Search

This chapter presents a fault detection method through uni- and multivariate hypothesis testing for wind turbine (WT) faults. A data-driven approach is used based on supervisory control and data acquisition (SCADA) data. First, using a healthy WT data set, a model is constructed through multiway principal component analysis (MPCA). Afterward, given a WT to be diagnosed, its data are projected into the MPCA model space. Since the turbu- lent wind is a random process, the dynamic response of the WT can be considered as a stochastic process, and thus, the acquired SCADA measurements are treated as a random process. The objective is to determine whether the distribution of the multivariate random samples that are obtained from the WT to be diagnosed (healthy or not) is related to the distribution of the baseline. To this end, a test for the equality of population means is performed in both the univariate and the multivariate cases. Ultimately, the test results establish whether the WT is healthy or faulty. The performance of the proposed method is validated using an advanced benchmark that comprehends a 5-MW WT subject to various actuators and sensor faults of different types.

© 20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works Industrial plants performance evaluation has become a difficult task due to the machinery complexity. Multi-chiller systems take up big proportion of energy in food and beverage companies. Complex refrigeration generation is usually hard to evaluate as the affectation of external signals plays an important role and also exist too many control features for the facility operator. Develop a method able to detect any deviation respect the optimal operation can provide the necessary information for the purpose of inefficiencies identification and a further optimization. In this paper, data-driven methods are used in order to describe a reliable coefficient of performance indicator (COP) in several known plant conditions. Self-organizing maps (SOM) are used to recognize different operating points among the multi-variable feature space for later performance evaluation. By the analysis of COP in each operating point, the potential energy saving can be illustrated. An experimental study is performed with refrigeration plant indicating the suitability of the proposed method

Industrial manufacturing plants often suffer from reliability problems during their day-to-day operations which have the potential for causing a great impact on the effectiveness and performance of the overall process and the sub-processes involved. Time-series forecasting of critical industrial signals presents itself as a way to reduce this impact by extracting knowledge regarding the internal dynamics of the process and advice any process deviations before it affects the productive process. In this paper, a novel industrial condition monitoring approach based on the combination of Self Organizing Maps for operating point codification and Recurrent Neural Networks for critical signal modeling is proposed. The combination of both methods presents a strong synergy, the information of the operating condition given by the interpretation of the maps helps the model to improve generalization, one of the drawbacks of recurrent networks, while assuring high accuracy and precision rates. Finally, the complete methodology, in terms of performance and effectiveness is validated experimentally with real data from a copper rod industrial plant.