Modern power grids become ever more complex due to the integration of renewable energy sources and decentralized energy production, among other factors. This presents significant challenges for monitoring, diagnosing, and explaining grid anomalies. To address such setting, the SENSE framework generates contextualized, user-centred explanations of anomalous events in smart grids. To that end, it makes use of semantic web technologies in order to integrate sensor data with domain expert knowledge, to handle system events, and to derive relevant explanations. We evaluate SENSE in two use cases: a real-world smart charging garage and a simulated local energy community. Our evaluation shows that SENSE delivers accurate explanations, with an average accuracy of 82%, while being perceived as highly usable by domain experts. The results demonstrate the potential of semantic web technologies to increase transparency in smart grids leading to important, longer-term economic and environmental impacts.