Forecasting errors in power markets, even as small as 1%, can have significant financial implications. However, even high-performance artificial intelligence (AI) based electricity price forecasting (EPF) models have instances when their prediction error is much higher than those shown by mean performance metrics. To date, explainable AI has been used to enhance the model transparency and trustworthiness of AI-based EPF models. However, this paper demonstrates that insights from explainable AI (XAI) techniques can be expanded beyond its primary task of explanatory visualizations. This work presents a XAI-based error compensation approach to improve model performance and identify irregular predictions. The first phase of the proposed approach involves error quantification through a Shapley additive explanations (SHAP) based corrector model that fine-tunes the base predictor’s forecasts. Using this corrector model’s SHAP explanations, the proposed approach distinguishes high-accuracy predictions from lower ones in the second stage. Additionally, these explanations are more simplified than the base model, making them easier for non-expert users like bidding agents. Performance enhancement and insightful user-centric explanations are crucial for real-world scenarios like price spikes during network congestion, high renewable penetration, and fluctuating fuel costs. Case studies discussed here show the efficacy of the proposed approach independent of model architecture, feature combination, or behavioral patterns of electricity prices in different markets.

Solar power generation through photovoltaic technology is one of the most popular renewable energy sources. But solar energy is a non-dispatchable source and it is dynamic in nature. Hence from the power system operation point of view, solar power forecasting becomes imperative for a stable grid operation. In this paper, a novel hybrid machine learning approach is proposed for forecasting solar power generation through a hybrid Ensemble Averager technique which exploits the advantages of different machine learning approaches and incorporates them into a single model. Missing values in insolation have been dealt with using a univariate regression-based imputation technique. The ensemble averager is a weighted average model of five individual models, namely - a non-linear autoregressive neural network (NAR-NN), a non-linear autoregressive neural network with exogenous signal (NARX-NN), a least square boosted decision tree model, a support vector regressor with RBF kernel and an Extreme Learning Machine (ELM). The proposed model is tested on a real-world dataset of a 1 MW solar park situated in Gujarat, India (23°09'15.1”N 72°40'00.8”E). Proposed model shows better performance as compared to other models.

In this paper necessary changes, to a recently developed non-communication adaptive single pole auotreclosure scheme based on adaptive cumulative sum method (ACUSUM), are proposed. The work intends to design a real time implementable, simple and effective single pole adaptive autoreclosing scheme for extra high voltage (EHV) transmission line. A proposed change provides necessary immunity against noise. Also, the concept of deionization time is introduced to achieve successful reclosure. The computational redundancy is avoided to an extent by reducing monitored output variables, coupled with a simple hardware inclusion. Steps are incorporated against possibility of maloperation during non faulty condition of power system under consideration. The suggested changes are required to be incorporated for successful implementation of the method.

Insular grids are fragile owing to lower inertia and the absence of interconnection with other grids. With the increasing penetration of non-dispatchable renewable energy sources, the vulnerability of such insular grids increases further. The government of India has proposed several projects to improve the photovoltaic systems (PV) penetration in the Andaman and Nicobar Islands’ grid. This paper investigates joint stochastic scheduling of energy and reserve generation for insular grids fed from diesel and gas-based generators, PV, and battery energy storage systems (BESS). The proposed stochastic scheduling model considers a wide range of probabilistic forecast scenarios instead of a deterministic model that assumes a single-point forecast. Hence, it provides an optimal solution that is technically feasible for a wide range of PV power forecast scenarios. The striking feature of the model developed in this work is the inclusion of stochastic constraints that represent (i) the coordination between PV and BESS, (ii) reserve constraints, (iii) battery charging/discharging limit constraints, and (iv) non-anticipatory constraints that ensure technical viability of scheduling decisions. The proposed model is validated on the dataset for South Andaman Island. Results reveal the applicability and feasibility of the proposed stochastic dispatch model for different generation mix scenarios.