Abstract
Condition Monitoring (CM) is the process of determining the state of a system according to a certain number of parameters. This ‘condition' is tracked over time to detect any developing fault or non desired behaviour. As the Information and Communication Technologies (ICT) continue expanding the range of possible applications and gaining industrial maturity, the appearing of new sensor technologies such as Macro Fiber Composites (MFC) has opened a new range of possibilities for addressing a CM in industrial scenarios. The huge amount of data collected by MFC could overflow most conventional monitoring systems, requiring new approaches to take true advantage of the data. Big Data approach makes it possible to take profit of tons of data, integrating in the appropriate algorithms and technologies in a unified platform. This chapter proposes a real time condition monitoring approach, in which the system is continuously monitored allowing an online analysis.
TopBackground
Wind energy is inexhaustible, ecologically and environmentally friendly. It is becoming one of the most widespread and productive methods for generating electrical energy (see Figure 1). Today, it is a mature technology and this energy source is applied to both large scale and small installations. It certainly has become a mainstay within the energy systems of many countries, and is recognized as a reliable and affordable source of electricity (Beattie & Pitteloud, 2012).
Figure 1. Wind turbine research center in Ohio.
In 2013, wind energy represented 3.5% of total energy demand. And by 2016 is expected to be the global installed capacity of 500,000 MW. In addition to onshore wind farms, wind farms are built in the sea (offshore), several kilometres from the coast, to take advantage of the best wind conditions to overcome the negative relief effects. In these installations it is common to find much more powerful machines than which are installed onshore. The diameter of the turbine is a crucial parameter: longer blades, more swept area and more energy produced.
This trend to building ever larger blades carries out certain problems. The blades have to bear more and more weight and strength due to its greater sweep area. This means an increased fatigue in the blade structure, and therefore any blade failure entails very high costs. It has been estimated that the time between failures in wind turbine blades is 5 years. The time spent in repairing one of these blades is 2 days on average in onshore wind turbines. However, in the case of offshore wind turbines the downtime could increase up to a month.
Key Terms in this Chapter
Acoustical Emission: Non-destructive technique used for detecting elastic waves produced by a material as repetitive loads are applied on it.
Data Mining: A subfield of computer science focused on patterns discovery in large data sets involving methods at the intersection of artificial intelligence, machine learning, statistics, and database systems.
Parallel Computing: A computation technique consisting on the concurrent execution of subtasks for the accomplishment of a bigger task.
Big Data: The set of methodologies and technologies that allow the capture, management and exploitation of considerable amounts of data within a tolerable period of time.
Macro Fiber Composite (MFC): A kind of low profile actuator and sensor consisting of a rectangular piezo-ceramic rods stacked between layers of adhesive, electrodes and polyimide film.
Cloud Computing: A set of internet-based computing services, resources, technologies and infrastructures that are remotely available for end-users.