Predictive maintenanceThe latest products
Collecting and analyzing data: These are the basic requirements for condition monitoring and predictive maintenance solutions. Computer&AUTOMATION presents new products from these areas.
Mathworks today announces Release 2018a (R2018a) with a range of new features in Matlab and Simulink. The new Predictive Maintenance Toolbox enables engineers to tag data, design condition indicators, and predict and prevent machine failures. Machine data from local files, cloud storage and distributed file systems can be imported for analysis. The toolbox contains reference examples for motors, gearboxes, batteries and other machines, which provide helpful guidelines for developing your own predictive maintenance and condition monitoring algorithms. Further updates to the 2018a release in the area of data analysis include the ability to visualize high-density data using scatter plots. Texts can now be analyzed even more accurately, as the Text Analytics Toolbox now recognizes sentences, email addresses and URLs and can extract and count expressions from multiple words. There are also new functions in the area of deep learning: the Neural Network Toolbox now provides a support package that can be used to implement deep learning layers and networks designed in TensorFlow Keras. Optimization techniques such as Adam, RMSProp and gradient clipping ensure better training of networks. In addition, networks in the form of directed acyclic graphs (DAG) can be trained more quickly using multiple GPUs.