2015 is looking to be the year that the Internet of Things comes into focus and is attainable, all be it with some serious work, by every type of manufacturing plant. Just yesterday Microsoft released the Cortana Analytics Suite with a nice focus on the internet of things and relevant use cases within the manufacturing environment.

At a first look at the market I want to take a deep dive into the McKinsey report released in June, “The Internet of things: Mapping the value beyond the hype.” You should definitely read the full report as well. In subsequent posts I will focus on the various use cases within manufacturing to explore the IoT space with you.

According to McKinsey factories (defined as any standardized production environment) will reap the largest economic value from the adoption of the internet of things. Their estimates lay the economic impact at $1.2 trillion to $3.7 trillion per year. The area that I would like to explore further is within the productivity and optimization improvements.

Technology

To gather data, which we will define as well, low-power sensors will be utilized throughout the production environment using radio frequency identification (RFID) tags and other hardware tracking devices. The sensors, energy requirements, data storage and analytical computing and visualization must be relatively inexpensive at the onset; however the impact and savings through the optimization and potential automation of the current process will provide the financial benefit to the supplier.

A couple of examples of sensors for connection and collecting data currently on the market include; Connect the Dots from Microsoft (https://github.com/MSOpenTech/connectthedots) and B-Scada (http://votplatform.com/en-US/Sensors).

A key point the report raises is the issue of interoperability. Wherein the benefits of the internet of things can be realized when the devices, sensors and application work together and speak the same technological language. The open standard for communication MtConnect (http://www.mtconnect.org/) would be an excellent starting point.

The example brought by McKinsey outlines an offshore oil rig with 30,000 sensors that would be used to monitor machines to schedule maintenance. The ability to combine data across all machines will increase the reach of the predictive maintenance analysis and improve the effectiveness of equipment maintenance by 100 to 200 percent.

Data & Analytics

In today’s current environment data is already being gathered in real-time, but is only utilized within anomaly detection and control. A known analytical method within the manufacturing space would be at process and quality control on end products or touch points along the line. The value added as mentioned several times throughout the report lies within optimization and predication. Anomaly detection is used to assess cases where the part is out of tolerance or the process has veered out of control within a certain standard deviation from the historical mean. Data from multiple machines and processes combined to illustrate a thorough subsection of the process or the entire process should be able to provide the insight into optimization. All departments within a plant would have to play nice for this to truly work.

For data analysis McKinsey views a multivariate structure and brings the example of condition based maintenance, where multiple sensors monitoring separate aspects of a water pump can be analyzed together to establish which pump is in need of replacement which would be understood long before the pump actually needed replacing. The concept being that the company would monitor various points in the process that lead to the pump breakdown, such as different chemicals. Other interesting applications are in optimizing the performance of the process or machines to improve quality levels, yield, energy consumption and inventory levels at each stage.
Within inventory control a fine balance is placed between carrying costs and the risk of being out of stock. Real time inventory monitoring and automated triggers from bins on the factory floor based on the weight and height of the material and on specific condition based algorithms. McKinsey estimates that IoT based inventory control could have a financial impact of $5 to $15 billion per year by the year 2025. Automated Inventory control and optimization would be a great starting point for any factory. A great example brought is that of Wurth USA who replenishes bins by means of camera technology used to monitor fill levels. The bin itself then wirelessly transmits inventory data to a management system that can reorder supplies as required. A second example within retail is the company called Trax who also utilizes image recognition for inventory control. Have a look at their concept here, http://traxretail.com/retail_technology/.

Summary of a few areas for optimization:

Inventory control
Equipment settings and their impact on product quality & yield
Energy consumption
Production & supply chain

As McKinsey puts it, with IoT manufacturers can gain a comprehensive view of every point in the production process to maintain an uninterrupted flow of finished goods and avoid defects. The combination of the process and quality is the key value as it pertains to optimizing the factory floor and continuously satisfying the customer base with high quality output.
We will further explore out the user cases provided by Microsoft next.

At an economic benefit of $4 trillion dollars for factories, the IoT is a space that should be highly focused on within every manufacturing environment.