Part One — The Intelligence Within
As a society we now expect products to be intelligent and conform to the norms that are being established. This has huge implications for all sectors of our economy, but if you are a manufacturer this expectation is amplified on several fronts. Consumers now expect an element of intelligence in the products you make, from fridges and lawn mowers, to door bells, trainers and even socks.
When done successfully, like Amazon’s smart speaker Alexa, it transforms an industry. But if approached incorrectly, a product can become dangerous both physically and virtually. A recent example showed how an electric scooter can be remotely controlled to accelerate and brake from 100 metres away. A great example of innovation but while security standards and laws are still being established, it could have its downsides.
Within the factory walls, owners and employees expect safety and augmentation, high availability of equipment and efficiency. The need to raise productivity to stay competitive is increasing and those that can apply technology to build higher quality products, faster and with features that consumers didn’t even realise they needed, will ultimately succeed.
For years we have had ERP systems feeding downline manufacturing execution systems (MES) and supervisory control and data acquisition (SCADA) systems, but while these have been connected, it has mostly been ad-hoc, one way and with minimal intelligence being applied. Even in the best cases it is a closed loop isolation and not used to redefine the product.
To give factories the confidence to take the plunge in full scale IoT adoption and realise the benefits, they need to be reassured about security and stability. Failing to start with a secure foundation, or utilising best practices and platforms is an expensive mistake to make. The United States Congress has just recently introduced the Internet of Things Cybersecurity Improvement Act aimed to change the ‘convenience over security’ consensus that seems to have emerged and define security standards for IoT companies that sell to the US government. This type of legislation is only set to increase across the globe, with the UK looking to introduce new laws regarding password protection for IoT gadgets.
Azure Sphere from Microsoft is an example of a technology used to build highly secure connected intelligent products. Security is at the heart of the microcontroller and combined with the operating system and cloud services provides in-depth defence against threats and the ability for manufacturers to focus on the real features of IoT, not the underlying concerns. Combine this with Epicor IoT and each manufactured product can be coupled with its digital twin, and bi-directional events can flow between the core ERP system, providing a next generation platform allowing manufacturers to move quickly with confidence and reimagine their business.
Part Two — Getting Innovation Right
We are seeing the rise of more connected enterprises applying intelligence to their factories to gain efficiency and operational excellence. You may have not heard of Sistema but like me you may notice you already have their products in your kitchen cupboards. The have millions of customers in 82 countries around the world and help make people’s lives easier with containers.
Based out of New Zealand and running at a relatively constant 20 percent year-on-year growth they switched to Epicor ERP in 2015, and previously had our advanced MES solution siloed for the last 10 years. Greg Heeley CTO has been on a mission with their team to use intelligence technology to support their growth and lay the foundations needed to operate with their 80 machines, 10 production lines and the shipping of 200x40ft containers per month.
While the industry is not yet at the price point for connected microcontrollers to embed into plastic containers we are seeing the adoption of passive RFID sensors supporting cold chains. For Greg and Sistema, Epicor IoT provides them with the missing piece of the puzzle, a way to connect the dots and experiment with further automation while harnessing the data to work faster and smarter.
And this is just the beginning for the manufacturing industry when it comes to realising the value of smart products and machine learning. Our next article delves a little deeper into the power of connected products and how to effectively harness the data the produce, for real business benefit and growth.
Part Three — The Power of Data
Intelligent things are fast becoming an expectation at home and in the workplace. Children are growing up with technology and expect to be able to swipe the TV like an iPad and talk to devices to control them. Even the older generation are getting to grips with the benefits of Echo and Alexa to make their lives easier.
But intelligent products are in fact more valuable to the manufacturer than the end consumer. Take Tesla for example. Every car connects back to them. The ability to perform over-the-air updates to the vehicle, correct security exploits, add new features and resolve other issues is incredibly valuable. It can delight users by keeping their experience fresh and innovative, but the true value is the by-product of this connectedness—data, and lots of it.
By capturing and harvesting anonymous data on usage, performance, failures, locations and events, the manufacturer can gain invaluable insights into consumers habits and expectations, product performance and false assumptions. It’s the virtuous cycle of improvement we see through artificial intelligence (AI)—more data yields a better product which yields more users, and so it repeats. AI must be used to harness that volume of data to build correlations, anomalies and suggestions that were not possible before.
Manufacturing an intelligent product or even enabling your intelligent factory has only become a reality for mid-market manufactures because of the advances and the maturity of IoT and cloud. It is now possible to integrate off the shelf components and on-demand computation to build products that can be deployed across the globe with intelligence tied back into your ERP systems. These products can be spoken to, updated over-the-air, perform translation on the fly and embed intelligence on the device at the edge. They can constantly mine data to realise value. To be successful you need to use the right combination of tools.
For example, when we buy a white goods appliance, such as a dishwasher (of which in the US alone saw 8.69m units shipped), we tend to think that each model is the same product. It has the same code on the front, looks the same, has the same manual. But, in reality it’s not a clone. Each one is unique, discrepancies in the raw materials and potential assembly aspects will change the behaviour and the failure profile over time, even when operated identically. A squeak, a worn seal, an over-heating component can all occur in different ways and sometimes its preventable and often its detectable.
By giving each product a nervous system, data from hundreds or even millions of devices can establish a feedback loop to the cloud and the manufacturing system of record, enriching the as-built bill of materials and traceability data. This allows manufacturers to learn from that data, usage patterns, predict failures, review design changes and establish new business models along with completely new experiences for their customers.
Part Four—A Data Crystal Ball
It is fascinating to look at how larger companies are utilising machine learning (ML) to be more efficient for their facilities. Google as an example showcased how it utilised widely available weather and historical turbine data to predict wind power output 36 hours ahead of actual generation and used this insight to provide commitments to the power grid. The result increased the value of its wind energy by 20 percent.
The ability to see around corners and infer trends, otherwise undetectable, results in a competitive advantage and better asset utilization. While Google has some of the best data scientists in the industry at DeepMind, there are increasing examples of small start-ups and manufacturers applying technology to be more efficient.
Epicor IoT utilises Azure to help manufacturers bridge data and IoT, allowing for not only enabling sensors and machines within the confines of the factory floor but also intelligent products that have been born. The advanced rule processing can trigger business process automation based on sensed events from requesting a service, to sending spare parts or raising a ticket—it’s not prescribed but is flexible to your agile needs.
The ability to see the digital twin data with the full manufacturing details and service history all in one place in real time provides a next generation platform for your business to grow. While AI and machine learning (ML) are becoming more viable for mid-market manufacturers that do not have the resources like Google to crunch numbers and build experiments, transferable ML models and common use cases are emerging that provide the stepping stones needed to take real advantage.
Intelligent factories building intelligent products making everyone’s life a little easier—sounds like a great future to me.
About the Author
Andy Coussins, Senior VP and Head of Sales, International, Epicor Software. Andy Coussins brings senior international sales, operations and enterprise software industry executive experience to Epicor. In the role of senior vice president and head of international, Coussins is responsible for driving sales, focusing on accelerating company growth throughout Europe, Middle East, and Africa (EMEA), and Asia Pacific (APAC).
Featured image: Lutsenko Oleksandr