In my last post on the future of cutting tools, I discussed a vision and roadmap document that I created and refined over the years. This roadmap was created by imagining what perfection (or utopia) looked like utilizing what we know to be technically possible today. It was a glimpse into a futuristic system for handling cutting tools in machining operations including robotic automation, copious amounts of data, and artificial intelligence in the form of machine learning. Today, I’m going to describe a vision for a futuristic production management system where data silos do not exist, predictive analytics provide glimpses into the future, and algorithms optimize throughput to balance costs and demand.
The Current State of Production Scheduling
Frequently, production decisions are made with partial information and what may seem like the “optimal” solution on a local-level creates costly disturbances on the macro-level. For example, manufacturing equipment needs preventative maintenance but production schedules typically do not include machine downtime for these events because of lack of visibility to them. This often delays preventative maintenance to permit production to continue operating. If this continues for too long, the equipment may fail prematurely and cause significant cost disruptions. There is also no visibility to the health of equipment to detect potential failures or changes to the production schedule. Some common production scheduling problems include:
- Indirect materials (cutting tools and other consumables) may run out of inventory.
- Personnel may become ill and cross-training may be limited causing the process to operate less efficiently.
- External influences like demand, weather, and supplier issues.
- The schedule could also be managed in a spreadsheet with limited data visibility.
The Future of Connected Data Systems
In the future, all data systems will seamlessly share information with one another. There will be a contiguous data record that exists throughout sales, manufacturing, customer monitoring and diagnostics, services, and product development. Manufacturing production systems will have immediate visibility to sales, product design, supply chain (internal and external), personnel availability, personnel qualifications, and customer issues. Inside the shops, the individual spreadsheets to manage production schedules will not exist and the state of all production resources, equipment, personnel, indirect, and direct material will be known at every moment in time. All data will be linked with proper context.
For example, a component failure on a product at a customer’s site will be linked to the manufacturer's equipment and personnel inside of a shop, the supplier who supplied the raw material, the quality checks throughout the manufacturing process, the design of the product and the analysis and testing that was conducted on the design, and the custom order that drove the product configuration. It eliminates any information silos and provides full transparency throughout the supply chain.
The Future of Predictive Analytics and Production Optimization
In the future, as manufacturing equipment ages, sensors will monitor the components of the equipment and analytics will run on the sensor data to determine the “health” of the equipment. As the equipment’s components reach the end-of-life, the data systems will be able to query the maintenance inventory to see if the necessary replacement components are on-hand. If the parts are not on-hand, the part supplier will be notified and a requisition to purchase the parts will be created. Then production risk will be calculated and machine downtime will be added to the schedule at a time where risk is as low as possible.
With regards to direct and indirect materials, the data systems will predict potential disruptions by comparing the demand based on production schedule with current inventories and scheduled delivery dates from the suppliers. As the product demand fluctuates, the manufacturing data systems will notify supplier data systems ahead of time so that suppliers can prepare to meet the new demand accordingly.
In the case of a machining/fabrication shop, many of the processes are developed using general purpose equipment. That equipment could be utilized for multiple different operations on multiple different products. In the essence of being “flexible” many shops have developed redundant processes so that if one machine goes down unexpectedly, the production could quickly be shifted to a different machine. By coupling asset condition with material inventories and resource staffing, routings can be dynamic and the system will be able to analyze multiple routes to complete a part and make suggestions to minimize overtime, outsource, and/or delivery penalty costs.
There will also be significant opportunities from using predictions or forecasts from sources beyond traditional manufacturing sources. For example, by analyzing production schedules in conjunction with information like flu outbreaks, the data system can begin to predict the probability that personnel will be out of the shop and assess the risk of missing production deadlines and the costs associated with those risks. By making those assessments, the system can identify where cross-training may be necessary to reduce the overall risk and costs. By including weather forecasts and supplier’s geographic locations, the manufacturing data system can predict which product schedules may have a higher probability of being impacted due to weather events that disrupt the supply chain, like a blizzard in the South.
This is another glimpse into one manufacturing geek’s view of how production scheduling may function in the future. It is very exciting to think about how shop management decision making can be augmented by the data systems and analysis. There are numerous possibilities and this does not even come close to representing all of them.