Research by IndustryWeek shows that equipment failure is the cause of 42 per cent of unplanned downtime. This raises questions about why more manufacturers aren’t using technology to improve machine and factory performance. Here, Stefan Reuther, Chief Sales Officer at industrial software expert COPA-DATA, explains how to leverage predictive analytics to reduce factory downtime and increase business competitiveness.
A report from Statistics MRC estimates that the global predictive analytics market will grow from $3.89 billion in 2016, to $14.95 billion in 2023. You could argue that this growth is a result of advancing technology. However, I believe that a larger contributor to this adoption is the growing understanding of the benefits predictive analytics can offer.
Traditionally, maintenance in manufacturing facilities has been performed on a fixed cycle, often referred to as scheduled maintenance. Plant managers schedule time to conduct checks on machinery and plan downtime for more in–depth maintenance. However, this can mean that parts are replaced too late, or even too early, increasing the risk of failure.
As an alternative, predictive analytics empowers plant managers to detect potential defects before they materialize — and avoid unnecessary maintenance for parts that do not need repair. It also prevents unnecessary downtime by allowing potential problems to be fixed without the need to stop the machinery.
These analytics help plant managers spot potential hardware or system problems by comparing historical machine data with current performance. However, there is much more to the software than simply identifying future equipment failures. One example is the integration of new machinery.
Industry 4.0–compliant systems and technologies are changing the factory floor. While advancement is great, investment in this machinery can be costly -and how can manufacturers ensure that these changes will be positive?
This has been the question plaguing plant managers in recent years. Thankfully, using analytics, they no longer have to accept a sales person’s guarantee that their offering is the best option. Analytics can predict how the process will react as a result of a change in specific parameters, giving manufacturers an accurate view of whether the investment has been beneficial.
To support this, predictive analytics software, like COPA-DATA’s zenon, can statistically evaluate production quality. The technology, when implemented to function at its fullest capability, can monitor historical processes to predict potential production variances at a very early stage. Supporting quality assurance (QA) from the start of a process reduces wasted time, materials and money.
Predictive analytics is the tool that will make sense of all other Industry 4.0 technology. Connecting devices and collecting masses of data is only useful if that data is then presented in a way that is understandable and perhaps more importantly, actionable.
COPA-DATA’s “zenon,” for example, takes data from various sources, such as a PLC or smart meter, and uses predictive analytics reports to produce smart analyses. This enables decisions to be made based on past values with the assistance of prediction models. It supports two types of prediction: time-based prediction, which forecasts how a value will develop over time, and value-based prediction, which shows how one value will behave if another is changed.
This includes classic uses for predictive analytics, by monitoring how machinery will perform over time, but it can also help in other areas, such as reducing energy consumption. The system can take data about previous energy usage and predict how it will be affected in the future based on changes to production tools or processes.
Predictive maintenance is incredibly useful, and it will help manufacturers reduce the risk of unplanned downtime. However, we shouldn’t limit predictive analytics just to this area of manufacturing. There is more to remaining competitive in industry than machine performance, and predictive analytics is the key to unlocking success.