Automated Manufacturing: Invest in Artificial Intelligence

Advanced predictive maintenance tools are leading the way for investment from many automation manufacturers. These tools use artificial intelligence to detect faults early and prompt timely repairs. This brings down the overall cost of owning automation while also raising its availability.

Platforms for the rollout of bespoke machine learning models are also becoming more common. These models manage specific production processes. From laser cutting to iron casting, there are many processes involved. Making processes that are both quicker and more effective while requiring less oversight from experts.

ABB and Viking Analytics

ABB is a market leader in industrial automation and electrical systems. It has observed an increase in the demand for predictive maintenance tools from its existing and potential customers. In response, it has bought a stake in a Swedish condition monitoring software startup, Viking Analytics.

MultiViz Vibration is the name of the programme that Viking has developed for measuring vibrations with the assistance of artificial intelligence. ABB has integrated the predictive analytics engine from Viking into the deployment of its own Ability Asset Manager. The combined solution has the ability to forecast and prevent faults in the machinery.

In the year 2020, Viking Analytics won the ABB Electrification Startup challenge. Since then, the two companies have worked together on applications and the visualisation of data from condition monitoring sensors.

In the past, Viking Analytics has worked on other applications of artificial intelligence in industrial automation. To accomplish this, it is necessary to make iron foundries capable of implementing automated pouring procedures. It developed machine-learning algorithms to make rapid adjustments that would otherwise need intervention from an experienced operator.

Siemens Xcelerator 

Xcelerator is a software platform under which Siemens deploys many different artificial intelligence applications. Including, MindSphere and Predictive Services.

MindSphere is Siemens’ main industrial automation and internet-of-things solution. It puts into action a wide array of smart solutions. From predictive maintenance, to the location of obstacles in sewer lines.

Siemens has used MindSphere machine-learning models on many of its own production sites. This is to detect developing faults, plan maintenance, assess process stability, and predict product quality.

Additionally, an AI model that MindSphere has pre-trained is capable of foreseeing sewage clogs. It monitors how sewer outlets behave in response to heavy rainfall, and alerts operators to any unusual patterns of behaviour.

Predictive Services is Siemens’ solution for monitoring the condition of drive systems, presses, and depaneling machines. It uses artificial intelligence to analyse a machine’s operating condition and assists with the planning of maintenance activities. It has resulted in a 10% increase in line availability at a company that manufactures tissue paper. This is done by predicting motor failures and prioritising maintenance during production shut-downs.

Mitsubishi Maisart

Maisart is the brand name under which Mitsubishi markets its applications for artificial intelligence. This includes its Compact AI, automated design deep learning, and AI with smart learning features.

Compact AI launched as an AI-based, predictive maintenance application. It reduces maintenance costs and premature failures in Mitsubishi's servo systems and variable frequency drives. The application uses artificial intelligence to detect both mechanical wear and signs of corrosion on electronic circuits.

Mitsubishi has developed other solutions that include artificial intelligence in industrial automation applications. It has used pre-trained machine-learning algorithms in its electrical discharge machining and laser cutting products. The algorithms optimise machining strategies and continuously adapt machining parameters. It has also developed natural language models to enable its human-machine interfaces to interpret vague user instructions.

Looking well into the future, Mitsubishi has worked on a prototype application based on quantum machine learning. It does so by recognising features that have been detected by contactless, ultra-high frequency image sensors. The quantum machine learning models are quicker and more efficient than conventional models.

Schneider EcoStruxure

Schneider Electric’s building and industrial automation software, EcoStruxure, is the basis for its artificial intelligence offering. It allows users of industrial automation to develop and deploy their own AI applications. Using an edge computing platform running Linux, intelligent algorithms run right next to the managed equipment. Making it ideal for distributed automation and environments that use legacy automation infrastructure.

Schneider Electric has trialled this artificial intelligence solution on oil pumpjacks. Each pumpjack has a monitoring device called a dynamometer. These measure load variations during each pumping cycle. Expert operators analyse the dynamometer data to determine the pumpjack’s status and performance.

Investing a large amount of time and money in human resources is necessary for this. Schneider Electric has used artificial intelligence to automatically identify patterns in the data that might show faults, leaks, and worn components. The method lowers the likelihood of unexpected equipment failures and lowers costs on maintenance expenses.