As one of the most important materials in global manufacturing supply chains, aluminum has been subject to changing regulations. For example, in 2017, the USA raised tariffs on imported aluminum by 10%. This had a major impact on domestic production, which grew by 67% one year after the tariff was imposed. This is a significant increase considering that, in the period between 2010 and 2017, 18 smelters shut down across the country. With such a steep increase in demand for domestic aluminum, US-based producers are having to invest billions of dollars in new mills, which take years to be completed.
We are expecting a further increase in aluminum demand, which is driven by both the automotive industry’s need for the material as well as the sustainable properties that it has:
- Aluminum is a lightweight replacement for steel and most automotive producers are adopting it to create lighter and more fuel-efficient vehicles.
- Aluminum is easily recyclable, which helps manufacturers achieve their sustainability goals.
Whilst building new mills creates a stepped increase in aluminum production capacity, manufacturers can ensure they continue to meet the increasing aluminum demand with their current resources by maximizing existing plants’ output quality. To achieve this, aluminum producers can look to build or identify and implement state-of-the-art technologies that will optimize their production processes and bolster growth.
Optimizing aluminum milling processes using the data already available
In the automotive industry, precision is crucial in maintaining vehicle safety, reliability, and quality. Inconsistencies in the structure or smoothness of an aluminum sheet may require significant rework for painting or assembly. Even if aluminum is becoming the go-to material for carmakers to build lightweight and fuel-efficient vehicles, new car models require even higher levels of precision to meet demands in terms of performance and aerodynamics.
Today’s specialized aluminum production equipment can control thickness within a few thousandths of a millimeter. This would not be achievable without considerable investment in sensors and analysis platforms that can help manufacturers monitor their production processes. Modern control systems for aluminum production constantly measure and store process data and can record more than three gigabytes of data daily during mill operations. This volume of data is one of the reasons manufacturing as the commercial sector produces the most data year to year – and positions the manufacturing environment as ideal for AI solutions.
However, the volume of data poses its own challenge when creating value from it. Traditional methods simply aren’t well suited to process that amount of data. While the data infrastructure is steadily becoming commoditized through more and more robust IoT gateways and sensors – value can only be found if action is taken from the data. Advanced controllers and artificial intelligence can support by suggesting a set of actionable insights to optimize the production process.
Aluminum mills are very familiar with level 1 controllers to assist in rapid adjustments in the rolling process to arrive at the desired product thickness. However, this data often is not used outside of that context due to its volume and velocity making traditional analysis upon it very difficult.
AI-powered solutions have the capability to ingest that quantity of data, and can manufacturers on their journey towards autonomous manufacturing, and can work in cooperation with existing level 1 controllers. The beginnings of automation were marked by shifting physical work from humans to machines. Now, we’re starting to outsource low-level operational decision making from engineers to AI-powered systems. As we continue developing more sophisticated AI systems, we will move closer towards an optimized manufacturing process that requires little to no human intervention.
All our efforts around plant digitalization – through automation, data gathering, and data processing – are aimed at achieving higher quality output, predictable results, and an improved rate of high-quality products with fewer resources and raw material expenditures. For example, AI solutions can aim at lowering the industry standard of 18% of lost material in the hot rolling process by minimizing the number of edge cracks and surface defects through carefully provided prescriptive actions passed to the control environment or the operators of the machinery.
We’ve been observing how manufacturers who share this vision and have been adopting Industry 4.0-level technologies are already gaining a competitive advantage. Manufacturers can no longer apply the paradigm of “producing more for growth’s sake”. Industry-leading producers that leverage the correct technology can reconfigure their setup with ease and adapt accordingly.
AI algorithms are creating today’s ‘smart plants’, which analyze data to learn the interactions between process set points and the resulting products. A smart plant is aware of its own state and can identify adjustments based on its knowledge of the process. They rely on Expert Execution Systems (EES) to perform AI-driven analyses that guide process control changes.
Creating AI-powered Smart Plants
Expert Execution Systems are a stepping stone in the journey towards autonomous manufacturing. EES are used to prescribe optimal setpoints for enhancing plant efficiency and apply corrective actions to parameters that negatively affect the quality of output. The AI systems which power EES can make intelligent cross-parameter associations between configurations and results.
EES is analogous to a network of connected vehicles that are aware of the state and intentions of all surrounding cars, making it easy for them to coordinate efficiently and avoid accidents.
As manufacturing processes have many moving parts, they require an EES to unify their actions. Without such a system, when operators make changes to process set points, they don’t have visibility of the thousands of data points that are produced upstream or downstream. Poorly configured set points early in the production process will have a knock-on effect on the output quality. Even if the processes are closely monitored and defective products are identified as soon as they happen, manufacturers will still suffer losses in terms of unusable products and lost production time.
This is why real-time data is not sufficient for optimizing processes and achieving up to 0% defects. For modern requirements, we must move away from a reactive approach and move towards a proactive one as pre-emptive actions can stop defects from taking place altogether.
Shifting from a reactive to a proactive approach by analyzing data sources, identifying ideal control bounds, and guiding operators on the optimal process control changes requires an intelligent system that is aware of its own state.
DataProphet’s AI solution is highly applicable for manufacturers today and provides pre-emptive actions to help move towards an intelligent, data-driven approach to plant optimization.
DataProphet’s Enterprise-ready Expert Execution System
By powering their Expert Execution System with a leading AI solution, DataProphet’s PRESCRIBE delivers an enterprise-ready application that analyzes relationships across multi-step processes to provide operators and machines with pre-emptive process control changes.
DataProphet PRESCRIBE can connect to existing process and quality databases which may be various Roll, Pass & Planning Databases, Process Historians, MES systems and quality databases, and even manual inspection logs. DataProphet’s AI solution has been designed to help users capitalize on Industry 4.0 technology. As an Expert Execution System, PRESCRIBE’s capabilities are a critical element in enabling aluminum producers to achieve fully autonomous and intelligent manufacturing.