Cristina Verde

Multi-leak detection and isolation in fluid pipelines

Abstract A multi-leak detection system for pipelines is designed and tested. The multi-leak detection problem is solved using only sensors of flow and pressure at the extremes of the duct, and using the analytical redundancy given of these measurements. The leak detection design is based on a distributed pipeline model that is discretized in space and assumes a set of leaks distributed through the duct. Leak location is accomplished by evaluating the residuals of a bank of unknown input observers that are robust against one leak and sensitive to the rest. Simulation and experimental results are reported to demonstrate the effectiveness of the proposed approach when two leaks appear simultaneously.

High accuracy operon prediction method based on STRING database scores

We present a simple and highly accurate computational method for operon prediction, based on intergenic distances and functional relationships between the protein products of contiguous genes, as defined by STRING database (Jensen,L.J., Kuhn,M., Stark,M., Chaffron,S., Creevey,C., Muller,J., Doerks,T., Julien,P., Roth,A., Simonovic,M. et al. (2009) STRING 8–a global view on proteins and their functional interactions in 630 organisms. Nucleic Acids Res., 37, D412–D416). These two parameters were used to train a neural network on a subset of experimentally characterized Escherichia coli and Bacillus subtilis operons. Our predictive model was successfully tested on the set of experimentally defined operons in E. coli and B. subtilis, with accuracies of 94.6 and 93.3%, respectively. As far as we know, these are the highest accuracies ever obtained for predicting bacterial operons. Furthermore, in order to evaluate the predictable accuracy of our model when using an organisms data set for the training procedure, and a different organisms data set for testing, we repeated the E. coli operon prediction analysis using a neural network trained with B. subtilis data, and a B. subtilis analysis using a neural network trained with E. coli data. Even for these cases, the accuracies reached with our method were outstandingly high, 91.5 and 93%, respectively. These results show the potential use of our method for accurately predicting the operons of any other organism. Our operon predictions for fully-sequenced genomes are available at http://operons.ibt.unam.mx/OperonPredictor/.

Original article: Exponential nonlinear observer for parametric identification and synchronization of chaotic systems

This work proposes the use of a new exponential nonlinear observer for the purpose of parametric identification and synchronization of chaotic systems. The exponential convergence of the observer is guaranteed by a persistent excitation condition. This approach is shown to be suitable for a wide variety of chaotic systems. In order to illustrate the observer design procedure, several examples with simulation results are presented.

Leak detection using parameter identification

Abstract This work proposes an approach to detect single and multiple leaks. The task is carry out by identifying the parameters of finite models associated with the leak events. The identification problem is attacked by using the Prediction Error Method (PEM). In addition, a frequency evaluation is realized to check the conditions for implementing the PEM or any other method which require an excitation condition.

Monitorability Analysis for a Gas Turbine Using Structural Analysis

Abstract This paper studies the monitorability of a gas turbine in a combined cycle power plant with respect to sensors fault. The analysis is carried on using the structural analysis where the structurally relations between known and unknown variables are obtained from a nonlinear complex dynamic model described by 37 equations with twelve sensors. The structure decomposition in just-constrained and over-constrained subsets allows a separation of the monitorability analysis in two parts and a reduction in the dimension of the structure from which the causal incidence matched matrices and the redundancy relations are obtained. The analysis concludes that all the turbine variables can be evaluated from measurements and control actions in normal conditions and this capability is only robust with respect to fault in two sensors. Moreover, the monitorability of the structure can be recovered for a set of three sensors fault by redundancy. However, the capability to compute all variables of the structure is destroyed if one of a set of seven sensors is not reliable even some of them are in the over-constrained subsystem. From the just-constrained subsystem one identifies that the plant components without redundancy are associated to the variables of the proper gas turbine.

Liénard type model of fluid flow in pipelines: Application to estimation

This paper highlights how fluid flow in a pipeline can be represented as a nonlinear model of so-called Liénard type. It is then shown how the structure of this model is suitable for the design of algorithms to identify parameters of a pipeline or estimate unmeasurable states. This approach is illustrated by simulation results, for instance showing how to estimate the Darcy-Weisbach friction coefficient or the fluid acceleration.

Multi-leak Reconstruction in Pipelines by Sliding Mode Observers

In this paper a solution for the multileak reconstruction issue is introduced using a sliding mode observer scheme. Since fluid model in a pressurized single pipeline with only known pressures and flows at the ends does not satisfy the sliding mode observer conditions for leak reconstruction, a cascade scheme to satisfy such conditions is proposed. The key of the proposal is the estimation of pressures at the leak points through successive estimations of internal states up to those in which faults affect. The performance of the technique and its robustness with respect to operation point changes after the leak occurrences are illustrated by simulated and experimental results.

Proportional Integral Derivative Based Fault Tolerant Control for a Gas Turbine

This paper presents a fault tolerant control switching scheme with embedded fault detection and isolation system for the gas turbine of a combined cycle power plant. Mechanical faults of the turbogenerator are considered as case study. Previous to the control design, the detection and isolation capabilities of the gas turbine are studied by applying structural analysis to a first principles based gas turbine model. As a result, a new sensor was suggested to improve the detectability and isolability with respect to mechanical faults and sensors faults. Furthermore, the active fault tolerant control developed is based on stabilizing families of proportional integral derivative (PID) controllers, which are tuned off-line and the plant switching scheme preserves the stability of the whole closed-loop system, thanks to a careful selection of controller according to fault conditions. Simulation results with nonlinear model show the potential of the procedure.

Analytical redundancy for a gas turbine of a combined cycle power plant

This paper studies the structural properties related with the supervision of a gas turbine unit of a combined cycle power plant with sensors faults. The analysis is carried on using the structure defined by constraints and variables contained in a non-linear complex dynamical model. Considering the model described by 37 algebraic and differential equations, the structural analysis based on the matching procedure of the Boolean incidence matrix is used to study the analytical redundancy of the turbine and its redundancy degrees for fault diagnosis. The conclusions obtained with the analysis allows to suggest which sensors and variables have to be considered as the most critical in the turbine from integrity point of view

Structural Analysis for Fault Diagnosis in a Pipeline 1

Abstract This paper deals with generic properties of dynamic systems structure. The major part of the paper discusses conditions for the solvability of the Fault Detection and Isolation, FDI, issues using the Structural Analysis as framework. From a generic paths of the output to the unknown variables, one shows how to extract available inherent redundant information of the system in terms of its graph. The key of the analysis is the concept of monitorability which is related to the Analytical Redundancy Relations and can be checked by means of direct graphs. We review results concerning the FDI and Fault Tolerance Control problems and show with an illustrative example of a transport phenomena process with 14 constraints, the simplicity and advantages of the Structural Analysis.