Laurent Hartert

A semi-supervised dynamic version of Fuzzy K-Nearest Neighbours to monitor evolving systems

Data issued of most real-world applications are evolving; they change constantly over time. In such applications, it is difficult to induce correctly a model (classifier) using traditional classification methods. Thus, it is important to use an adapted classification method to build a classifier and to update its parameters as new data is available. In this paper, we propose an adaptive classification approach based on the Fuzzy K-Nearest Neighbours (FKNN) method to monitor online evolving systems. The developed method, named semi-supervised Dynamic FKNN, comprises the following phases. In the first phase (detection phase), a class evolution can be detected and confirmed after the classification of each new pattern. Then in the second phase (adaptation phase), the parameters of the evolved class are updated incrementally. In the last phase (validation phase) the adapted classes are validated in order to keep only the representative ones. This approach is illustrated using an example of system which switches between several functioning modes.

Dynamic detection of nuclear reactor core incident

Surveillance, safety and security of evolving systems are a challenge to prevent accident. The dynamic detection of a hypothetical and theoretical blockage incident in the Phenix nuclear reactor is investigated. Such an incident is characterized by abnormal temperature rises in the neighbourhood of the concerned reactor core assembly. The dataset is the output temperature map of the reactor, it is provided by the Atomic Energy and Alternative Energies Commission (CEA). A real time approach is proposed, based on a sliding temporal window, it is divided into two steps. The first one behaves like a sieve, its function is to detect simultaneous temperature evolutions in a close neighbourhood which may induce a potential incident. When such evolutions are detected, the second step computes the temperature contrast between each assembly having these evolutions and its neighbourhood. This method permits to monitor the system evolution in real time while only few observations are required. Results are validated on various noisy realistic simulated perturbations.

Dynamic K-Nearest Neighbors for the monitoring of evolving systems

In this article, a new Pattern Recognition (PR) approach is proposed to monitor the functioning modes evolutions in dynamic systems. When a functioning mode evolves, the system characteristics change and the observations, i.e. the patterns, obtained on the system change too. In this case, classes representing the system functioning modes have to be updated by keeping representative patterns only. The developed PR approach is based on the K-Nearest Neighbors (KNN) method. It is named Dynamic KNN (DKNN) and comprises two phases: a detection phase to detect and confirm classes evolutions and an adaptation phase realized incrementally to update the evolved classes parameters and reduce the dataset. To illustrate this approach, the monitoring of weldings quality (good or bad) is realized on an industrial system, based on acoustic noises issued of weldings operations.

Dynamic Fuzzy Pattern Matching for the monitoring of non stationary processes

Abstract In the domain of diagnosis by Pattern Recognition, a pattern is a simplified observation about the system state and each class, containing similar patterns, represents a functioning mode. Classes issued of non stationary processes are dynamic and their characteristics vary over the time. In this paper, the classification method Incremental Fuzzy Pattern Matching (IFPM) is developed for the monitoring of non stationary processes. This development is based on the monitoring of the accumulative changes in the data distribution. When these changes reach a threshold prefixed by an expert, the data distribution will be recursively updated online using the recent and useful patterns.

Statistical Dynamic Classification to Detect Changes in Temperature Time Series

This article deals with the problem of change detection in the output temperature time series of the Phenix nuclear reactor core assemblies. These time series are provided by the Atomic Energy and Alternative Energies Commission (CEA). A hypothetical and theoretical blockage of an assembly cooling system could lead to a temperature rise of its nearest neighbours. To detect such a rise, first a signal preprocessing has been realized in several steps: simulation of a blockage, filtering, interpolation and re-sampling. Then, several statistical estimators have been calculated on sliding windows. The feature space has been determined based on the most discriminant parameters, including a derived third order moment. Finally, a set of classification rules has been defined to detect an assembly blockage. Thus, a statistical dynamic classification is realized online to obtain at most two classes. Results have been validated on several assemblies with different realistic perturbations.

Semisupervised Dynamic Fuzzy K -Nearest Neighbors

This chapter presents a semi-supervised dynamic classification method to deal with the problem of diagnosis of industrial evolving systems. Indeed, when a functioning mode evolves, the system characteristics change and the observations, i.e. the patterns representing observations in the feature space, obtained on the system change too. Thus, each class membership function must be adapted to take into account these temporal changes and to keep representative patterns only. This requires an adaptive method with a mechanism for adjusting its parameters over time. The developed approach is named Semi-Supervised Dynamic Fuzzy K-Nearest Neighbors (SS-DFKNN) and comprises three phases: a detection phase to detect and confirm classes evolutions, an adaptation phase realized incrementally to update the evolved classes parameters and to create new classes if necessary and a validation phase to keep useful classes only. To illustrate this approach, the diagnosis of a welding system is realized to detect the weldings quality (good or bad), based on acoustic noises issued of weldings operations.

Dynamic pattern recognition for the diagnosis of evolving systems

In this paper, we propose an approach to achieve the monitoring of the functioning (normal, faulty) of the Steam Generator (SG) of the nuclear Prototype Fast Reactor (PFR). This approach is based on three steps: signal analysis, clustering and classification. The first step analyzes the acoustic signals measuring the noises issued of the injection of water or Argon in the SG. These injections simulate a leakage representing a faulty functioning mode of the steam generator. The goal of the signal analysis is to determine the minimal set of parameters required to discriminate the normal and faulty modes in the feature space. In the clustering step, the patterns obtained by the acoustic signals analysis are labeled as belonging to the first class (non-injection) or to the second class (injection) corresponding respectively to normal and faulty functioning modes. Finally, the decision function is generated in the third step in order to assign a new pattern (new acoustic signal) to one of the two learned classes. We use the Semi-Supervised Dynamic Fuzzy K-Nearest Neighbours (SS-DFKNN) method to achieve the clustering and the online classification of the new incoming patterns.

A Hybrid Multi-classifier to Characterize and Interpret Hemiparetic Patients Gait Coordination

The characterization of inter-segmental coordination patterns in hemi paretic gait is interesting to improve the management of hemiparetic patients. Indeed, the analysis of the coordination patterns can help clinician to establish patient diagnosis and to choose a treatment. The coordination patterns used in this article were obtained from the Continuous Relative Phase (CRP) measure in the sagittal plane. The CRP correlates angle positions and velocity of two segments, i.e. parts of the patient leg, over each phase of the gait cycle. Thigh-shank and shank-foot CRPs were measured for 66 hemiparetic patients, 27 healthy subjects and 14 patients pre and post treatment. CRPs signals are classified using a multi-classifier. This classification permits to discriminate gait patterns for hemiparetic and healthy subjects. The multi-classifier is based on a structural and a statistical approaches used in parallel. The structural part of the proposed hybrid method keeps links between the data issued from CRPs and the statistical part converts CRPs into spatial scalar parameters. Then, using a similarity measure this approach permits to quantify the global gait coordination improvement of patients after a therapeutic treatment. The proposed approach uses only interpretable parameters in order to let the classification results be physically understandable.