The key challenge for household load forecasting lies in the high volatility and uncertainty of load profiles. Traditional methods tend to avoid such uncertainty by load aggregation (to offset uncertainties), customer classification (to cluster uncertainties) and spectral analysis (to filter out uncertainties). This paper, for the first time, aims to directly learn the uncertainty by applying a new breed of machine learning algorithms—deep learning. However, simply adding layers in neural networks will cap the forecasting performance due to the occurrence of over-fitting. A novel pooling-based deep recurrent neural network is proposed in this paper which batches a group of customers’ load profiles into a pool of inputs. Essentially the model could address the over-fitting issue by increasing data diversity and volume. This paper reports the first attempts to develop a bespoke deep learning application for household load forecasting and achieved preliminary success. The developed method is implemented on Tensorflow deep learning platform and tested on 920 smart metered customers from Ireland. Compared with the state-of-the-art techniques in household load forecasting, the proposed method outperforms ARIMA by 19.5%, SVR by 13.1% and classical deep RNN by 6.5% in terms of RMSE.

In the future power system, an increasing number of distributed energy resources will be integrated including intermittent generation like photovoltaic (PV) and flexible demand like electric vehicles (EVs). It has been long thought to utilize the flexible demand to absorb the PV output locally with the technical solution proposed while an effective commercial arrangement is yet to be developed due to the significant uncertainty associated with local generation. This paper proposes a peer-to-peer (P2P) local electricity market model incorporating both energy trading and uncertainty trading simultaneously. The novelty is to match the forecast power with demand having time flexibility and the uncertain power with demand having power flexibility. This market enables more PV uncertainty to be balanced locally rather than propagating to the upper layer system. In the test case, 55.3% of PV forecast error can be balanced locally in the proposed joint market. In comparison, 43.6% of PV forecast error is balanced locally when the forecast power and uncertain power are traded separately in a day-ahead market and a real-time market. The proposed P2P market can also motivate PV owners to improve forecast accuracy.

Phase identification is a process to determine which of the three phases a particular house is connected to. The state-of-the-art identification methods usually exploit smart metering data. However, the data sets are not always available and the major challenge is hence to identify phases with incomplete data set. This paper proposes a novel spectral and saliency analysis identification method to overcome this hurdle. Spectral analysis is first performed to extract the high-frequency features from the incomplete data. Saliency analysis is then adopted to extract salient features from the variations of high-frequency loads in the time domain. Correlation analysis between customer features and the phase features is used to determine customers' phase connectivity. The method is executed iteratively until all customers with smart meters have been allocated to a specific phase or no salient features can be found. It is validated against real data from over 6000 smart meters in Ireland and achieves an accuracy of over 93% with only 10% smart meter penetration ratio in a 100-household network.

The interconnected converter (IC) links the ac and dc microgrids (MGs) and exchanges powers to maintain the hybrid system stability. However, inertia stored in the ac or dc side cannot be transferred through IC to support each other flexibly according to demand. This paper proposes a novel bidirectional virtual inertia (BVI) control strategy for the IC with an inertia transferring mechanism. IC actively transfers the inertial power to balance the transient performance of both sides of the interconnected MGs. The adaptive strategy is proposed for coefficients of the BVI control to change the inertial power flexibly according to the inertia gap caused by disturbances. Experimental results demonstrate that the proposed strategy of transferring inertia strikes a transient balance between the dc voltage and ac frequency and hence, enhance the transient stability of the hybrid system.

Hierarchical load forecasting (HLF) is an approach to generate forecasts for hierarchical load time series. The performance of HLF can be improved by optimizing the forecasting model and the hierarchical structure . Previous studies mostly focus on the forecasting model while the hierarchical structure is usually formed by clustering of natural attributes like geography, customer type, or the similarities between load profiles. A major limitation of these natural hierarchical structures is the mismatched objectives between clustering and forecasting. Clustering aims to minimize the dissimilarity among customers of a group while forecasting aims to minimize their forecasting errors. The two independent optimizations could limit the overall forecasting performance. Hence, this paper attempts to integrate the hierarchical structure and the forecasting model by a novel closed-loop clustering (CLC) algorithm. It links the objectives of forecasting and clustering by a feedback mechanism to return the goodness-of-fit as the criterion for the clustering. In this way, the hierarchical structure is enhanced by re-assigning the cluster membership and the parameters of the forecasting models are updated iteratively. The method is comparatively assessed with existing HLF methods. Using the same forecasting model, the proposed hierarchical structure outperforms the bottom-up structure by 52.20%, ensemble-based structure by 26.89%, load-profile structures by 19.90%, respectively.

More accurate forecasts may not necessarily lead to better decision-making. To address this challenge, decision-focused learning (DFL) has been proposed as a new branch of machine learning that replaces traditional statistical loss with a decision loss to form an end-to-end paradigm. Applications of DFL in power systems have been developed in recent years. However, existing applications remain fragmented without systematic analysis of methodologies or comparative benchmarks. This review addresses this gap by performing a set of scenario analysis, taxonomy analysis, application analysis and comparative analysis. It first illustrates the inherent mismatch between statistical accuracy and operational decisions through power system example scenarios. It then establishes a structured taxonomy of DFL techniques, categorizing methods by model structure (direct/indirect) and gradient handling (gradient-based/free). An application-based analysis reviews existing DFL applications by forecasting targets and decision contexts. Furthermore, an open-source comparative benchmark is developed to assess different DFL models through power system-specific metrics like cost reduction, forecasting accuracy, decision speed, providing a baseline for future research. Finally, this paper identifies the challenges to adopting DFL in power systems and presents future research directions, offering researchers a roadmap to advance DFL beyond theoretical analysis into power grid-tailored models.