Liisberg,Julian Lemos-Vinasco

2024-03-26

Abstract:Demand response (DR) plays a critical role in ensuring efficient electricity consumption and optimal use of network assets. Yet, existing DR models often overlook a crucial element, the irrational behaviour of electricity end users. In this work, we propose a price-responsive model that incorporates key aspects of end-user irrationality, specifically loss aversion, time inconsistency, and bounded rationality. To this end, we first develop a framework that uses Multiple Seasonal-Trend decomposition using Loess (MSTL) and non-stationary Gaussian processes to model the randomness in the electricity consumption by residential consumers. The impact of this model is then evaluated through a community battery storage (CBS) business model. Additionally, we apply a chance-constrained optimisation model for CBS operation that deals with the unpredictability of the end-user irrationality. Our simulations using real-world data show that the proposed DR model provides a more realistic estimate of end-user price-responsive behaviour when considering irrationality. Compared to a deterministic model that cannot fully take into account the irrational behaviour of end users, the chance-constrained CBS operation model yields an additional 19% revenue. Lastly, the business model reduces the electricity costs of solar end users by 11%.

Systems and Control,Information Theory

What problem does this paper attempt to address?

The key problem that this paper attempts to solve is how to effectively integrate the irrational behaviors of electricity end - users in the Demand Response (DR) model, especially loss aversion, time inconsistency, and bounded rationality. Existing DR models usually assume that users are fully rational, that is, they will adjust their electricity consumption according to price signals to minimize electricity bills, but this assumption may not hold in the long run. Behavioral economics points out that in the actual decision - making process, users often lack full rationality, which may lead to wrong predictions made by models based on the assumption of rational users, thus bringing significant economic losses.

To solve this problem, the author proposes a new price - responsive DR model, which specifically takes into account the irrational behaviors of users and is implemented through the following methods:

1. **Introducing Irrational Behavior Characteristics**: The model integrates three key aspects: loss aversion, time inconsistency, and bounded rationality. Specifically:
- **Loss Aversion**: Users will feel uncomfortable when reducing their electricity consumption, while the comfort brought by increasing electricity consumption is relatively small.
- **Time Inconsistency**: Users have biases in evaluating their behaviors in future time periods, tending to overestimate the importance of current decisions and underestimate the impact of the future.
- **Bounded Rationality**: Users' decisions are restricted by cognitive limitations, causing them to choose sub - optimal solutions.

2. **Modeling Randomness Using Non - stationary Gaussian Processes**: To capture the random components in users' electricity consumption behaviors, the author adopts Multi - Seasonal Trend Decomposition (MSTL) combined with Non - stationary Gaussian Processes. This method can handle the uncertainties in users' behaviors and generate more realistic predictions.

3. **Applying Chance - Constrained Optimization Models to Manage Community Battery Storage (CBS) Systems**: Considering the unpredictability of users' behaviors, the author proposes a chance - constrained optimization model to guide the operation of CBS, ensuring that it can still operate efficiently in the face of users' irrational behaviors.

Through these improvements, this model not only improves the prediction accuracy of users' price - responsive behaviors but also significantly enhances the economic benefits of CBS projects. For example, compared with deterministic models, this model increases the CBS operating income by 19% and reduces the electricity bill costs of solar power users by 11% at the same time.

In conclusion, the main objective of this paper is to develop a more realistic demand - response model to better understand and predict the irrational behaviors of electricity end - users, thereby improving the efficiency and economy of energy systems.