Yu Fujimoto

Detection of Cyber Attacks Against Voltage Control in Distribution Power Grids With PVs

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.

A generalization of independence in naive bayes model

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.

On detection of cyber attacks against voltage control in distribution power grids

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.

Distributed Energy Management for Comprehensive Utilization of Residential Photovoltaic Outputs

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.

Pattern sequence-based energy demand forecast using photovoltaic energy records

Considering recent trends in energy technology development, consumers 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.

A grouped ranking model for item preference parameter

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.

Home energy management based on Bayesian network considering resident convenience

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.

A generalisation of independence in statistical models for categorical distribution

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.

An estimation of generalized bradley-terry models based on the em algorithm

The Bradley-Terry model is a statistical representation for ones 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.

Energy disaggregation based on semi-binary NMF

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.