Design of an Innovative Test Rig for Industrial Bearing Monitoring with Self-Balancing Layout

The remote prognosis and diagnosis of bearings can prevent industrial system failures, but the availability of realistic experimental data, being as close as possible to those detected in industrial applications, is essential to validate the monitoring algorithms. In this paper, an innovative bearing test rig architecture is presented, based on the novel concept of “self-contained box”. The monitoring activity is applicable to a set of four middle-sized bearings simultaneously, while undergoing the independent application of radial and axial loads in order to simulate the behavior of the real industrial machinery. The impact of actions on the platform and supports is mitigated by the so-called “self-contained box” layout, leading to self-balancing of actions within the rotor system. Moreover, the high modularity of this innovative layout allows installing various sized bearings, just changing mechanical adapters. This leads to a reduction of cost as well as of system down-time required to change bearings. The test rig is equipped with suitable instrumentation to develop effective procedures and tools for in- and out-monitoring of the system. An initial characterization of the healthy system is presented.

Machine Learning for Wind Turbine Fault Prediction through the Combination of Datasets from Same Type Turbines

Early fault detection in wind turbines is key to reduce both costs and uncertainty in the generation of energy and operation of these structures. The isolation of many wind farms, especially those offshore, makes scheduled maintenance very costly and on many occasions inefficient. In addition, the downtime of these structures is typically long and a predictive solution is much needed to 1) help prepare for the maintenance procedure beforehand, for instance to avoid delays when waiting for the required resources and components for maintenance to be available and, 2) avoid the possibility of more destructive system failures. Predicting failures in such complex systems requires modeling of multiple components in isolation and as a whole. Physics-based and data-based models are used for this purpose, which have been proven useful in this regard. Specifically, Machine Learning algorithms are proven to be a valuable resource in a wide range of problems in this industry, however a solution capable of accurately predicting the range of faults of a particular type of wind turbine is still a challenge. In this paper, we will introduce the capabilities of machine learning for wind turbine fault prediction, as well as a technique to predict different types of faults. We will compare the performance of two well established machine learning algorithms (namely K-Nearest Neighbour and Random Forest classifiers) on real wind turbine data which have produced great levels of prediction accuracy. We also propose data augmentation methods to help enhance the training of ML models when wind turbine data is scarce by merging data from turbines of the same type.

Estimation of Maximum Potential Losses for Digital Banking Transaction Risks Using the Extreme Value-at-Risks Method

The application of industry 4.0 in banking presents many challenges, with several operational risks related to downtime and timeout services due to system failures. One of the operational risk management steps is to estimate the value of the maximum potential losses. The purpose of this study is to estimate the maximum potential losses for digital banking transaction risks. The method used for estimating risks is the EVaR method. There are several steps in this study. The first step is to resample the data using MEBoot. This process is a simulation of the operational risk loss data of digital banking. Next, the threshold value is determined to obtain the extreme data value. Then, a Kolmogorov–Smirnov test is conducted to fit the data with the GPD. Afterward, the GPD parameter is estimated. Then, EVaR is calculated using a portfolio approach to obtain a combination of risk values as maximum potential losses. The analysis results show that the maximum potential loss is IDR144,357,528,750.94. The research results imply that the banks need to pay attention to the maximum potential losses of digital financial transactions as a reference for risk management. Therefore, banks can anticipate the adequacy of reserve funds for these potential risks.

A new remote monitoring device to track magnetic resonance imaging machine cooling system failures

The magnetic resonance imaging (MRI) machine cooling system has a vital role in the conduct of MRI examinations because a shutdown of the MRI cooling system in the absence of the manipulators can lead to grave consequences over time, like quench, which is the vaporization of helium liquid in the MRI tank, and it's the most expensive MRI failure. To limit the risks of this problem, several companies have tried to develop a monitoring system to track MRI cooling system failures but all solutions proposed are complicated and demand many connections with MRI. The proposed solution is simple, easy, and efficient requires only one joint with the helium compressor, and it has a humidity and temperature sensor to detect quench incident, it works using an advanced monitoring algorithm that evaluates the status of the cooling system and identifies breakdowns, in case of failure our system will send short message service (SMS) notifications and emails to the customer service team. The proposed solution shows the potential for starting the research to understand the relationship between the behavior of the MRI cooling system and the quench using machine learning algorithms.

Qualitative Analysis of State/Event Fault Trees Based on Interface Automata

State/Event Fault Tree (SEFT) can be used for safety modeling and assessment. However, SEFT does not provide adequate semantics for analyzing the minimal scenarios leading to system failures. In this paper, we propose a novel qualitative analysis method for SEFT based on interface automata. Firstly, we propose the concept of guarded interface automata by adding guards on interface automata transitions. Based on this model, we can describe the triggers and guards of SEFT simultaneously. Then, a weak bisimilarity operation is defined to alleviate the state space explosion problem. Based on the proposed guarded interface automata and the weak bisimilarity operation, the semantics of SEFT can be precisely determined. After that, a qualitative analysis process is presented on the basis of the formal semantics of SEFT, and the analyzing result is the minimal cut sequence set representing the causes of system failures. Finally, a fire protection system case study is illustrated step by step to demonstrate the effectiveness of our method.