Risk assessment and management are some of the major tasks of urban power-grid management. The growing amount of data from, e.g., prediction systems, sensors, and satellites has enabled access to numerous datasets originating from a... more
Risk assessment and management are some of the major tasks of urban power-grid management. The growing amount of data from, e.g., prediction systems, sensors, and satellites has enabled access to numerous datasets originating from a diversity of heterogeneous data sources. While these advancements are of great importance for more accurate and trustable risk analyses, there is no guidance on selecting the best information available for power-grid risk analysis. This paper addresses this gap on the basis of existing standards in risk assessment. The key contributions of this research are twofold. First, it proposes a method for reinforcing data-related risk analysis steps. The use of this method ensures that risk analysts will methodically identify and assess the available data for informing the risk analysis key parameters. Second, it develops a method (named the three-phases method) based on metrology for selecting the best datasets according to their informative potential. The meth...
Research Interests:
Research Interests:
ABSTRACT
As complexity and safety requirements of current and future nuclear power plants increase, innovative methods are being investigated to perform accurate and reliable system diagnoses. Detecting malfunctions, identifying their causes and... more
As complexity and safety requirements of current and future nuclear power plants increase, innovative methods are being investigated to perform accurate and reliable system diagnoses. Detecting malfunctions, identifying their causes and possibly predicting their consequences are major challenges, especially if extended to the whole plant. Monitoring plant performances by means of data reconciliation techniques has proved successful to detect anomalies during operation, provide early warnings and eventually schedule maintenance. At the same time, the large amount of information provided by large-scale monitoring systems is hard to handle manually. In this HWR, a function-oriented modelling approach, called Multilevel Flow Modelling (MFM), is proposed for performing an automatic analysis of the outcomes of the monitoring systems with the aim of identifying the root causes of the possibly detected anomalies. Reasoning in MFM models derives its power from representation of process knowl...
The operation of any industrial plant is based on the readings of a set of sensors. The ability to identify the state of operation, or the events that are occurring, from the time evolution of these readings is essential for tasks such as... more
The operation of any industrial plant is based on the readings of a set of sensors. The ability to identify the state of operation, or the events that are occurring, from the time evolution of these readings is essential for tasks such as supervisory control, detection and diagnosis of faults, and process quality control. Reasoning in time, however, is very demanding, because time introduces a new dimension with significant levels of additional freedom and complexity. The real-time history of scores of variables can be displayed and monitored in most computerized process monitoring and control systems. However, when the process is in significant transience or crises have occurred, the displayed trends of interacting variables and alarms can easily overwhelm an operator. In this paper we describe how a combined use of wavelets and recurrent neural networks improves on our previously proposed solutions to the transient classification problem. In particular, the newly developed system ...
Research Interests:
Recent advances in IT-related fields are opening up a broad range of novel applications. This is especially true in the energy sector, where Smart Grid solutions are offering new opportunities for the monitoring of power transmission and... more
Recent advances in IT-related fields are opening up a broad range of novel applications. This is especially true in the energy sector, where Smart Grid solutions are offering new opportunities for the monitoring of power transmission and distribution in electrical grids. However, optimal use of potentially accessible data sources is challenging, and most of the current Smart Grid projects continue to exhibit suboptimal utilization of heterogeneous information. This situation is also faced when it comes to the assessment of risks associated to operation of electricity transmission and distribution networks. As a consequence, current management systems fail to provide accurate estimations of risk levels in real-world situations. Our paper addresses this issue and contributes to the identification of possible solutions. The paper identifies a number of heterogeneous data sources which could be relevant for risk assessment, but which are currently not fully exploited. Furthermore, the p...
Research Interests:
With the emergence of cloud computing and sensor technologies, Big Data analytics for the Internet of Things (IoT) has become the main force behind many innovative solutions for our society's problems. This paper provides practical... more
With the emergence of cloud computing and sensor technologies, Big Data analytics for the Internet of Things (IoT) has become the main force behind many innovative solutions for our society's problems. This paper provides practical explanations for the question "why is the number of Big Data applications that succeed and have an effect on our daily life so limited, compared with all of the solutions proposed and tested in the literature?", with examples taken from Smart Grids. We argue that "noninvariants" are the most challenging issues in IoT applications, which can be easily revealed if we use the term "invariant" to replace the more common terms such as "information", "knowledge", or "insight" in any Big Data for IoT research. From our experience with developing Smart Grid applications, we produced a list of "noninvariants", which we believe to be the main causes of the gaps between Big Data in a laborator...
Research Interests:
In this paper, we equip Prototypical Networks (PNs) with a novel dissimilarity measure to enable discriminative feature normalization for few-shot learning. The embedding onto the hypersphere requires no direct normalization and is easy... more
In this paper, we equip Prototypical Networks (PNs) with a novel dissimilarity measure to enable discriminative feature normalization for few-shot learning. The embedding onto the hypersphere requires no direct normalization and is easy to optimize. Our theoretical analysis shows that the proposed dissimilarity measure, denoted the Squared root of the Euclidean distance and the Norm distance (SEN), forces embedding points to be attracted to its correct prototype, while being repelled from all other prototypes, keeping the norm of all points the same. The resulting SEN PN outperforms the regular PN with a considerable margin, with no additional parameters as well as with negligible computational overhead.
Research Interests:
Predicting the occurrence of failures in power grids through specific outage risk predictors is a primary concern for utilities nowadays. Wooden poles represent core items to focus on in this process. Millions of them are used worldwide... more
Predicting the occurrence of failures in power grids through specific outage risk predictors is a primary concern for utilities nowadays. Wooden poles represent core items to focus on in this process. Millions of them are used worldwide and they are all subject to the risk of crack formation. Analyzing the evolution of pole cracks is particularly relevant in reliability analyses of power grids for two main reasons. First: the cracks might highlight previously unconsidered or changing factors, such as unusual local weather conditions (e.g. overload of ice and/or wind). Second: as cracks provide an access for external threats (e.g. humidity, fungi, insects) to potentially non-treated internal parts of the poles, they might in turn accelerate the occurrence of further failures. Evaluating the role of crack formation is thus essential for estimating the risk of outages in power grids. As climatic variations are known to be among the most influencing factors in the initiation and propaga...
Research Interests:
Research Interests:
On-line sensor monitoring and diagnostics systems aim at detecting anomalies in sensors and reconstructing their correct signals during operation. Since 1994, research at the OECD Halden Reactor Project has focused on the problem of... more
On-line sensor monitoring and diagnostics systems aim at detecting anomalies in sensors and reconstructing their correct signals during operation. Since 1994, research at the OECD Halden Reactor Project has focused on the problem of sensor monitoring and diagnostics, eventually leading to the development of the PEANO system for signal validation and reconstruction. PEANO combines empirical techniques like Fuzzy Clustering and AutoAssociative Neural Networks and has proved to be successful in a variety of practical applications. Nevertheless, using one single empirical model sets a limit to the number of signals that can be handled at a time. Recently, efforts have been made to extend the applicability of PEANO to the whole plant, which requires the validation and reconstruction of thousands of signals. This has entailed moving from a single-model to an ensemble-of-model approach which has involved the investigation of new issues. This paper presents the method hereby developed for o...
Validity and accuracy of sensor signals are crucial for the enhancement of the safety, reliability and performance of complex industrial systems such as nuclear power plants. In this view, on-line sensor monitoring aims at detecting... more
Validity and accuracy of sensor signals are crucial for the enhancement of the safety, reliability and performance of complex industrial systems such as nuclear power plants. In this view, on-line sensor monitoring aims at detecting anomalies in sensors and reconstructing their correct signals during operation. Auto-associative regression models are commonly adopted to perform the signal reconstruction task. Nevertheless, on real scale applications the number of sensors signals is too large to be handled effectively by one single model. To overcome this problem, one may resort to an ensemble of reconstruction models, each one handling an individual (small) group of sensor signals. The outcomes of the models need then to be opportunely combined. In this work, three methods for aggregating the outcomes of the individual reconstruction models of a randomized ensemble are implemented, applied and compared on a case study concerning the reconstruction of 920 simulated signals of the Swed...
Research Interests:
On-line sensor monitoring systems aim at detecting anomalies in sensors and reconstructing their correct signals during operation. Auto-associative regression models are usually adopted to perform the signal reconstruction task. In full... more
On-line sensor monitoring systems aim at detecting anomalies in sensors and reconstructing their correct signals during operation. Auto-associative regression models are usually adopted to perform the signal reconstruction task. In full scale implementations however, the number of sensors to be monitored is very large and cannot be handled effectively by a single reconstruction model. This paper tackles this issue by resorting to an ensemble of reconstruction models in which each model handles a small group of signals. In this view, firstly a procedure for generating the signal groups must be set. Then, a corresponding number of signal reconstruction models must be built on the bases of the groups and, finally, the outcomes of the reconstruction models must be aggregated. In this paper, three different signal grouping approaches are devised for comparison: pure-random, random-filter and random-wrapper. Signals are then reconstructed by Evolving Clustering Method (ECM) models. The me...
Research Interests:
Research Interests:
In unmanned aerial vehicle (UAV) flights, power lines are considered as one of the most threatening hazards and one of the most difficult obstacles to avoid. In recent years, many vision-based techniques have been proposed to detect power... more
In unmanned aerial vehicle (UAV) flights, power lines are considered as one of the most threatening hazards and one of the most difficult obstacles to avoid. In recent years, many vision-based techniques have been proposed to detect power lines to facilitate self-driving UAVs and automatic obstacle avoidance. However, most of the proposed methods are typically based on a common three-step approach: (i) edge detection, (ii) the Hough transform, and (iii) spurious line elimination based on power line constrains. These approaches not only are slow and inaccurate but also require a huge amount of effort in post-processing to distinguish between power lines and spurious lines. In this paper, we introduce LS-Net, a fast single-shot line-segment detector, and apply it to power line detection. The LS-Net is by design fully convolutional, and it consists of three modules: (i) a fully convolutional feature extractor, (ii) a classifier, and (iii) a line segment regressor. Due to the unavailabi...