Increasing renewable energy (RE) penetration is indispensable for achieving a carbon-neutral society. In particular, a methodology to ensure profitability through the electricity market is a crucial technology for RE producers with unstable output, such as wind power, and a decision-making framework for bidding based on the uncertainty of generation and market prices is essential. This study proposes a day-ahead bidding scheduling method for RE generation targeting two Japanese markets. The proposed method includes functionalities to select appropriate bidding for markets from the viewpoint of maximizing expected revenues calculated based on probabilistic forecast results considering the uncertainty of RE generation and market transaction prices. The usefulness of the proposed method was evaluated through a numerical experiment using real-world historical data.

Recently, a utilization scheme of distributed energy resources (DERs) is discussed as a countermeasure for mitigation of congestion in distribution networks. This paper describes the technological trends in Japan and overseas based on pilot projects. Additionally, the topics about DERs utilization verified in research papers are surveyed and summarized.

The widespread introduction of functionally-smart inverters will be an indispensable factor for the large-scale penetration of distributed energy resources (DERs) via the power system. On the other hand, further smartization based on the data-centric operation of smart inverters (S-INVs) is required to cost-effectively achieve the same level of power system operational performance as before under circumstances where the spatio-temporal behavior of power flow is becoming significantly complex due to the penetration of DERs. This review provides an overview of current ambitious efforts toward smartization of operational management of DER inverters, clarifies the expected contribution of machine learning technology to the smart operation of DER inverters, and attempts to identify the issues currently open and areas where research is expected to be promoted in the future.

The amount of power generated by wind farms (WFs) varies depending on the wind conditions, and its use as the main power source within a power system is thus uncertain. This study proposes a balancing group (BG) scheme that combines the power from WFs with an existing variable-speed pumped-storage hydro generator (PSHG), which possesses a large capacity that can compensate for the variability in WF output. To maximize revenue, power suppliers need to schedule and operate the BG by considering the uncertainty in WF output and the PSHG's operational constraints. In the proposed method, probability density predictions (PDPs) are used to estimate WF uncertainty, and the BG output is scheduled to maximize the expected revenue. The PSHG water-level transition is managed within an operational limit using WF output scenarios derived from PDP results and historical WF output trends. The effectiveness of the proposed scheme was assessed based on revenue for three weeks with characteristic wind conditions, and the results of numerical simulations conducted using several operational schemes were compared using real-world WF output data. The results indicate that higher revenues can be achieved using the proposed operational scheme, thus validating the usefulness of the proposed method.

Polymer electrolyte fuel cell cogeneration systems (PEFC-CGSs) provide hot water by utilizing exhaust heat produced in electricity generation process. The energy saving potential of PEFC-CGSs can be maximized by optimal operational plans, and most state-of-the-art approaches implement operational planning function (OPF) based on energy demand time-series prediction by using machine learning techniques. In general, prediction of demand time-series with small expected average errors is regarded as the most important point in obtaining appropriate operational plans; however, several recent studies have revealed that other complex factors such as the direction and timing of forecast errors greatly affect the quality of operational plans in some cases. Core ideas proposed in these previous studies are broadly classified into seven types. The purpose of this study is to characterize these OPFs from the two aspects: the output form of prediction model and prediction target variable, and to clarify "what kind of uncertainty should be focused on" and "how this uncertainty should be handled" in designing OPF. The seven kinds of OPFs were comprehensively evaluated via numerical simulations using real-world data. The results show the significance of OPF based on prediction of expected operational cost surface using multiple output prediction model. (C) 2021 The Authors. Published by Elsevier B.V.

To utilize renewable energy systems (RES) as main generators in power system, power system operators need to maintain reserve power to compensate the uncertain output of RES, such as wind farms (WF). This paper focuses on the balancing group (BG) composed of WF and a pumped-storage hydro generator (PSHG). In the BG operation, power system operators need to schedule the operation of PSHG to minimize the imbalance risk caused by the prediction error of WF output while considering the system specific constraints of PSHG, e.g. the mechanical constraint on pump-up/down switching. This paper proposes a method to schedule the PSHG operation using the interval prediction results of the WF output for alleviating the imbalance risk the proposed scheduling scheme is formulated based on the linear programming. To verify the effectiveness of the proposed method, the amount of imbalance and the net generation of the BG were evaluated based on numerical experiments under two types of situations with different characteristics in WF output. The experimental results reveal the effectiveness of the proposed method.

This study evaluates the operational potential of pumped-storage hydro generators (PSHGs) as a regulating power source for the stable use of wind power generation. In recent years, the concept of a balancing group (BG), consisting of multiple generators, has been considered to alleviate the imbalance, defined as a discrepancy between planned and actual power supply values. Combining unstable power sources, such as wind power with controllable sources such as PSHG, is expected to reduce the imbalance in BG. For the effective utilization of PSHG, power system operators need to determine a schedule of operation considering the uncertainty in the wind farm (WF) output on the preceding day. In this study, the operation scheduling of PSHG, based on the interval prediction results of wind power generation, is discussed and the potential contribution of PSHG operation is assessed from the perspective of the imbalance reduction within this framework. The operational potential is evaluated from the viewpoint of the types of PSHG, that is, fixed-speed and variable-speed PSHGs, by the annual operational simulation using real-world WF output data. The results show that the appropriate operational scheduling of a single variable-speed PSHG can contribute to a reduction in imbalance comparable to the reduction achieved by two fixed-speed PSHGs.

This study proposes a novel operational planning method for polymer electrolyte fuel cell cogeneration systems (PEFC-CGSs). PEFC-CGSs provide hot water by utilizing waste heat produced in the electricity generation process, and hot water is stored in an attached tank. Generating and storing hot water based on an optimal operational plan according to household demand leads to further energy saving; therefore, operational planning methods based on household demand prediction have received significant attention. However, the improvement in the demand prediction accuracy does not necessarily lead to efficient PEFC-CGS operation in terms of operational costs; in other words, the accuracy in the demand prediction does not directly indicate the resulting operational efficiency. In this study, the authors propose a novel approach based on a surrogate model for deriving an appropriate plan that minimizes the expected operational costs among the operational plan candidates. In the proposed scheme, the error between expected and actual operational costs explicitly represents the relevance of the operational plan, so that the optimal operational plan can be selected directly from the perspective of the resulting operational efficiency. The practicality of the proposed approach is evaluated with the existing demand prediction-based approach via numerical simulations using real-world measurements of multiple customers in Japan. The proposed method reveals 30% reduction of the excessive operational costs by avoiding the inefficient operation of the auxiliary gas-heater in the experiments and will further enhance the value of introducing highly efficient residential fuel cell system that contributes to a low-carbon society.

Forecasting the electricity spot market price (i.e., the price of electricity traded for the next day) is important in planning efficient power generation and demand schedules. This research aims to discuss the following two points in the forecasting task: an explicit scheme for estimating high and low-price situations and an effective weighted integration of given multiple forecast products depending on the estimated situations. The authors propose an adaptive method of combining market price forecasts (ADAPTIVE) to utilize several types of given forecast results efficiently and improve forecasting accuracy. The method aims to capture the features of the component models' results considering high- and low-prices situations. To evaluate the effectiveness of the proposed ADAPTIVE, numerical experiments were conducted on electricity spot market prices in the Japan Electric Power Exchange. The results indicate that the proposed method is a promising way to improve the accuracy of the conventional method of combining forecasts.

This study proposes a nonintrusive appliance load monitoring framework for estimation of the power consumptions of individual residential appliances by using aggregated total consumption based on Gaussian-softmax restricted Boltzmann machine with temporal reservoir. The proposed method is expected to estimate the hidden states of the appliances well by coding the current situation in the nonlinear temporal dynamics of the power consumption in the reservoir units so as to estimate the appliance-wise consumptions well. The accuracy of the proposed framework is evaluated by using real-world power consumption data sets.

© 2019 IEEE. This study proposes a hot water demand prediction method for operational planning of polymer electrolyte fuel cell cogeneration systems (PEFC-CGSs). PEFC-CGSs provide hot water by utilizing waste heat produced in the electricity generation process. An optimal operational plan according to household demand leads to further energy saving. Therefore, operational planning methods based on household demand prediction have received intense focus. In particular, the prediction of the amount of hot water demand is important for efficient operation. The authors have attempted to improve the hot water prediction method based on multivariate random forest (MRF), which uses the average of many decision trees' outputs as the prediction result. However, some experimental results show that a prediction strategy based on averaging the outputs of decision trees does not always lead to the best solution. In this study, the authors propose a novel prediction method utilizing the quantile of the estimation results derived in MRF. By setting the appropriate quantile, we can evade the demand underestimation, which has a higher negative impact on operational efficiency than overestimation. The usefulness of the proposed approach is evaluated via numerical simulations using real-world demand data.

© 2019 IEEE. This study proposes an operation planning method for polymer electrolyte fuel cell cogeneration systems (PEFC-CGSs). PEFC-CGSs provide hot water by utilizing waste heat produced in the electricity generation process. An appropriate operation plan according to household demand will lead to further energy saving. The authors have discussed a framework for estimating expected operation costs of individual operation plans to select an appropriate candidate that minimizes the cost. In this scheme, the expected cost under each corresponding operation was independently estimated. This approach can be expected to provide accurate prediction of costs under individual operations. However, one open issue remains; it may not work well in estimating relative goodness compared to other operations. In this study, the authors propose a new cost estimation approach, focusing on relative appropriateness in candidate plans based on a multiple output prediction. The usefulness of the proposed approach is proved via a numerical simulation.

We consider enhancing cyber security in voltage regulation of distribution systems in the presence of malicious attacks. Due to the increase in distributed generation, the regulation of voltage has become more complicated, requiring more sensor data to be used for control. Recently, we have developed an attack detection algorithm to find data falsification attacks on voltage measurements transmitted by sectionizing switches in the feeders to a centralized controller. In this paper, the security level of the system is further improved by introducing a controller that is capable of operating the regulation in the presence of attacks by utilizing the detection results. In particular, it identifies abnormal behavior in the sensor data and ignores measurements taking extreme values among those received. Through detailed simulation studies on a small scale distribution system, we show the effectiveness of the proposed control and analyze the relation between the number of attacks and the detectability of the attacks.

© ECOS 2019 - Proceedings of the 32nd International Conference on Efficiency, Cost, Optimization, Simulation and Environmental Impact of Energy Systems. All rights reserved. For a daily-basis scheduling of an energy system, energy management system often enough to use not a global optimal scheduling but a near optimal scheduling. The article proposes an online scheduling framework without online optimization. The framework is built from two encoder-decoder architectures to extract features of time series; a multi-layer long short-term memory regression model for multi-step time-series forecasting, and multi-class and single-label classification model for on/off scheduling of a device. The models are estimated at offline, and return scheduling from historical time series as input, at online. We evaluate the accuracy of scheduling from the viewpoint of Kullback-Leibler divergence which measures the dissimilarity between two probability distributions. Through the numerical experiments, we demonstrate the usefulness of the proposed framework.

Home energy management systems (HEMSs) are the system to manage the energy usage in houses. The use of HEMSs, and especially those which are capable of automatically controlling home energy appliances such as air-conditioners (ACs), is expected to manage energy utilized in domestic field effectively. In the present study, we focused on automatic AC operation by HEMS with the combined goal of improving thermal comfort while reducing electricity costs. In general, the room temperature and electricity consumption of an AC are highly dependent on the characteristics of the installation environment. so that the derivation of an appropriate AC operation plan is generally a difficult task. To tackle this problem, an energy management method to provide AC operation plan tailor-made for the target AC installation environmental by learning the characteristics of the installation environment (CIE) from the historical operation result data is proposed. The efficacy of the proposed method is verified via numerical and real-world experiments. (C) 2018 Elsevier B.V. All rights reserved.

With growing attention to sustainability and recognition of the impact of global warming problems, energy supply and consumption have become critically important. This paper presents the construction of a modeling platform accommodating cooperative energy management systems (EMSs), which virtually produces the model of a smart city with a distribution network (DN) by using a wide range of data obtained from the real world. The platform involves models of various EMSs, governing the operation of a power system or controlling consumer-installed devices, and simulating the power flow, electrical losses, and voltage in the DN. In addition, indices measuring the sustainability of the model city, such as CO2 emission, are estimated from scenarios, for example, photovoltaic system installation, electric vehicle penetration, etc. The results can be visually displayed and the platform is highly versatile and applicable to various types of issues associated with smart cities. Two case studies are presented in detail.

The introduction of photovoltaic power systems is being significantly promoted. This paper proposes the implementation of a distributed energy management framework linking demand-side management systems and supply-side management system under the given time-of-use pricing program for efficient utilization of photovoltaic power outputs; each system implements a consistent management flow composed of forecasting, operation planning, and control steps. In our framework, demand-side systems distributed in the electric distribution network manage individual energy consumption to reduce the residential operating cost by utilizing the residential photovoltaic power system and controllable energy appliances so as not to inconvenience residents. On the other hand, the supply-side system utilizes photovoltaic power maximally while maintaining the quality of electric power. The effectiveness of the proposed framework is evaluated on the basis of an actual Japanese distribution network simulation model from both the supply-side and demand-side viewpoints.

© 2017 IEEE. A low-voltage regulator (LVR) is investigated as a countermeasure for voltage violation caused by the integration of photovoltaic (PV) systems in distribution systems. To achieve appropriate voltage control by the LVR, the appropriate control parameters need to be determined. We have proposed schemes to determine the appropriate control parameters using PV forecast profiles. In a previous work, the control parameters were determined on the basis of one-scenario PV forecast, which derives one forecast value in each time step. However, voltage violation could not be avoided by this method in a time step with a large forecast error. In this paper, we propose a method for determining the voltage control parameters of the LVR based on the forecast interval of the PV outputs to quantitatively consider the PV forecast error. The numerical simulation results show that the proposed method can reduce the amount of voltage deviation compared with the conventional method.

Under the large penetration of photovoltaic systems, the voltage control in a distribution system will contain a difficulty due to the fluctuated reverse power flow, so that the appropriate deployment of voltage regulators is required to suppress the voltage violation. However, derivation of an appropriate long-term deployment plan of voltage regulators is a difficult task because of the uncertainties in spatial-temporal penetration scenario of the photovoltaics systems connected to the distribution system. This paper focuses on a simple approach for deploying step voltage regulators considering uncertainties in penetration transition of photovoltaic systems in order to discuss the necessity of long-term plan. The impact of long-term deployment plans is evaluated by numerical experiments.

© 2018 University of Minho. All rights reserved. There is a high demand that aggregating Home Energy Management System (HEMS) as demand response service. Aggregation scheme is divided into the centralized and decentralized approach. On the one hand, in a centralized fashion, a controller manages all device. On the other hand, in a decentralized fashion, a local controller manages own devices and exchanges information for achieving the global optimum. Our previous work has proposed decentralized HEMS aggregation with Alternating Direction Method of Multipliers (ADMM) to address scalability and privacy issue. The main idea of this method is that we decompose the large-scale aggregated scheduling problem into individual HEMS scheduling problem by introducing local upper limit which stands for the maximum purchasable electricity from an electrical grid. This article reports that we extend the previous decentralized HEM method to asynchronous fashion in order to improve convergence time efficiency from the practical implementation perspective. We propose the waiting ratio which represents the minimum HEMS number for updating the next local upper limit. We also evaluate the algorithm processing time with changing waiting ratio and the number of households. As a result, it is confirmed that though the result shows 53% of acceleration at the most accompanied by barely 2% increase cost, the asynchronous process does not always accelerate the processing time. It may be reasonable to suppose that the degree of acceleration is decreased, as the degree of asynchrony increases, because iteration number until fulfilling converging criteria is increased.

As demand-side energy management system (EMS) has not only a function to maximize owner's utility by automatic control but also a potential capability to respond a demand activation, demand-side EMS is expected to fulfill the key role for intelligent control of community energy management. To achieve the stable control result of EMS against the uncertain environment, it needs approach from both forecast and operation sides. The objective of this study is to construct a control scheme for a home EMS from a practicable implementation perspective and to evaluate the control performance of the scheme proposed. The scheme is composed of operational planning with receding horizon and energy demand forecasting. The goal of the scheme is to maximize householder's utility only by retrofitting the home EMS to a residential household. Throughout numerical results, it is revealed that the better control performance comes from the conservative operational strategy derived from the multiple scenarios.

Visualizations that depict a consumer's utilization of domestic appliances aid them in effectively thinking about energy conservation. Energy disaggregation aims to break down the total power consumed into the amount consumed by each individual appliance through the use of a single smart meter. This estimation technique provides detailed information about energy consumption and is cheaper than techniques that directly measure each appliance. In order to improve the accuracy of disaggregation, we utilize the consumer feedback information without placing any special burden on the consumer. We propose a semi-supervised shift-invariant weighted non-negative matrix factorization method with auxiliary feedback that records the on or off status of each individual appliance. The experimental results obtained by applying our proposed method to household datasets show that our proposed method and the associated auxiliary information generated contribute to the improvement of the disaggregation accuracy.

Wind power prediction is necessary for stable operation of a power grid under the introduction of significant wind power generation. Wind power prediction approaches based on the results of numerical weather prediction (NWP) have been developed successfully in recent years. However, the high dimensionality of NWP results can be a major obstacle when training models. This paper proposes a feature extraction scheme based on convolutional neural networks to compress high-dimensional NWP results by deriving critically important low-dimensional information for wind power prediction. The experimental results show that the proposed feature extractor can contribute to the improvement of wind power prediction accuracy.

This paper proposes a voltage control method during service restoration in the distribution networks with photovoltaic (PV) generator systems. In the current distribution automation system (DAS) process in Japan, voltage dips and surges occur during service restoration because PVs are disconnected simultaneously after a fault and subsequently reconnected after service restoration. However, in the current DAS, voltage regulators such as an on-load tap changer (OLTC) and step voltage regulators (SVRs) are not controlled during service restoration. The proposed DAS estimates the voltage in a distribution network during service restoration, and it controls the tap position of OLTC and/or SVRs according to the predicted voltage. The numerical simulation results using a real-world distribution system model on a real map and PV output profiles derived by actual square kilometer solar radiation data will be shown. Those results indicate that the proposed DAS prevents voltage deviation that occurs as long as the current DAS is used. The results also show high spatial resolution PV output data are needed to prevent voltage deviation absolutely. (C) 2016 American Society of Civil Engineers.

An increased penetration of wind energy into the power system will lead to instability of local voltage and global frequency in Japan. Power flow simulation is a powerful tool for understanding the electrical behavior in the power system caused by wind energy; however, plausible and various wind power profiles are required to perform a meaningful simulation to evaluate the effect of wind generators. This paper proposes a procedure of generating synthetic wind power profiles that involve the plausible short-term fluctuation for a power flow simulation based on the spatial kriging method and the bootstrap method. (C) 2016 American Society of Civil Engineers.

© 2017 IMEKO As communication standard, such as SEP 2.0 and ECHONET Lite, for digital communication of home appliances has been established, the research of networked control system is progressing in the domain of residential house. Instead of building new power station to fulfil new electricity requirement, such as electric vehicle charging, the concept, that an aggregator coordinates massive amount of home energy management system (HEMS) to supply negawatt power, is proposed. The aggregator requires to manipulate electricity consumption through massive amount of HMES as networked control system. As fundamental study for control algorithm of distributed architecture, this paper constructed one of the networked supervisory control system formulated as stochastic model predictive control scheme, and analysed primary energy consumption sensitivity against update interval of energy demand forecasting and operational strategy in addition to changing number of scenarios. The mainly results are concluded that: as the energy saving ratio in the case of 6 hours update interval has enough high point, which is 9.5% in comparison with the case of 24 hours update interval, the stochastic model predictive control scheme proposed in the case using six scenarios can downsample until 6 hours update interval.

The world is more aware of the need for saving energy because of increasing world energy consumption and environmental problems. To promote saving energy in the domestic field, the use of home energy management systems (HEMSs) is rapidly spreading. The HEMS which has automatically controlling function can control domestic electrical appliances including air-conditioners (ACs). In this research, we focus on AC operation plans to improve thermal comfort and reduce electricity costs for residents. However, AC control planning is generally a difficult task because the operation results greatly depend on the environmental characteristics in which the HEMS is installed. To solve this problem, we proposed an AC planning method that accounts for environmental characteristics and uncertainty in prediction by using historical data.

The introduction of wind power generation has been promoted in Japan. However, wind power is an unstable power source because its output varies according to the weather. Particularly, sudden changes in output, which could adversely affect the power system, are called ramps and may cause serious problems in the power system. In this paper, the authors discuss the ramp event forecast by using classifiers. A serious issue in this setup is that classification based forecast tends to be inaccurate since the occurrence of such a ramp is relatively rare. This problem is called the class imbalance problem in the machine learning field. To overcome the class imbalance problem in ramp forecast, several data sampling approaches are implemented. The effectiveness of these sampling approaches is experimentally evaluated by using a real-world wind power generation dataset. The results show that the implemented approaches drastically improved the forecast accuracy.

In this paper, we consider the impact of cyber attacks on voltage regulation in distribution systems when a number of photovoltaic (PV) systems are connected. We employ a centralized control scheme that utilizes voltage measurements from sectionizing switches equipped with sensors. It is demonstrated that if measurements are falsified by an attacker, voltage violation can occur in the system. However, by equipping the control with a detection algorithm, we verify that the damage can be limited especially when the number of attacked sensors is small through theoretical analysis and simulation case studies. In addition, studies are made on attacks which attempt to reduce the output power at PV systems equipped with overvoltage protection functions. Further discussion is provided on how to enhance the security level of the proposed algorithm.

Photovoltaic (PV) systems have recently attracted considerable attention in the context of environmental problems, antinuclear power movements, and energy problems. Therefore, the large-scale introduction of PV systems is expected in the near future. But the connection of many PV systems to the power system leads to problems. For example, the system voltage often drifts from the norm when reverse power flows increase. Accordingly, optimal system operation is required in order to make maximum use of solar energy to the maximum by installing energy buffers such as storage batteries. In particular, in forecast information, knowledge of the reliability as well as the predicted solar irradiance is essential for effective operation. In this paper, we propose a way of estimating the prediction interval of solar irradiance as an index of reliability by using just-in-time modeling. We consider the accuracy of the prediction interval under many conditions and derive a high-precision estimation method. In addition, we also discuss about the future subjects of study. (C) 2016 Wiley Periodicals, Inc. Published online in Wiley Online Library (wileyonlinelibrary.com).

© 2016 University of Ljubljana. The degree of operational freedom of home appliances is increasing with introducing controllable demand-side appliances by using Home Energy Management System with some communication standard. This situation increases the demands on methodology achieving customer utility maximization, which is able to plan operational strategies of a variety of appliances under various exogenous variables such as energy demand, outlet temperature, and power output of Photovoltaic power generator. Energy demand and photovoltaic output have some uncertainty or variation in the parameter. The objective of this study is to analyze a solution behavior of an operational planning problem adapting Mixed Integer Second-Order Cone Programming extended from traditional Mixed Integer Linear Programming model. The objective function of the mathematical optimization involves mean and its variance by extracting naturally Markowitz model. Above proposed model is intended to sophisticate robust operational planning method, and is also evaluated on the EMS simulation platform that we are developing. As a result, it is found that involving variance to objective function affects to computational time meaning computational load rather than robustness of planned schedule. Additionally, using several scenario results in more robust schedule than using plausible one scenario.

The large-scale introduction of renewable energy resources will cause instability in the power supply. Residential energy management systems will be even more important in the near future. An important function of such systems is visualization of appliance-wise energy consumption; residents will be able to consciously avoid unnecessary consumption behavior. However, visualization requires sensors to measure appliance-wise energy consumption and is generally a costly task. In this paper, an unsupervised method for non-intrusive appliance load monitoring based on a semi-binary non-negative matrix factorization model is proposed. This framework utilizes the total power consumption patterns measured at the circuit breaker panel in a house, and derives disaggregated appliance-wise energy consumption. In the proposed approach, the energy consumption of individual appliances is estimated by considering the appliance-specific variances based on an aggregated energy consumption data set. The authors implement the proposed method and evaluate disaggregation accuracy using real world data sets.

Maintaining voltage levels in distribution networks (DNs) with photovoltaic systems (PVs) is a complicated task for conventional voltage control schemes that only use a load ratio control transformer (LRT) and step voltage regulators (SVRs), because of local voltage fluctuation caused by the introduction of PVs. This paper proposes a coordinated voltage control scheme consisting of multiple kinds of load tap changers (LTCs). The proposed scheme determines the location at which to introduce low-voltage regulators (LVRs) and provides control parameters for the LTCs by considering the behavior of the other LTCs. We carried out numerical simulations using a DN model including PVs to verify the validity of the proposed scheme. The results specify the characteristics of the voltage deviation that cannot be prevented by a conventional voltage control scheme, and the proposed scheme significantly reduces such local voltage deviation.

This paper proposes an optimal predictive control of 0.75 kW PEM fuel-cell cogeneration with home appliances. This paper also models fuel cell system for design and operation evaluation of building equipment based on actual measurement of residential fuel cell system on sale. As one application of constructed model and proposed control method, this paper discusses concerning home EMS for efficient PV utilization.

The subject of this study is to propose a power interchange system in a collective housing with residential solid oxide fuel cells (SOFCs) and a robust operation planning method for integrated SOFCs against uncertain energy demand forecast. In this method, the future operation plan for multiple SOFCs is optimized and determined to minimize the expected total primary energy consumption in the collective housing considering uncertainty in demand forecast. If the forecast energy demand includes forecast errors, the result of SOFCs operation will corrupt from the viewpoint of the primary energy consumption. Thus, the output decision problem for SOFCs is formulated by considering the corruption caused by forecast errors, so that the decided SOFC outputs have the robustness against uncertainty in demand forecast. The validity of the proposed method is examined based on numerical simulations from the perspective of the robustness.

Unexpected fluctuation of wind power output will become serious problems from the viewpoint of stable supply for an electricity grid. Operating a battery system installed in the grid for mitigating the short-term fluctuation is one of the new approaches for grid stabilization. In this paper, we propose a method of generating synthetic wind power profiles with high temporal resolution for power flow simulation which aims to estimate the impact of wind power fluctuation and specify the required battery system. We numerically show the plausibility of the synthetic wind power profiles from the viewpoints of statistical properties. (C) 2016 The Authors. Published by Elsevier Ltd.

Compensating the voltage within the appropriate range becomes difficult when a large number of photovoltaic (PV) systems are installed. As a solution to this problem, the installation of a low-voltage regulator (LVR) has been studied. In this paper, we propose a method for rapidly and accurately determining the line drop compensator method (LDC) parameters as a part of a voltage management scheme, which consists of prediction, operation, and control. In the proposed method, candidates of the appropriate LDC parameters are selected with low computational cost by using classifiers that learns the relation between power series data and the properness of LDC parameters. We performed numerical simulations to evaluate the validity from the viewpoints of computational time and classification accuracy for determination of the LDC parameters, and verified the voltage control performance of the proposed method.

PV system recently attracts much attention on the back of environmental problems, antinuclear power movements and energy problems. Therefore, a large scale introduction of PV system is expected in the near future. On the other hand, a lot of PV systems connected to the power system bring on some problems. For example, a system voltage often drifts from the norm when the reverse power flow increase. Accordingly, it is necessary to perform an optimal system operation in order to utilize a solar energy to the maximum by installing energy buffers, e.g. storage batteries. Especially, the forecast information i.e., the reliability as well as predicted solar irradiance is essential for the effective operation. In this paper, we propose a way to estimate the prediction interval of the solar irradiance as an index of reliability by using Just-In-Time Modeling (JIT Modeling). We consider the accuracy of the prediction interval under many conditions and derive the high-precision estimation method. In addition, we also talk about the future outlook of this study.

This paper studied power and hour capacity requirement for energy storage by two approaches. First one is evaluations without constrained condition of power and hour capacity with 80 sets of maximum ramp rates and time windows. And second one is evaluations with constrained condition of power and hour capacity for 2 grid codes (less than 0.1pu change of wind farm (WF) capacity in 20min time window, showing 0.1pu/20min in this paper, and 0.3pu/360tnin) to see details. The first evaluation in case of 3 sigma resulted 1.0 pu power capacity (charge plus discharge capacity for all result) and 1.3 to 2.2 pu-h hour capacity are required to satisfy maximum ramp rate in time window of 0.1pu/20min. And 1.2 to 1.3 pu and 13 to 18 pu-h are required for 0.3pu/360min. And from the second evaluation in case of 3 sigma, 0.7 to 0.8 pu power capacity and 1 to 2 pu-h for 0.1pu/20min and 1.0 pu and 10 to 15 pu-h are required for 0.3pu/360min. This paper provides relationship between maximum ramp rate in time windows and power and hour capacity requirement. the results help to answer questions whether power and hour capacity requirement satisfy grid codes with simple operating method of an energy storage system.

This paper proposes a new method for determining the LDC (Load Drop Compensator) parameters of LRT (Load Ratio Transformer) and SVR (Step Voltage Regulator). Since the voltage control effect of LRT and SVR depends on the LDC parameter, the proper parameters should be chosen in order to control voltage within proper range. When the photovoltaic generation system is installed in the distribution system, voltage control becomes difficult and the optimal parameters in the daytime differ from those in the nighttime. In the proposed method, the parameters are reset every hour. The parameters are determined by choosing from the database including all the optimal parameters for each time period in the past. The database also includes the PV output profiles and load profiles in the past, and the proposed method searches the similar time periods in the database to the next one hour period by comparing forecasted PV output profiles with that in the database. And the common parameters among the similar time periods are chosen for the next one hour. The simulation results will be shown to evaluate the performance of proposed method.

Demand-side electricity saving is an important factor in the reduction of the installed capacity of power-supply facilities. In order to save electricity automatically while maintaining comfort levels, home energy management systems (HEMS) have attracted attention. These systems can control residential energy equipment cooperatively to reduce electricity consumption while considering benefits to consumers. Although many researchers have evaluated HEMS, no one has conducted a study which considers the control of various types of residential energy equipment in real time along with the uncertainties of energy demands. This paper proposes a single HEMS method which connects prediction, operational planning and control steps and enables the evaluation of operational planning methods of HEMS connected with many kinds of residential energy equipment currently in use in Japan while considering the uncertainties. The purpose of this study is to evaluate the economic potential of residential energy systems based on the proposed method with the uncertainties of energy demands and photovoltaic (PV) output under time-of-use prices. The results allowed us to establish a framework to quantitatively evaluate the operational planning methods of HEMS with the uncertainties of energy demands and PV output. In addition, the usability of the proposed method was confirmed by comparing the operational costs to those of a reference method.

Residential houses are in the process of introducing power generators such as photovoltaic (PV) power generators and fuel cell cogeneration systems. Under current laws in Japan, surplus electricity from a residential PV system can be sold by feeding it back into the electrical grid. However, when a lot of neighboring power generators make and feed back electricity at the same time, there is an issue of an upper voltage violation of the provisions of the laws and regulations relating to the Electricity Business Act in the distribution system of the electrical grid. The issue has been solved by stopping power generators when they reach an upper voltage limit, 107V. Another solution is to acknowledge demand response signal from electrical grid operator to store electricity in residential batteries using Home Energy Management System (HEMS). The research question is that how to collaborate HEMS and Grid Energy Management System (GEMS). This paper developed an evaluation framework for the cooperative behavior between HEMS and GEMS. Using this evaluation framework, this paper demonstrated the optimal operational strategy of HEMS including PV in the case that HEMS is informed voltage profile from GEMS, and also assessed amount of PV suppression quantitatively in residential sector in practical aspect.

This paper presents a distribution automation system (DAS) for service restoration in the distribution network with photovoltaic (PV) generator systems, which are disconnected simultaneously after a fault and subsequently reconnected after service restoration. Because the reverse power flow of PVs affects voltage in the distribution system, voltage dips and surges occur during the service restoration. However, current DAS do not control voltage regulators such as an on-load tap changer (OLTC) and step voltage regulators (SVRs) during the service restoration. The proposed DAS estimates the voltage in a distribution network considering the simultaneous disconnection of PVs by performing power flow calculations, and it controls the tap position of OLTC and/or SVRs according to the predicted voltage deviation. The voltage after the disconnection of PVs is calculated by estimating the PV output utilizing square kilometer solar radiation data calculated using satellite image data in real time.

This paper presents a new method for determining the load drop compensator (LDC) parameters of On-load tap changer (OLTC) and step voltage regulators (SVRs) in the distribution system. Introduction of photovoltaic generation systems (PVs) into a distribution system makes determining the appropriate LDC parameters difficult, which may cause voltage deviation because reverse power flow from PVs affects voltage. Especially in the distribution system in which multiple SVRs are installed, determining the appropriate combination of LDC parameters becomes difficult. In the proposed method, the parameters are reset every hour by choosing one LDC parameter from a database including feasible parameters in the past. To create the database, the proposed method hourly enumerates the feasible parameters utilizing data acquired by SCADA. To evaluate the performance of the proposed method, the amount of voltage deviation of the proposed method will be compared with that of a centralized control method.

Output of the photovoltaic (PV) system drastically fluctuates depending on weather conditions. Therefore users of PV systems should manage their energy usage with forecast information of the PV output. Due to the reasons above, we have developed an estimation method of X% prediction interval of solar irradiance. We have achieved a high-accuracy prediction interval for any prediction coefficient X%. However, we found that there are some errors in a case where the prediction coefficient is high. In this paper, we improve the estimation algorithm of the prediction interval.

In this paper, we consider the impact of cyber attacks on voltage regulation in distribution systems. We employ a centralized control scheme which utilizes voltage measurements from sectionizing switches equipped with sensors for connecting distributed generation. Through detailed case studies by simulations, it is demonstrated that if measurements are falsified by an attacker, voltage violation can occur in the system. However, by equipping the control with a detection algorithm, we verify that the damage can be limited especially when the number of attacked sensors is small. Further discussion is provided on how to enhance the security level of the proposed algorithm.

Total electricity consumption in Japan increased rapidly and the power consumption per household is also continuing to increase. The framework of demand response (DR) to promote the reduction of electricity consumption in the household sector by regulating the price of the electricity will be introduced in the future. In this situation, residents must operate their appliances so as not to affect much to their lifestyles while taking into account the power cost. A home energy management system (HEMS) will have an essential role to control appliances such as air conditioners (ACs), battery energy storage systems (BESSs), electric vehicles (EVs), and heat pump water heaters (HPWHs) and automatically match their operations to the behavior of a resident when the electricity price changes. In this study, a Bayesian network, a fundamental tool of machine learning, is adapted to an HEMS to learn the behavior of the resident and appropriate operations of controllable appliances.

Forecast of the energy demand is an important topic for the realization of effective energy management. In this paper, the authors focus on the K-nearest neighbor approach for forecast of the energy demand pattern, introduce an idea of distance metric learning to select appropriate K-nearest neighbors, and propose some learning frameworks for forecasting. The proposed frameworks are evaluated based on the real-world datasets from the viewpoint of selection accuracy of the true neighbors.

A complicated voltage fluctuation in distribution systems and a decline in power quality occur when a large number of photovoltaic (PV) systems are installed. In this paper, the installation of a low-voltage regulator is assumed, and a method for instantly and accurately determining the line drop compensator (LDC) parameters is proposed to perform efficient voltage management, which consists of prediction, operation, and control. In the proposed method, the computational cost to derive the LDC parameters can be reduced by learning the optimality of the parameters in a series of load demands and the PV output using multiple classifiers. We performed numerical simulations to verify the validity of the proposed method. From the results, the classification accuracy is found to improve by considering the majority vote of multiple classifiers. Additionally, the improvement in the voltage control performance is verified.

Highly accurate predictions of load demand and photovoltaic (PV) output have become possible in recent years because of improved measuring instruments and the increase of databases on load demand and PV output. The appropriate control parameters for actual power system operation can be determined by using these predictions. Although parameters determined by conventional methods are accurate, they may not be determined in time before the beginning of operation because extensive time is required for the calculations. In this paper, the support vector machine-a machine learning method that solves the two-class classification problem-is used to determine the line drop compensator (LDC) parameters instantly. To verify the validity of the proposed method, we carried out numerical simulations to determine the LDC parameters. From the simulated results, we found that the proposed method can instantly and accurately determine the LDC parameters.

In this paper, the authors propose an electric power interchange system in collective housing with fuel cells (FCs) and a determination method of operation plan for FCs in the collective housing. In the method, the operation plan for FCs is determined from an evaluation point, primary energy consumption of all houses in the collective housing, based on an enumeration method and particle swarm optimization (PSO) which is one of non-liner optimization methods. In order to examine the validity of the determination method, numerical simulations are carried out for the collective housing model, primary energy consumptions of the model are compared to those of a standard system (without FCs), and the reduction effects are evaluated taking uncertainty into consideration by using 10,000 demand patterns which are represented by 40 observational demand data based on the bootstrap method. In addition, the reduction effects by reducing introduced FCs by half are also evaluated from the point of primary energy consumptions, operating efficiency and running efficiency.

This paper verifies, for distribution systems with residential photovoltaic generators (PV), the distribution loss reduction effect on the loss minimum configuration selected from all possible configurations. Methods that determine the loss minimum configuration of distribution systems have been developed. However, loss reduction effects and guidelines for future PV deployment to effectively reduce distribution loss have not been discussed extensively, because of difficulties in enumerating possible configurations and in comparing distribution losses for each. We have previously proposed a method for selecting loss minimum configurations from among all possible configurations. In this paper, we use the proposed method to verify the loss reduction effect on loss minimum configurations of a distribution system composed of 468 switches for three PV penetration patterns. We also evaluate the loss minimum configuration by comparing its distribution loss with that of other configurations and provide a guideline for future PV penetration position and system operation. © 2013 IEEE.

Nonnegative Matrix Factorization (NMF) is broadly used as a mathematical tool for processing tasks of tabulated data. In this paper, an extension of NMF based on a generalized product rule, defined with a nonlinear one-parameter function and its inverse, is proposed. From a viewpoint of subspace methods, the extended NMF constructs flexible subspaces which plays an important role in classification tasks. Experimental results on benchmark datasets show that the proposed extension improves classification accuracies. © 2012 Springer-Verlag.

This paper proposes a computation method to determine a optimal planning and operation of residential Photovoltaic Power Generation systems (PV systems) and Battery Energy Storage Systems (BESS) based on the weather forecast to each customer. In this proposed method, the optimal planning and operation are decided by changing the amount of charge in the BESS in midnight when the electric rate is cheap. This proposed method can determine the optimal planning (determination of optimal rating capacities of PV and BESS) and the optimal operation (determination of three values of the amount of charge at midnight on sunny day, cloudy day and rainy day) so as to minimize the annual cost of the household. In order to check the effectiveness of the proposed method, the optimal planning and the operation are compared on the data of two years.

Considering recent trends in energy technology development, consumer's energy demand could be influenced by the renewable energy supply in any way. A simple extension of pattern sequence-based forecasting (PSF) enables us to predict demand curves based on the correlated bidimensional time-series by using co-occurrence patterns of energy supply and demand. However, prediction accuracy of PSF deeply depends on the clustering result, which is used for pattern matching. In this paper, a promising clustering method based on nonnegative tensor factorization is applied for this task and evaluated experimentally from the viewpoint of prediction accuracy.

In this paper, generalised statistical independence in statistical models for categorical distributions is proposed from the viewpoint of generalised multiplication characterised by a monotonically increasing function and its inverse function, and it is implemented in naive Bayes models. This paper also proposes an idea of their estimation method which directly uses empirical marginal distributions to retain simplicity of calculation. This method is interpreted as an optimisation of a rough approximation of the Bregman divergence so that it is expected to have a kind of robust property. Effectiveness of proposed models is shown by numerical experiments on some benchmark datasets. Copyright © 2012 Inderscience Enterprises Ltd.

The Bradley-Terry model is a statistical representation for one's preference or ranking data by using pairwise comparison results of items. For estimation of the model, several methods based on the sum of weighted Kullback-Leibler divergences have been proposed from various contexts. The purpose of this letter is to interpret an estimation mechanism of the Bradley-Terry model from the viewpoint of flatness, a fundamental notion used in information geometry. Based on this point of view, a new estimation method is proposed on a framework of the em algorithm. The proposed method is different in its objective function from that of conventional methods, especially in treating unobserved comparisons, and it is consistently interpreted in a probability simplex. An estimation method with weight adaptation is also proposed from a viewpoint of the sensitivity. Experimental results show that the proposed method works appropriately, and weight adaptation improves accuracy of the estimate.

Given a set of rating data for a set of items, determining preference levels of items is a matter of importance. Various probability models have been proposed to solve this task. One such model is the Plackett-Luce model, which parameterizes the preference level of each item by a real value. In this letter, the Plackett-Luce model is generalized to cope with grouped ranking observations such as movie or restaurant ratings. Since it is difficult to maximize the likelihood of the proposed model directly, a feasible approximation is derived, and the em algorithm is adopted to find the model parameter by maximizing the approximate likelihood which is easily evaluated. The proposed model is extended to a mixture model, and two applications are proposed. To show the effectiveness of the proposed model, numerical experiments with real-world data are carried out.

In this paper, generalized statistical independence is proposed from the viewpoint of generalized multiplication characterized by a monotonically increasing function and its inverse function, and it is implemented in naive Bayes models. This paper also proposes an idea of their estimation method which directly uses empirical marginal distributions to retain simplicity of calculation. Our method is interpreted as an optimization of a rough approximation of the Bregman divergence so that it is expected to have a kind of robust property. Effectiveness of our proposed models is shown by numerical experiments on some benchmark data sets.

To describe the relation between some values, arithmetic operations like multiplication and division are important and conventional tools. These arithmetic operations for probabilities are characterized by the KL divergence, and hence, they can be generalized by using the Bregman divergence instead of the KL divergence. With this idea, independence of random variables is modified by generalized product rule, and a joint probability model is proposed based on this modified weak independence. Effectiveness of weak independent models is shown by numerical experiments on toy examples, and discussed from a geometrical viewpoint.

The EM algorithm is a sophisticated method for estimating statistical models with hidden variables based on the Kullback-Leibler divergence. A natural extension of the Kullback-Leibler divergence is given by a class of Bregman divergences, which in general enjoy robustness to contamination data in statistical inference. In this paper, a modification of the EM algorithm based on the Bregman divergence is proposed for estimating finite mixture models. The proposed algorithm is geometrically interpreted as a sequence of projections induced from the Bregman divergence. Since a rigorous algorithm includes a nonlinear optimization procedure, two simplification methods for reducing computational difficulty are also discussed from a geometrical viewpoint. Numerical experiments on a toy problem are carried out to confirm appropriateness of the simplifications.

Modeling of a large joint probability table is problematic when its variables have a large number of categories. In such a case, a mixture of simpler probability tables could be a good model. And the estimation of such a large probability table frequently has another problem of data sparseness. When constructing mixture models with sparse data, EM estimators based on the beta-likelihood are expected more appropriate than those based on the log likelihood. Experimental results show that a mixture model estimated by the beta-likelihood approximates a large joint probability table with sparse data more appropriately than EM estimators.