Improving costs and scale reflect looming opportunities

This paper describes the connected photovoltaic (PV) power generation system's grid overvoltage protection function and summarizes the occurrence of the output power loss due to the grid voltage rise. Power injection from the PV system will raise the voltage at the power distribution line. A power conditioning subsystem (PCS) needs to regulate its output if the voltage becomes higher than the upper limit in order to avoid the overvoltage at the power grid. Thus, a PV system cannot generate electricity under the high grid voltage. There are 553 residential PV systems installed in Ota, Japan, for the demonstration research project of clustered PV systems. Measurement data of these 2.1-MW grid-connected PV systems are used for the analysis. Only the limited number of PV systems experienced a significant amount of output energy loss due to the high grid voltage in a particular day, whereas the other system's outputs also raise the grid voltage. The causes of this maldistribution of the output energy loss are the difference of the line impedance, the difference of the starting voltage of the PCS's grid overvoltage protection function, and the imbalance of the load in single-phase three-wire power distribution systems. The present control of the PCS successfully avoids the overvoltage on the grid but cannot share the loss.

The Japanese earthquake and tsunami of 11 March 2011 spurred numerous discussions and debates regarding the national energy policy of Japan. One of the outcomes of this process was a new feed-in tariff (FIT) program to promote renewable energy, including photovoltaic (PV) systems, that went into effect on 1 July 2012. This new incentive program is expected to completely change the PV market in Japan, which has historically been dominated by residential PV applications. The FIT will drive not only the residential sector but also the nonresidential sector, including the new entrants to the Japanese power generation business: megawatt-scale PV power plants. Accelerating the introduction of renewable energy is important not only for diversifying Japan's sources of energy (for energy security) and combating global warming (for the environment) but also for the development of green industries (for the economy). It is essential that Japan create an environment conducive to the expansion of renewable energy in Japan by identifying the appropriate mix of regulatory measures (e.g., FITs), public support, and private-sector voluntary efforts best suited to each energy source. This article describes the current status of the PV market in Japan; the new FIT program and its impacts on the PV market; and further institutional, technical, and R&D challenges for PV dissemination.

Load forecasting is an essential task in the operation management of a power system. Electric power companies utilize short-term load forecasting (STLF) technology to make reasonable power generation plans. A forecasting model with low prediction errors helps reduce operating costs and risks for the operators. In recent years, machine learning has become one of the most popular technologies for load forecasting. In this paper, a two-stage STLF model based on long short-term memory (LSTM) and multilayer perceptron (MLP), which improves the forecasting accuracy over the entire time horizon, is proposed. In the first stage, a sequence-to-sequence (seq2seq) architecture, which can handle a multi-sequence of input to extract more features of historical data than that of single sequence, is used to make multistep predictions. In the second stage, the MLP is used for residual modification by perceiving other information that the LSTM cannot. To construct the model, we collected the electrical load, calendar, and meteorological records of Kanto region in Japan for four years. Unlike other LSTM-based hybrid architectures, the proposed model uses two independent neural networks instead of making the neural network deeper by concatenating a series of LSTM cells and convolutional neural networks (CNNs). Therefore, the proposed model is easy to be trained and more interpretable. The seq2seq module performs well in the first few hours of the predictions. The MLP inherits the advantage of the seq2seq module and improves the results by feeding artificially selected features both from historical data and information of the target day. Compared to the LSTM-AM model and single MLP model, the mean absolute percentage error (MAPE) of the proposed model decreases from 2.82% and 2.65% to 2%, respectively. The results demonstrate that the MLP helps improve the prediction accuracy of seq2seq module and the proposed model achieves better performance than other popular models. In addition, this paper also reveals the reason why the MLP achieves the improvement.

Large-scale penetration of photovoltaic (PV) power generators and storage batteries is expected in recently constructed power systems. For the realization of smart energy management, we need to make an appropriate day-ahead schedule of power generation and battery charge cycles based on the prediction of demand and PV power generation, which inevitably involves nontrivial prediction errors. With this background, a novel framework is proposed to maintain the balance among the total amounts of power generation, demand, and battery charging power with explicit consideration of the prediction uncertainty, assuming that consumer storage batteries are not directly controllable by a supplier. The proposed framework consists of the following three steps: 1) the day-ahead scheduling of the total amount of generation power and battery charging power; 2) the day-ahead scheduling of utility energy consumption requests to individual consumers, which aim to regulate battery charging cycles on the consumer side; and 3) the incentive-based management of the entire power system on the day of interest. In this paper, we especially focus on the day-ahead scheduling problems in steps 1) and 2), and show that they can be analyzed in a manner originating from spatiotemporal aggregation. Finally, we demonstrate the validity of the proposed framework through numerical verification of the power system management.