Tri Kurniawan Wijaya

Electricity load forecasting for residential customers: Exploiting aggregation and correlation between households

The recent development of smart meters has allowed the analysis of household electricity consumption in real time. Predicting electricity consumption at such very low scales should help to increase the efficiency of distribution networks and energy pricing. However, this is by no means a trivial task since household-level consumption is much more irregular than at the transmission or distribution levels. In this work, we address the problem of improving consumption forecasting by using the statistical relations between consumption series. This is done both at the household and district scales (hundreds of houses), using various machine learning techniques, such as support vector machine for regression (SVR) and multilayer perceptron (MLP). First, we determine which algorithm is best adapted to each scale, then, we try to find leaders among the time series, to help short-term forecasting. We also improve the forecasting for district consumption by clustering houses according to their consumption profiles.

Matching demand with supply in the smart grid using agent-based multiunit auction

Recent work has suggested reducing electricity generation cost by cutting the peak to average ratio (PAR) without reducing the total amount of the loads. However, most of these proposals rely on consumers willingness to act. In this paper, we propose an approach to cut PAR explicitly from the supply side. The resulting cut loads are then distributed among consumers by the means of a multiunit auction which is done by an intelligent agent on behalf of the consumer. This approach is also in line with the future vision of the smart grid to have the demand side matched with the supply side. Experiments suggest that our approach reduces overall system cost and gives benefit to both consumers and the energy provider.

Minimizing Human Effort in Reconciling Match Networks

Schema and ontology matching is a process of establishing correspondences between schema attributes and ontology concepts, for the purpose of data integration. Various commercial and academic tools have been developed to support this task. These tools provide impressive results on some datasets. However, as the matching is inherently uncertain, the developed heuristic techniques give rise to results that are not completely correct. In practice, post-matching human expert effort is needed to obtain a correct set of correspondences. We study this post-matching phase with the goal of reducing the costly human effort. We formally model this human-assisted phase and introduce a process of matching reconciliation that incrementally leads to identifying the correct correspondences. We achieve the goal of reducing the involved human effort by exploiting a network of schemas that are matched against each other.We express the fundamental matching constraints present in the network in a declarative formalism, Answer Set Programming that in turn enables to reason about necessary user input. We demonstrate empirically that our reasoning and heuristic techniques can indeed substantially reduce the necessary human involvement.

Mining complex activities in the wild via a single smartphone accelerometer

Complex activities are activities that are a combination of many simple ones. Typically, activities of daily living (ADLs) fall in this category. Complex activity recognition is an active area of interest amongst sensing and knowledge mining community today. A majority of investigations along this vein has happened in controlled experimental settings, with multiple wearable and object-interaction sensors. This provides rich observation data for mining. Recently, a new and challenging problem is to investigate recognition accuracy of complex activities in the wild using the smartphone. In this paper, we study the strength of the energy-friendly, cheap, and ubiquitous accelerometer sensor, towards recognizing complex activities in a complete real-life setting. In particular, along the lines of hierarchical feature construction, we investigate multiple higher-order features from the raw sensor stream (x, y, z, t). Further, we propose and evaluate two SVM-based fusion mechanisms (early fusion vs. late fusion) using the higher-order features. Our results show promising performance improvements in recognizing complex activities, w. r.t. prior results in such settings.

SmartD: smart meter data analytics dashboard

The ability of smart meters to communicate energy consumption data in (near) real-time enables data analytics for novel applications, such as pervasive demand response, personalized energy feedback, outage management, and theft detection. Smart meter data are characterized by big volume and big velocity, which make processing and analysis very challenging from a computational point of view. In this paper we presented SmartD, a dashboard that enables the data analyst to visualize smart meter data and estimate the typical load profile of new consumers according to different contexts, temporal aggregations and consumer segments.

Symbolic representation of smart meter data

Currently smart meter data analytics has received enormous attention because it allows utility companies to analyze customer consumption behavior in real time. However, the amount of data generated by these sensors is very large. As a result, analytics performed on top of it become very expensive. Furthermore, smart meter data contains very detailed energy consumption measurement which can lead to customer privacy breach and all risks associated with it. In this work, we address the problem on how to reduce smart meter data numerosity and its detailed measurement while maintaining its analytics accuracy. We convert the data into symbolic representation and allow various machine learning algorithms to be performed on top of it. In addition, our symbolic representation admit an additional advantage to allow also algorithms which usually work on nominal and string to be run on top of smart meter data. We provide an experiment for classification and forecasting tasks using real-world data. And finally, we illustrate several directions to extend our work further.

SMART: A tool for analyzing and reconciling schema matching networks

Schema matching supports data integration by establishing correspondences between the attributes of independently designed database schemas. In recent years, various tools for automatic pair-wise matching of schemas have been developed. Since the matching process is inherently uncertain, the correspondences generated by such tools are often validated by a human expert. In this work, we consider scenarios in which attribute correspondences are identified in a network of schemas and not only in a pairwise setting. Here, correspondences between different schemas are interrelated, so that incomplete and erroneous matching results propagate in the network and the validation of a correspondence by an expert has ripple effects. To analyse and reconcile such matchings in schema networks, we present the Schema Matching Analyzer and Reconciliation Tool (SMART). It allows for the definition of network-level integrity constraints for the matching and, based thereon, detects and visualizes inconsistencies of the matching. The tool also supports the reconciliation of a matching by guiding an expert in the validation process and by offering semi-automatic conflict-resolution techniques.

An economic analysis of pervasive, incentive-based demand response

Demand response (DR) emerges as one of the cheapest and greenest solutions to match supply and demand in the electricity sector. While DR has been focused on large and industrial consumers, pervasive implementation (by including residential consumers) is needed to maximize its potential. This paper presents theoretical analysis of pervasive, incentive-based DR from the economics perspective. Our analysis consider cases whether (1) DR is used to encourage consumers to decrease or increase their demand, and (2) utility companies have access to a single or multiple energy sources. We determine the necessary conditions and derive the optimal incentives to benefit from DR events.

Online unsupervised state recognition in sensor data

Smart sensors, such as smart meters or smart phones, are nowadays ubiquitous. To be “smart”, however, they need to process their input data with limited storage and computational resources. In this paper, we convert the stream of sensor data into a stream of symbols, and further, to higher level symbols in such a way that common analytical tasks such as anomaly detection, forecasting or state recognition, can still be carried out on the transformed data with almost no loss of accuracy, and using far fewer resources. We identify states of a monitored system and convert them into symbols (thus, reducing data size), while keeping “interesting” events, such as anomalies or transition between states, as it is. Our algorithm is able to find states of various length in an online and unsupervised way, which is crucial since behavior of the system is not known beforehand. We show the effectiveness of our approach using real-world datasets and various application scenarios.

Methodologies for effective demand response messaging

Demand Response (DR) is considered an effective mechanism by utilities worldwide to address demand supply mismatch and reduce energy consumption, peak load and emissions. Consumer participation is central to realize the full potential offered by DR programs. The communication between a utility company and consumers participating in DR is through DR messages. However, despite the importance of DR messages in the context of residential DR programs, only a limited number of relevant experimental studies have been reported in literature so far. To address this gap, in this paper, we report findings from 6-month long DR field trials involving residential participants in Luleå, Sweden. The trials specifically focus on four aspects related to DR messages - notification mechanism, message type, associated incentive, and participation feedback. The primary outcome of these trials is a set of guidelines and recommendations for design of effective DR programs.