Data Analytics: How Data Represents Reality

Data Analytics: How Data Represents Reality
Marginal Column
Content

November 2017

 

The wealth of data in the Internet of Things provides numerous opportunities for improving the quality of decision-making. But only if genuine information can be drawn from all those zeros and ones. This is where data analytics comes in.

Having a lot of data is a long way from being a guarantee of better decisions. Higher quality is only achieved when data analytics is used to obtain a representation of reality from the wealth of raw data. Data analytics refers to IT-based analysis methods - intelligent algorithms that turn bits and bytes into valuable information. Data analytics involves sorting data by validity, relevance and topicality, and then linking it appropriately. This allows processes to be controlled in real time - in a factory or on the road.

No accidents in traffic

Let’s stay on the road and look at the issue of assistance systems in cars - radar, lidar, ultrasound and stereo cameras - cars have numerous eyes scanning their surroundings. They can detect other vehicles, crossing traffic, pedestrians and obstructions - but only recently. Valid data was actually available more than a decade ago but its meaning remained a puzzle for the software. Researchers from the development departments of car manufacturers have traveled an arduous path. First of all, they had to use sensor fusion to combine the impressions from different systems. Only then could intelligent programs be developed to reliably detect acute situations in traffic at any time.

While ten years ago it was only possible to establish that there was “something” in front of the vehicle, modern driver assistance systems are now able to use the position of the feet, the head posture and the direction of vision to reliably identify whether a pedestrian will remain at the side of the road or is about to start crossing. The systems replicate the human intuition of an experienced driver. In both cases, measures are initiated to avert danger, for example braking or controlled swerving.

Preventive methods in production

In factories, systems are already in use that generate information from data at a very high level. Professor Felix Wortmann, scientific head of the Bosch IoT lab at St Gallen University, believes that industrial production is on the verge of new business models. “The next major service wave will be triggered by data analytics. It is an instrument that can create added value as part of a hybrid value creation process.” One example of this is predictive maintenance. For example, forward-looking maintenance is increasingly being used in rolling mills in the rubber industry. Because these production lines normally run continuously on a three-shift system, machine availability plays a hugely important role. It is ensured by predictive maintenance - for example using Rexroth’s Online Diagnostics Network (ODiN).

Instead of maintaining plants according to rigidly defined cycles, maintenance with ODiN is forward-looking. The core idea of this service package is to carry out maintenance by using a combination of sensors, cloud-based applications, and machine learning methods before a stoppage occurs.

Clever prediction

Based on a range of sensor data, a machine learning algorithm determines a normal healthy condition for the plant. ODiN then uses a data-based model to continuously calculate the rolling mill’s health index. If just a single measured value moves outside the tolerance range for a short time, this does not necessarily lead to an intervention. Ultimately, wear can rarely be identified from a single signal. However, if the health index deteriorates due to changes in the data from several sensors, the system warns of a problem. The health index is determined not only for the overall plant but for all system-critical components. This allows a targeted search for the cause of any faults. In the health index reports created on a regular basis, ODiN uses machine learning to provide appropriate information and helps to create specific recommendations for action.

Retrofit for the Internet of Things

To ensure that the method is not only the preserve of new installations, Bosch Rexroth is offering retrofit options for existing plants. The IoT Gateway links existing plants without sufficient connectivity to Industry 4.0 topologies. This allows them to be automatically integrated into higher-level systems and to be used for condition monitoring and data mining. Implementation can be carried out without changing the existing automation solution. Configuration of the IoT Gateway is web-based and requires no specialist programming.

Back to the future

The Bosch IoT Gateway enables operators of older machines to enjoy the many advantages of the Internet of Things (IoT). To illustrate its capabilities, Robert Bosch’s pedal-powered lathe from 1887 has been dragged into the Industry 4.0 age using sensors and software. Thanks to technical support, the lathe is ready for Industry 4.0. The Bosch Rexroth IoT Gateway combines sensors, software, and an IoT-capable industrial control, enabling the condition of the lathe to be monitored.

The global market for retrofit solutions is worth billions. Market observers unanimously agree that the quality of data analytics will be crucial to a company’s future success. In times that are shaped by huge complexity, rapid dynamism, volatility, and disruption, only those who are agile when it comes to reacting to changes will succeed. The only way to do this is with real time control.