Abstract:This paper studies the use of conformal prediction (CP), an emerging probabilistic forecasting method, for day-ahead photovoltaic power predictions to enhance participation in electricity markets. First, machine learning models are used to construct point predictions. Thereafter, several variants of CP are implemented to quantify the uncertainty of those predictions by creating CP intervals and cumulative distribution functions. Optimal quantity bids for the electricity market are estimated using several bidding strategies under uncertainty, namely: trust-the-forecast, worst-case, Newsvendor and expected utility maximization (EUM). Results show that CP in combination with k-nearest neighbors and/or Mondrian binning outperforms its corresponding linear quantile regressors. Using CP in combination with certain bidding strategies can yield high profit with minimal energy imbalance. In concrete, using conformal predictive systems with k-nearest neighbors and Mondrian binning after random forest regression yields the best profit and imbalance regardless of the decision-making strategy. Combining this uncertainty quantification method with the EUM strategy with conditional value at risk (CVaR) can yield up to 93\% of the potential profit with minimal energy imbalance.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to study the use of **Conformal Prediction (CP)**, an emerging probabilistic prediction method, to increase the participation of photovoltaic (PV) power in the electricity market. Specifically, the goal of this research is to improve the day - ahead prediction of PV power through the CP method, thereby helping market participants make better decisions in the electricity market. #### Main problems 1. **Improve the reliability of PV power prediction**: - The prediction of PV power is crucial for power suppliers and grid operators. Accurate prediction can reduce the imbalance cost in the real - time market and optimize the trading strategies in the electricity market. - Traditional machine - learning models have difficulty accurately estimating the uncertainty of prediction, while the CP method can quantify this uncertainty by providing prediction intervals or cumulative distribution functions. 2. **Optimize the bidding strategies in the electricity market**: - In the day - ahead market (DAM), suppliers need to provide the amount of electricity they can supply and their minimum selling price one day in advance. If the actual power generation does not match the bid amount, suppliers will face high imbalance costs. - The paper explores several bidding strategies, including Trust - the - forecast, Worst - case, Newsvendor, and Expected Utility Maximization (EUM), to determine the optimal bid amount, so as to achieve high profits and minimal energy imbalance. 3. **Combine the CP method and bidding strategies**: - The research found that combining the CP method with certain bidding strategies (such as EUM and Conditional Value at Risk (CVaR)) can achieve up to 93% of the potential profit while ensuring minimal energy imbalance. #### Method overview 1. **Point prediction models**: - Use machine - learning models such as simple linear regression (SLR), multiple linear regression (MLR), and random forest regression (RFR) for point prediction. 2. **Uncertainty quantification methods**: - Implement multiple CP variants, including Basic CP, k - nearest neighbor (KNN) - based Conformal Prediction, Mondrian - binning - based Conformal Prediction, and Conformal Prediction System (CPS). 3. **Bidding strategies**: - Apply different bidding strategies, such as Trust - the - forecast, Worst - case, Newsvendor, and Expected Utility Maximization, to determine the optimal bid amount. #### Data and experiments - Use the public data set of 175 PV systems in Utrecht province, the Netherlands, with a time span from 2014 to 2017. - Divide the data into training set (2014 - 2015), calibration set (2016), and test set (2017). - Combine weather forecast data and other physical characteristics (such as solar radiation, wind speed, etc.) for prediction. Through these methods, the paper shows how to use the CP method to improve the reliability of PV power prediction, and combine different bidding strategies to optimize the participation in the electricity market, thereby achieving higher profits and smaller energy imbalances.

Conformal Prediction for Stochastic Decision-Making of PV Power in Electricity Markets

Conformal Prediction Interval Estimations with an Application to Day-Ahead and Intraday Power Markets

A Hybrid Approach for Probabilistic Forecasting of Electricity Price.

Probabilistic Forecasting of Photovoltaic Generation: an Efficient Statistical Approach

Nonparametric Probabilistic Forecasting of Regional Photovoltaic Power Based on Spatial Clustering and Combining Quantile Regression

On-line conformalized neural networks ensembles for probabilistic forecasting of day-ahead electricity prices

Forecasting-based electricity tariff selection for resident users with photovoltaic and energy storage considering forecast uncertainties

Cost-Oriented Prediction Intervals: on Bridging the Gap Between Forecasting and Decision

Combining predictive distributions of electricity prices: Does minimizing the CRPS lead to optimal decisions in day-ahead bidding?

Peer-to-Peer Energy Trading Using Prediction Intervals of Renewable Energy Generation.

Probabilistic day-ahead prediction of PV generation. A comparative analysis of forecasting methodologies and of the factors influencing accuracy

Probabilistic forecasting of photovoltaic power supply — A hybrid approach using D-vine copulas to model spatial dependencies

Short-Term Photovoltaic Power Probabilistic Forecasting Based on Temporal Decomposition and Vine Copula

Short-term photovoltaic power prediction on modal reconstruction: A novel hybrid model approach

Conformal Uncertainty Quantification of Electricity Price Predictions for Risk-Averse Storage Arbitrage

Electricity price forecast in wholesale markets using conformal prediction: Case study in Mexico

Forecasting and uncertainty analysis of day-ahead photovoltaic power using a novel forecasting method

Postprocessing of point predictions for probabilistic forecasting of day-ahead electricity prices: The benefits of using isotonic distributional regression

Review on probabilistic forecasting of photovoltaic power production and electricity consumption

An Advanced Hybrid Boot-LSTM-ICSO-PP Approach for Day-Ahead Probabilistic PV Power Yield Forecasting and Intra-Hour Power Fluctuation Estimation

Forecasting Intraday Power Output by a Set of PV Systems using Recurrent Neural Networks and Physical Covariates