In this article, you can learn about:

• Use cases where Project Bonsai has proven to be valuable
• Industry examples of optimization opportunities
• Key benefits delivered
• How Tietoevry works with Project Bonsai
• Examples of project deliveries

Introduction: What is Project Bonsai?

The move towards autonomous industrial systems is often seen as being both prohibitively complex and expensive for many manufacturers. While automation has been around for years – and is the proven ‘staple’ for controlling processes – the decisions and resulting actions will always be made based on a pre-defined set of rules. This means that continuous optimization can be challenging, especially if the process inputs and environment can be variable.

Autonomous AI makes a world of a difference. It learns and can adapt to environmental changes and thus evolve as needed to ensure that optimization is always in focus. This ability to ‘learn’ results in the potential for huge optimization possibilities as it explores new ways to solve problems and prevent unwanted results. In other words, Autonomous AI can both optimize existing controls (and operator capabilities) and also hunt and explore for new ways for any given industrial process to be improved. It is this ability to hunt for new solutions that makes it so potentially invaluable in terms of cost savings, productivity, and human capital capability.

The key difference between traditional AI technologies and Project Bonsai is that it is a complete toolchain and a platform for building Autonomous AI with a low code approach. This means that your systems and control experts are now able to create their own Bonsai Brains with just a little training, and without the need of Data Science expertise.

Use-cases showcasing how you might benefit from Microsoft Project Bonsai

In order to ensure a high level of optimization success, the selection and definition of the use-case is critical. Below are just some examples of use-case characteristics where a Bonsai Brain can bring quick wins and measurable results.

	High level of human-operator dependency in processes In any instance where the operator needs to change machine/process settings according to e.g., raw material feed, quality requirements or environmental changes, AI driven autonomous automation can have a significant impact. A Bonsai Brain can improve upon and complement the operator’s performance autonomously whilst also ensuring that you always have an excellent performance standard regardless of the skill level of the human operating/supervising the machine/process.
	Simulation data available The key advantage of a Bonsai Brain is that it simulates learning. This learning can unlock not only skilled operator-like performance but also the opportunity to exceed the performance of your best human capital. The more the Brain learns in Simulation, the better the end performance is when put into a production environment. So, if your machine/process can be simulated, then it is possible to apply a Bonsai Brain to enhance and compliment your optimization capabilities.
	Systems and processes where you have gained a lot of process knowledge Human expert knowledge is invaluable in order to optimize. Much of this knowledge can be used to fast-track and accelerate the learning process for a Bonsai Brain, which in turn yields faster and more accurate results for you. Feeding the existing process expertise into the simulation environment ensures that the nuances of any given process and its parameters are understood and applied in the right way. It also means that this knowledge and capability will not remain only in the minds of your skilled and experienced personnel, but rather can be replicated and scaled across your processes.
	Processes and industrial applications, where optimization can bring significant savings potential Across the Nordics, we have a long history of automation in multiple different industries. Many of these industries rely on optimization as a means of competitive advantage and meeting sustainability targets. Sometimes optimization benefits may result from optimizing one process step or machine, and at other times it is a complete end-to-end process optimization that yields the desired results.

Industry examples of optimization opportunities

Pulp & paper industry

• In-feed and out-feed optimization for Bulk Material Handling applications such as conveyor belts and screw drives
• Digesters for pulp production
• Wash Presses for pulp purification

Plastics applications

• Die extrusion processes
• Plastics moulding
• Vacuum forming

Metals

• Smelting furnaces
• Chemical extraction and electrolysis process

Energy production and management

• Wind farm optimization
  • Hydro and turbine optimization
  • Load management
    
    

Key benefits of Microsoft Project Bonsai

The benefits of Project Bonsai are numerous and varied, but let’s focus on four key areas:

Compliment your best-skilled Human Operators

Imagine a world in which all of your processes are operated by people with the same excellent skill level as your best operator. A Bonsai Brain on your machine/system can make that a reality. It is getting harder and harder to both train and retain highly skilled operators, especially those with many years of experience and expert knowledge. A Bonsai Brain can be an excellent complement to your workforce and can result in huge savings via autonomous AI optimization.

New and innovative ways to optimize

Autonomous AI can hunt for new solutions to a problem by trying new and different approaches and solutions millions of times over. It can iterate solutions in simulation and learn based on a set of defined rewards. This means that it can discover better ways of solving a challenge over time and predict when to apply a different strategy according to a changing environment (raw material quality, temperature, pressure, etc). It is this ability that really makes this technology so invaluable and exciting.

Low Code and designed for engineers

Project Bonsai virtually eliminates the need for pure Data Science when using an AI based approach. It is low code, meaning that with a relatively small amount of training someone with good process and system expertise can design, build, simulate and implement a Bonsai Brain without having to rely solely on a team of data scientists. This also facilitates a large element of independence and self-sufficiency for you as an organization in terms of the rapid development of new capabilities.

Reduce risk and accelerate the development

Bonsai learns in a simulation environment which in turn results in low risk. The brain can practice and learn safely without the need to deploy directly to a production environment. Only when the brain proves it can improve upon the old/current process can it be deployed accordingly. This also has the side-effect of dramatically reducing the lead-time for new AI use-case development and testing. Further to this, the brain can be designed to work both with your operators or also on its own, making it highly scalable.

How do we work with Microsoft Project Bonsai?

Tietoevry and Microsoft have a strategic collaboration with Project Bonsai. We are Microsoft’s partner of choice in Northern Europe for the delivery of AI-powered solutions for autonomous industrial control systems based on Microsoft’s Project Bonsai low-code Platform. The partnership will help the industry to capture the benefits of AI-driven automation.

We have developed a simple and pedagogic process for fast-tracking your Bonsai project:

Click image to enlarge

We place high importance on defining the business case and ROI before any Bonsai Brain design and development happens. Thereafter we focus on the Brain Design with our Bonsai experts and in tandem arrange for the training of your Subject Matter Experts.

Simulation as expressed above is a crucial stage, and we take care to ensure that your simulation environment and data will fulfil the necessary criteria to guarantee the best possible Brain design and outcome.

Project Bonsai Proof-Points

Whilst Bonsai is a new technology, it has proven itself successful in many cases. Some examples include:

• Chemical process optimization: Continuously stirred tank reactor – 74% Improvement
  
  
• Food industry process optimization: PepsiCo extrusion optimization
  
  
• Optimizing machining processes and operators: NOV (National Oilwell Varco) reducing operator training times
  
  
• Improving mining operations by maximizing extraction, minimizing equipment failures, and reducing downtime

Read more about Microsoft Project Bonsai industrial automation