Guide to lean production for shopfloors with AI and automation integrations
An overview of lean manufacturing approaches and how automation and AI tools can be integrated by shopfloor managers to achieve tangible results immediately.
Lean production, or lean manufacturing, is based on the idea of working to continuously improve processes and eliminate waste. Waste includes time, money, and materials – anything that increases your cost of operations and does not add value to the customer. How can shopfloor and production managers make strategic decisions when investing in new technologies like AI and cobots? This guide will give an overview of lean production principles as well as an overview of where automations can be integrated effectively in a cost-effective and low-risk environment to immediately bring value to your business and customers.
What is the value of lean manufacturing techniques?
Many of the concepts of lean production come from Japanese post-war production processes, made famous with the Toyota just-in-time production line. Lean manufacturing enables teams to deliver higher quality products at significantly lower costs. It is an approach to manufacturing that has many different tools you can use, depending on what your needs are.
You do not need to apply everything. You also will not need to integrate every new AI solution or technology. Instead, take what works to increase the efficiency, effectiveness, and profitability of your manufacturing operations across specific topics. Below, we will go over some commonly understood lean production terms, and where digitalisation, AI, or Industry 4.0 robots can support operations.
Aiming for Continuous Flow
The goal of lean production is continuous flow, where everything flows smoothly through the stages of production with minimal (or no) buffers between steps. When resource planning, planning for buffer is realistic for estimating delivery times. The goal of continuous flow is to reduce this buffer (waste) as much as possible through standardisation, just-in-time delivery, and automation that does not compromise on quality.
5S
One of the basic principles of lean production is 5S, which aims to eliminate waste that results from a poorly organized work area. By only having what is needed, and finding out the most efficient setup and arrangement of the tools, your staff can save a few seconds or minutes, which builds up into hours and days over the year.
5S refers to what to do with a work area:
Sort: eliminate what isn’t needed (declutter)
Straighten: organize remaining items (reduce chaos)
Shine: clean and inspect work area (ensure tools can be used)
Standardize: write standards for above (share best practices)
Sustain: regularly apply the standards
Improvements for 5S may not immediately need technology. Instead, improvements can be made through the principle of gemba, which we explain below. Through careful observation and optimisation, improving the workspace itself will have effects. At the team level, this can be manually tracked through time to produce a piece or complete a task before and after the 5S approach is applied. At production-scale level, it is possible to collect the data from all shopfloor teams to analyse for further optimisations.
Some principles behind lean manufacturing
Because lean manufacturing comes from post-war manufacturing in Japan, made famous by the Toyota Production System. Many of the terms used are borrowed from their original Japanese phrases, which we outline below:
Gemba (現場): The idea that all management should have spent time on the plant floor, where the real action occurs. It promotes a first-hand understanding of real-world manufacturing issues through first-hand observation, by talking with plant floor employees, and having experienced the space itself.
Heijunka (平準化): Also from the Toyota production system, this is a production scheduling that manufactures smaller batches through sequencing (mixing) product variants within the same process to reduce lead times and inventory.
Hoshin Kanri (方針管理): 7-step process used in strategic planning in which the goals of the company (strategy) are communicated throughout the company, become plans in middle management (tactics) and then put into action on the production floor.
Jidoka (自働化): It is not just automation, but specifically equipment designed for partial automation, which improves operational efficiency, without the cost and risk of full automation. Workers are freed up to monitor multiple stations, reducing labour costs, and the process can automatically stop when defects are detected.
Kaizen (改善): Often translated as continuous improvement, where employees work together proactively to achieve incremental improvements throughout the manufacturing process. It combines collective talents within a company and gives workers the agency to contribute to process improvements.
Kanban (かんばん): Regulating the flow of goods with suppliers and inside the factory with automatic replenishments through signal cards to show that more goods are needed. It keeps track of actual need, reduces overproduction and inventory space.
Poka-yoke (ポカヨケ): Error detection and prevention built into production processes with the aim of having 0 defects. As defect detection gets costlier with each stage of production, designing checks at every stage increases the chance that something gets caught early, thereby saving downstream costs.
Muda (無駄): The Japanese term for waste. The main goal of lean production is to reduce this.
Lean production concepts for planning and strategy
We have broken down lean production concepts into planning and strategy, which requires time for thinking, data gathering, and documentation of ideas to share. Below are the things that you as production and shopfloor managers can plan, even with pen and paper. After having a better understanding of goals, performance indicators, and what baseline data there is, you can track the improvement of processes, as well as look for tools to help automate the process later on.
Setting effective manufacturing KPIs (Key Performance Indicators)
KPIs are numbers-based goals that are set to encourage progress to reach critical goals. After setting the KPIs, there are usually check-ins to track the progress towards the goal. Well-set targets are extremely powerful drivers and motivators, while unrealistic targets can be demotivating.
Shopfloor and production managers can set KPIs that help show their contribution to the company’s strategic goals. Examples can include:
Increased overall output to reach the company’s strategic growth goals
Increased efficiency in production, as measured by days, human hours, cost
Quantifying and showing reduced waste (for example, OEE)
Ensure that the KPIs can be directly influenced by plant floor employees, so they can drive results. KPIs are a common metric used in for human resources to mark employee performance and there are entire HR software solutions such as Leapsome, or BambooHR that include feedback and performance assessment in addition to payroll. However, you can get started on KPIs with a simple document or spreadsheet template every month or quarter that you assess with your staff.
Mapping the Six Big Losses
The Six Big Losses are six categories of productivity loss that all shopfloor managers and production lines need to consider:
Breakdowns
Setup/Adjustments
Small Stops
Reduced Speed
Startup Rejects
Production Rejects
Collecting data on these factors will help you make better delivery estimates, put a cost on poor quality, and advocate for investment in improvements. An example of this is the overall equipment effectiveness (OEE) calculations, which can be found online. This helps shopfloor managers advocate for equipment upgrades, such as why a cobot would be more cost effective within a year for a specific task.
Analysing bottlenecks and root causes
An important part of lean production is identifying inefficiencies. Bottleneck Analysis looks for the step in the manufacturing process that limits the overall throughput. By releasing the bottleneck, the performance and delivery time is immediately improved: it aims to strengthe the weakest link in the manufacturing process.
Root Cause Analysis is an problem solving approach that digs into the underlying problem instead of applying quick fixes to the immediate problem that is seen. A common approach is to ask why five times: Why did this happen? Why is the answer to the first question happen? So on, and so forth. Each why should get deeper into the root cause.
Value Stream Mapping
Value Stream Mapping is a tool used to visually map the flow of production. Shows the current and future state of processes in a way that highlights opportunities for improvement. Value Stream Mapping exposes waste in the current processes and provides a roadmap for improvement through the future state.
Setting SMART goals
Even though experience and intuition can go a long way, putting them in a framework ensures they meet business goals and helps us communicate with colleagues. SMART Goals are: Specific, Measurable, Attainable, Relevant, and Time-Specific. Taking the time to write down what the goal is, how success is measured (through a KPI), with a realistic target against baseline data, which business goals it contributes to, and by when it should be achieved, helps align everyone in the team and set expectations. This can be done with a simple document.
Processes that support lean production
Although all lean production is oriented towards process optimization, the concepts that we’ve outlined below are focused on execution and operations.
Setting up a PDCA (Plan, Do, Check, Act)
PDCA is an iterative methodology inspired by scientific methods for experiments. PDCA helps implementing improvements in a systematic way by ensuring that what is planned has a measurable impact:
Plan: Make a plan and expected results (the KPIs)
Do: Execute what you planned
Check: See if your KPIs, or other benefits, were achieved
Act: Assess whether the impact is worth continuing, or if another solution is needed
Standardized Work
Standardized work refers to documented procedures during the manufacturing process that are best practices (including the time to complete each task). These “living” documents should be easy to update, and scheduled for regular reviews (monthly, quarterly, annually) by responsible persons.
Standardized Work builds on the PDCA approach to eliminate waste by documenting what works best (best practices) and ensuring they are consistently applied as the standard procedure.
Checking your Overall Equipment Effectiveness (OEE)
Overall Equipment Effectiveness (OEE) is a framework for measuring productivity loss for a manufacturing process based on these three categories:
Availability (such as downtime)
Performance (such as slow cycles)
Quality (such as rejects)
Knowing your OEE gives you a baseline to track progress in eliminating waste from a manufacturing process. Achieving 100% OEE means a perfect production process (manufacturing only good parts, as fast as possible, with no downtime). Having routine checks for OEE and setting incremental improvements over time will help cut your bottom line operational costs.
How to integrate AI and Industry 4.0 into lean production
Manufacturing as an industry has been using technology to optimize processes for years. The difference now is the sheer number of technology solutions to augment lean strategies with Industry 4.0, and the increasingly accessible prices of solutions. This can include Software-as-a-Service (SaaS) products, collaborative robots (cobots), machine learning and data analytics solutions, and AI-powered solutions like automated quality control. Each area of optimization for production now has its own myriad of options. For example, you can check out our resource for machine vision systems and solutions providers.
We recommend researching solutions based on the specific area you are trying to solve, rather than trying to learn about every latest new technology. You can begin by considering automation solutions in each of the below categories:
Production: Specialised or general robots can increase throughput and reduce injury for repetitive or high-stakes tasks.
Quality control: Machine vision systems help automate quality control, improving efficiency, reducing human error
Supply chain management: Machine learning applied to your inventory data optimises stock levels, demand forecasting, and replenishment planning
Predictive maintenance: AI-based equipment monitoring to help your OEE
The solutions mentioned are not exhaustive. In general, solutions can also be bucketed into smart sensors and devices for error proofing processes, digitization and automation for standardized work, data collection, and quality assurance for finalised goods. With the technologies available today, it is worth asking if any repetitive task can be replaced or augmented with a machine – either a robot to lift and sort, a sensor system for automated quality assurance, or an app for automatic data analysis.
Assessing AI and technology providers for manufacturing
In prior decades, industrial-scale technology was the domain of a handful of companies. Many solutions were proprietary, came as a complete package, had large upfront costs, and were expensive to train staff for and maintain. Now, manufacturers can choose which providers they want to solve each of their specific cases. There is no need to be locked into one provider’s system for everything on the shop floor.
Software solutions can come as full-service packages, but also as monthly subscriptions based on users or usage rates. Likewise, even robotics solutions have different tiers and levels of software support.
We recommend that production managers approach solutions providers with the following ideas in mind
Is it helping me with production, quality assurance, supply chain, equipment maintenance, or something else (like worker safety)?
Which of the Six Big Losses is it helping me with, and how much will it save me?
Can the solution be tested on-site before you commit and how fast?
Is the solution easy for staff to use and adjust or does it require special training?
Are there specific systems the solution needs to integrate with?
What are the ongoing costs of the solution?
Is there flexibility for me to start small or is there a minimum purchase?
After doing research, meeting with different providers, and seeing the demos for their solutions, we recommend creating a KPI and setting a SMART goal for the solution that you choose. Ideally, you can try it with one team before scaling to different shifts or other areas of operation. This way, you can learn and adapt in months, not years.