Search

In this article, two methods for broken bar detection in induction motors are considered and tested using data collected from the LIAS laboratory at the University of Poitiers. The first approach is Motor Current Signature Analysis (MCSA) with Convolutional Neural Networks (CNN), in which measurements have to be processed in the frequency domain before training the CNN to ensure that the resulting model is physically informed. A double input CNN has been introduced to perform a 100% detection regardless of the speed and load torque value. A second approach is the Principal Components Analysis (PCA), in which the processing is undertaken in the time domain. The PCA is applied on the induction motor currents to eventually calculate the Q statistic that serves as a threshold for detecting anomalies/faults. Even if obtained results show that both approaches work very well, there are major differences that need to be pointed out, and this is the aim of the current paper.

The number and size of social networks have grown significantly as years have passed. With its 1.5 billion active users, Facebook is by far the most popular social networks on the planet. From kindergarten kids to grandparents to teenagers, Facebook attracts users of all ages, religions, personalities and social status. Facebook users are sharing their personal information, their lifestyle, their precious moments and their feelings online. In this paper, we download a set of comments from the page ‘Opposing Views’ from Facebook. These were then categorised into either a positive comment or a negative comment using the auto code feature in NVivo 11. Comments where no positive or negative sentiments are found are considered to be neutral. Out of 626 comments, 29.6% were found to contain positive sentiments while 62.0% were found to contain negative sentiments. The outcome of this work can be used by businesses to assess public reviews about their products. This will help them understand what is working and what is not. Thus, they can improve their products and respond to customer demands sufficiently quickly.

In hydrogeology, understanding the behavior of aquifers is an important issue. May it be for forecasting the availability of water on a long term or for estimating the impact of pollutants when injected at difierent regions, modeling of an aquifer is very useful for hydrogeologists. This paper gives an overlook of the difierent types of tests on a particularsite, namely the Hydrogeological Experimental Site (HES) of Poitiers (France). The aim is to show that the HES is a huge instrumented test bed allowing difierent types of experiments. Thus, system identification users can expect to gather large series of input-output datasets so as to experiment their own tools. Note that a large database exists concerning already-run tests, but what is interesting is that new experiments can be programmed so as to account for customized datasets. As a counterpart, the use of HES data should promote proper modeling of the underlying aquifer, thus contributing to the hydrological research field.

In this paper, a practical solution for the identification of Continuous-Time (CT) Input-Output (IO) Linear Parameter Varying (LPV) systems is proposed. For this particular class, we formulate an output error identification problem and present a parameter estimation scheme in which a prediction error based cost function is minimized using nonlinear programming. Because the cost function possesses local minima, the success of any iterative parameter estimation algorithm depends on appropriate initial seeds. One approach would be to generate an initial estimate using CT Reinitialized Partial Moments (RPM) models extended to CT IO LPV systems. We assume that the inputs, outputs and scheduling parameters are directly measurable, and that the functional dependence of the system coefficients on the parameters is of a polynomial form. The performance of the proposed method is evaluated by Monte Carlo simulation analysis.

In hydrogeology, estimating aquifer permeability is an important issue. This can be useful in understanding the flow of pollutants from one area of an aquifer to another. For this aquifer analysis sake, the Hydrogeological Experimental Site of Poitiers (France) covering a limestone aquifer is an appropriate instrumented test bed enabling measurement of hydraulic responses of a series of observation wells due to a step-type pumping out excitation at a given well. Given the input-output data, black-box continous-time modeling is quite a straight forward process as shown in this paper. The aim is then to be able to use the identified parameters to classify the different wells according to how sensitive they are to the one having been excited. A correlation between black-box parameters and hydrogeological ones is then established.

This research work deals with a new concept in the field of system identification and control. The common denominator to every proposed method is a particular class of systems that are characterised by some hybrid properties and by a piecewise functioning nature. These systems are particularly useful in designing tools for real time control. The present dissertation proposes a chapter on the concept of hybrid systems and the origin and nature of the piecewise functioning systems (PFS). The remaining chapters detail the theoretical and practical explanation concerning new methods of identification and con-trol based on PFS, and thus compatible with real time application. The online identifica-tion, which is carried out by a “cloning” method, is ensured by an adaptive algorithm that enables fast convergence. As for the control command, it is designed to perform sampled tracking of a given trajectory by the state of a linear plant, even in the case where the only available feedback comes from a digital sensor delivering the output of the plant in a delayed and sampled format. For each method, the reader can find a state of the art introduction, a mathematical description and examples coming from computer simulations and real time applications., Ce travail de recherche concerne une approche nouvelle d'identification et de commande de processus réels. Les travaux sont fondés sur une classe particulière de systèmes qui possèdent des propriétés hybrides et qui ont une dynamique caractérisée par un fonction-nement par morceaux. Ces systèmes permettent de développer des outils particulièrement adaptés à une architecture temps réel. Le mémoire consacre un chapitre au concept de système hybride et à l'origine et à la nature des systèmes à fonctionnement par morceaux (SFM). Les autres chapitres fournissent la mise en œuvre théorique et pratique de nouvel-les méthodes d'identification et de commande utilisant les SFM et donc adaptées au temps réel. L'identification en ligne, assurée par une méthode appelée « clonage », est régie par un algorithme adaptatif qui garantit une convergence rapide. La commande, quant à elle, vise à réaliser la poursuite échantillonnée d'une trajectoire consigne par l'état d'un système linéaire, même dans le cas où le seul retour possible correspond à l'information provenant d'un capteur numérique qui délivre la sortie du système sous forme retardée et échantillonnée. Chaque méthode est fournie avec une introduction permettant de la situer par rapport à l'existant, une formalisation mathématique et des exemples de simulation et d'implantation temps réel.

This paper presents a dynamic model of a counter flow water to oil heat exchanger when all inputs (inlet temperatures of the fluids and the mass flow rates) are simultaneously varying. Although interesting results about modeling of heat exchanger by linear parameter varying (LPV) can be found in [25] , several problems remain to be solved such as the structure estimation or a proper initial MISO model for the optimization algorithms. This paper introduces a new model structure called quasi LPV model which simulates accurately the temperature and flow transients in a counter flow heat exchanger (COFHX). The quasi LPV model is compared to a realistic numerical model of a counter flow heat exchanger adjusted with the test rig heat exchanger of the University of Valenciennes in France. Comparisons indicate that the developed quasi LPV model is capable of predicting the transient performance of the heat exchangers satisfactorily.

This article deals with the trajectory tracking of linear plants with sampled and delayed outputs. A class of sampled-data controllers with two time-scales is proposed which allows sampled tracking with a delay twice as much as that of the plant output. Numerical simulation results are presented to demonstrate the effectiveness of the proposed approach.

This paper presents a vision based Cart-Inverted Pendulum (CIP) system under a hybrid feedback configuration: the continuous cart’s position measured by encoder and the delayed & sampled inverted pendulum’s upper coordinates, obtained from a visual sensor. The challenge here is to stabilize the CIP from a big inclined initial angle by using a low cost CCD camera. Under this scheme, we propose a hybrid control which consists in a Jumpingup (Bang-Bang) control and a two causal stabilization loops control: the first one (inner loop) realizes a linearization and the stabilization control of the pendulum based on an innovative Piecewise Continuous Reduced Order Luenberger Observer coupled with a linearization module, the second one (the outer loop) realizes a Lyapunov based control for the unstable internal system with lower dynamics than that of the pendulum. This hybrid control method is capable of balancing the CIP system within small cart’s displacement. Performances issues of the proposed method are illustrated by the experimental figures and videos.

This paper presents a stabilization control system of a 3D inverted pendulum. The challenge is to realize a control via remote sensing based on a low cost CCD camera. The proposed method transforms the 3D problem into a 2D one by choosing a new balacing plane for the pendulum at each camera's sampling instant. For a given plane, two feedback loops are considered. The first one, which is an observation loop, processes the delayed and sampled angle information delivered by the artificial vision system so as to set up a linear and stable model of the pendulum (without any approximations) and to reconstitue its continuous state by means of a specific observer called Plecewise Continuous reduced-order Luenberger Observer (PCO). The second loop, which is a stabilitation loop , makes use of a Lyapunov function based control in order to stabilise the cart-pendulum system. Simulations and experimental results illustrate the method's performance.

This paper presents a vision based cart-inverted pendulum (CIP) system under a hybrid feedback configuration: the continuous cart's position measured by encoder and the delayed & sampled inverted pendulum's upper coordinates, obtained from a visual sensor. The challenge here is to stabilize the CIP from a big inclined initial angle by using a low cost CCD camera. Under this scheme, we propose a hybrid control which consists in a jumping-up (bang-bang) control and a two causal stabilization loops control: the first one (inner loop) realizes a linearization and the stabilization control of the pendulum based on an innovative piecewise continuous reduced order Luenberger observer (PCROLO) coupled with a linearization module, the second one (the outer loop) realizes a Lyapunov based control for the unstable internal system with lower dynamics than that of the pendulum. This hybrid control method is capable of balancing the CIP system within small cart's displacement. Performances issues of the proposed method are illustrated by the experimental figures and videos.

This paper presents a stabilization control system of a 2-DOF inverted pendulum. The challenge is to realize a control with a contact-less feedback including a low cost CCD camera. The global control system uses two feedback loops. The first one (inner loop) includes the delayed and sampled inverted pendulum's top coordinates, obtained from the CCD camera and the second one (outer loop) concerns the cart's position in continuous time (obtained from encoders). With this feedback scheme, we propose a specific control method by choosing an optimal balancing plane at each camera's sampling instant, in view of transforming the 2-DOF pendulum problem to a 1-DOF one. Then in this balancing plane, we realize two control loops: The inner loop, relative to the pendulum's angle, is controlled by a linearization and stabilization method based on an innovative observer called piecewise continuous reduced-order Luenberger observer (PCROLO). The outer loop uses a Lyapunov function based control scheme with slower internal dynamics compared to that of the pendulum, considering that the global controlled system (cart + pendulum) is an unstable non-minimum-phase system. Numerical simulations show that the stabilization control of the inverted pendulum on an x-y robot that are strongly non linear is successful.

Tracheostomies are the most common procedure performed in the intensive care units for patients who need prolonged respiratory support (>21 days).

International audience; This article deals with linear plants whose outputs are not available directly, but only via digital sensors which deliver them in a delayed and sampled format. First, we reconstitute the plant's state by using a Lyapunov-Krasovskii based observer. A sampled tracking control strategy is then proposed by combining the observer with a particular controller that belongs to a class of piecewise continuous systems. Computer simulation examples are presented so as to enhance the theoretical aspect. The method shows reliability and robustness against slight time-variations of the plant's parameters.

This article presents a control method to realize tracking of linear plants that deliver a delayed and sampled output. To do so, we make use of a class of piecewise functioning controller (PFC): bi-sampled controller. The use of this type of controller allows sampled tracking with a delay twice as much as that on the plant's output. Thus, we propose mathematical approach of the dynamics of our controller. We then give an adaptation of these mathematical approaches, using only the delayed and sampled output, by means of a reduced-order Luenberger observer. Computer simulation results are given so as to enhance the theoretical aspect of our method.

International audience; Cet article définit une stratégie de suivi de trajectoire pour les systèmes linéaires, dont le vecteur d'état n'est accessible que sous forme retardée et échantillonnée. Pour cela, on utilise une classe de contrôleurs à fonctionnement par morceaux (CFM) , les contrôleurs bi-échantillonnés. L'implantation de cette classe de contrôleurs garantit une poursuite échantillonnée avec un retard égal au double du retard sur l'état. Dans cette optique, nous proposons, dans un premier temps, un formalisme mathématique du contrôleur. Dans un second temps, en utilisant la commande optimale, nous proposons une adaptation de ce formalisme, pour un état retardé et échantillonné. Des résultats de simulations illustrent cette étude.

This article deals with the control of linear plants that deliver delayed and sampled output. We propose a control theory which makes use of a class of piecewise continuous systems and that enables sampled tracking. A mathematical approach of the dynamics of our control command is given. This approach uses only the delayed and sampled output vector as feedback. Moreover, a computer simulation example is presented so as to enhance the theoretical aspect. The method shows relialability and robustness against slight time-variations of the plant's parameters.