Rafael Palacios

Anomaly detection in onboard-recorded flight data using cluster analysis

A method has been developed to support Flight Operations Quality Assurance (FOQA) by identifying anomalous flights based on onboard-recorded flight data using cluster analysis techniques. Unlike current techniques, the method does not require pre-defined thresholds of particular parameters, but detects data patterns which differ from the majority of flights by considering all the available flight parameters. The method converts time series data from multiple flight parameters into a high dimensional data vector. Each vector captures all the available information for a single flight. Cluster analysis of the vectors is performed to identify nominal flights which are associated with large clusters and anomalous flights that do not belong to a specific cluster. The method was applied to a representative Digital Flight Fata Recorder (DFDR) dataset from an international airline. Detailed analysis was performed on takeoff and approach for 365 B777 flights. Abnormal flights were detected using the cluster technique which was able to identify anomalous behaviors including: high and low energy states, unusual pitch excursions, abnormal flap settings, high wind conditions. In addition, data clusters representing nominal conditions were also detected. Three distinct takeoff clusters were identified in the B777 data: one represented a majority of the takeoff cases, one correlated with a specific high altitude airport, one correlated with reduced power takeoffs. This initial evaluation indicates that cluster analysis is a promising approach for the identification of anomalous flights from onboard-recorded flight data.

Analysis of Flight Data Using Clustering Techniques for Detecting Abnormal Operations

The airline industry is moving toward proactive risk management, which aims to identify and mitigate risks before accidents occur. However, existing methods for such efforts are limited. They rely on predefined criteria to identify risks, leaving emergent issues undetected. This paper presents a new method, cluster-based anomaly detection to detect abnormal flights, which can support domain experts in detecting anomalies and associated risks from routine airline operations. The new method, enabled by data from the flight data recorder, applies clustering techniques to detect abnormal flights of unique data patterns. Compared with existing methods, the new method no longer requires predefined criteria or domain knowledge. Tests were conducted using two sets of operational data consisting of 365 B777 flights and 25,519 A320 flights. The performance of cluster-based anomaly detection to detect abnormal flights was compared with those of multiple kernel anomaly detection, which is another data-driven anomaly ...

Experiences learned from the on-line internal monitoring of the behaviour of a transformer

Knowledge of the health of power transformers is important to prevent high costs caused by failures and to maintain the quality of the service. On-line monitoring and diagnosis seem to be the right way to reach this objective. In this paper a set of experiences related to the installation of vibration and temperature sensors inside a test transformer are described. Different types of sensors have been tested, and some of them discarded because of induced electromagnetic noise. Some of the main results and conclusions are shown in this paper. These experiences are included in a more general project called TRAFES, whose main objective is the continuous monitoring and diagnosis of large power transformers.

Feedback-based architecture for reading courtesy amounts on checks

The processing of bank checks is one application that continues to rely heavily on the movement of paper. Checks are currently read by human eyes and physically transported to the bank of the payer, involving significant time and cost. Since paper checks constitute a popular mechanism for noncash payments, and the vol- ume of checks continues to be high, there is a significant interest in the banking industry for new approaches that can read paper checks automatically. We propose a new approach to read the nu- merical amount field on the check; this field is also called the cour- tesy amount field. In the case of check processing, the segmenta- tion of unconstrained strings into individual digits is a challenging task because one must accommodate special cases involving con- nected or overlapping digits, broken digits, and digits physically con- nected to a piece of stroke that belongs to a neighboring digit. The described system involves three stages: the segmentation of the string into a series of individual characters, the normalization of each isolated character, and the recognition of each character based on a neural network classifier.

Analysis of stochastic problem decomposition algorithms in computational grids

Stochastic programming usually represents uncertainty discretely by means of a scenario tree. This representation leads to an exponential growth of the size of stochastic mathematical problems when better accuracy is needed. Trying to solve the problem as a whole, considering all scenarios together, yields to huge memory requirements that surpass the capabilities of current computers. Thus, decomposition algorithms are employed to divide the problem into several smaller subproblems and to coordinate their solution in order to obtain the global optimum. This paper analyzes several decomposition strategies based on the classical Benders decomposition algorithm, and applies them in the emerging computational grid environments. Most decomposition algorithms are not able to take full advantage of all the computing power available in a grid system because of unavoidable dependencies inherent to the algorithms. However, a special decomposition method presented in this paper aims at reducing dependency among subproblems, to the point where all the subproblems can be sent simultaneously to the grid. All algorithms have been tested in a grid system, measuring execution times required to solve standard optimization problems and a real-size hydrothermal coordination problem. Numerical results are shown to confirm that this new method outperforms the classical ones when used in grid computing environments.

Test bench for measuring the electrical properties of commercial thermoelectric modules

One of the most promising applications of thermoelectricity is the recovery of waste heat for the production of electrical energy. Nowadays, several thermoelectric companies manufacture commercial thermoelectric modules (TEMs) based on Bi/sub 2/Te/sub 3/ compounds specially designed to perform as Seebeck modules.. This paper describes a test bench (geometry, materials, measuring equipment) to analyse the behaviour of this type of modules working under several temperature differences (/spl Delta/Ts). This allows to estimate the potential electric power generated in an application where the optimum AT cannot be achieved because the amount of heat supplied by the heat source is too small, or there is a limitation in the heat dissipation capacity at the cold side. The paper also shows the results obtained using two commercial modules tested under different working conditions. Plots of voltage, electrical power generated, and efficiency versus electric current generated are also included.

Speed estimation of vehicles approaching an intersection: a digital image processing method

Abstract Most severe car accidents that occur in urban environments involve side impacts at street intersections, even at those regulated with traffic lights. Hence, it is very common to implement a small delay since one road changes to red until the other road changes to green. This delay is intended to avoid accidents in which a vehicle decides to go through the intersection after the sequence green–yellow–red is started, underestimating the time required to overtake the intersection. A better approach is to adjust the delay dynamically, depending on the speed of the vehicles approaching to the intersection. Using the dynamic approach, it is possible to improve traffic flow by reducing unnecessary delays, and to improve safety by applying longer delays when needed. This paper proposes a speed estimation method based on digital image processing of pictures taken with wireless cameras installed on top of existing traffic lights. The algorithm finds a vehicle in two consecutive images (either in day or night condition) and computes its speed by correlation. When a traffic light turns red, the systems estimates the speed of the cars approaching and decides to change the other road to green immediately, or to wait until it is safe to do so. The system was tested with real traffic flow at a street located in the city of Talavera de la Reina, Spain, with vehicles at different speeds. The image processing method proved to be accurate for this application, and adding the advantage of low cost equipment and easy installation results in a very attractive solution.

Component stress evaluation in an electrical power distribution system using neural networks

This paper presents a procedure that permits a qualitative evaluation of the stress in components of an electric power distribution system. The core of this procedure is the development of a set of models based on neural networks that are able to represent and to predict the normal behavior expected of the components under different working conditions. The paper includes the application to the characterization of the thermal behavior of power transformers and the operation time of circuit breakers.

Neural network models to detect airplane near-collision situations

Abstract The US Federal Aviation Administration (FAA) has been investigating early warning accident prevention systems in an effort to prevent runway collisions. One system in place is the Airport Movement Area Safety System (AMASS), developed under contract for the FAA. AMASS internal logic is based on computing separation distances among airplanes, and it utilizes prediction models to foresee potential accidents. Research described in this paper shows that neural network models have the capability to accurately predict future separation distances and aircraft positions. Accurate prediction algorithms integrated in safety systems such as AMASS can potentially deliver earlier warnings to air traffic controllers, hence reducing the risk of runway accidents even further. Additionally, more accurate predictions will lower the incidence of false alarms, increasing confidence in the detection system. In this paper, different incipient detection approaches are presented, and several prediction techniques are evaluated using data from one large and busy airport. The main conclusion is that no single approach is good for every possible scenario, but the optimal performance is attained by a combination of the techniques presented.

Assisting tools for a new maintenance planning in a power distribution system

This paper describes a framework in which several tools have been integrated in order to help the maintenance planning in a power distribution network. This framework will help to improve the process of decision making in an asset management context of a power distribution company. Two kinds of tools have been developed, one has been oriented to the identification of possible failure risk according to the history of unavailabilities occurred, and another one has been designed for the continuous monitoring and analysis of the life of main components such as power transformers and circuit breakers of a power distribution network. The information supplied by both tools is useful to perform maintenance actions when they are really required by the components monitored according to their health conditions, and to orient the maintenance efforts to the components with certain risk of failure. The final objective of these tools is to apply the resources needed when they are required, optimizing the maintenance costs and keeping a high quality of the service for the clients.