5 September 2024
Outline
A production line is a series of orderly workstations where personnel perform specific production activities following a standard operating procedure (SOP) to transform raw materials into final products.
For example, in an EMS company, the production line involves soldering circuit boards and electronic components, assembling enclosures, and connecting circuits, followed by product sampling inspections. All these operations and inspections take place along the production line.
Production line management encompasses all resources, equipment, labor, and SOPs involved in the manufacturing process. The goal is to ensure products are produced efficiently according to the plan while maximizing resource utilization, minimizing costs, and maintaining consistent product quality. Effective production line management not only enhances production efficiency but also boosts the long-term competitiveness of the business.
Modern manufacturing production lines still rely on human operators. However, manual assembly and inspection by production line workers can often lead to errors that deviate from the SOP, negatively impacting production efficiency and product quality
80% of unplanned downtime is attributed to human error.
(Source: Worldmetric.org.)
As production lines become more complex and variable, coordinating different processes becomes increasingly challenging. Quality control personnel may struggle to track and address abnormalities in each production stage promptly, increasing the risk of undetected defects and delayed corrective actions.
Traditional production lines require manual data collection, consolidation, and analysis before making meaningful adjustments. Delayed data analysis in response to ongoing production issues leads to lower production efficiency and reduced market responsiveness.
To achieve efficient production line management, start with these four key steps:
Currently, many factories still rely heavily on experienced workers’ oral instructions and accumulated mistakes to gradually establish SOPs that align with the production line process. After repeated practice, experienced workers teach new operators the production process. However, this method of standardizing production line processes solely based on “human” input is greatly influenced by human variability and uncertainty. This leads to production fluctuations, inconsistent quality, and failure to achieve the expected production efficiency set by the SOP.
In reality, due to the different habits and experiences of various experienced workers, there are often deviations from the SOP, or even skipping necessary steps in the production process, resulting in unstable product quality. For new operators, unfamiliarity with the processes requires them to repeatedly stop and refer to manuals, affecting production speed.
As the length or complexity of the production line increases, if each workstation is operated based on varying understandings and habits, it will significantly impact overall production line efficiency.
To address this challenge, we have integrated AI vision technology into the production line.
By using AI vision to capture detailed records of the production line, we can analyze small details, such as the precise hand movements of workers, posture analysis, and habitual positions for placing workpieces. Based on this AI image analysis, we can more efficiently establish SOPs that align with the production line processes.
AI vision can automatically detect whether operators are following the standardized SOPs. If an operator deviates from the required steps, the AI alert system provides real-time warnings to prevent errors on the production line.
Additionally, when errors occur, the AI alert system captures the moments affecting production efficiency and notifies the production line manager, who can use these insights to redesign or improve the SOPs. For example, if most workers frequently place parts slightly off the standard location, the AI system identifies this issue and notifies the manager. The manager can then decide whether to retrain the workers or further optimize the SOP to better fit practical operations
After implementing the AI vision system for four weeks, the production line saw a 5.2% increase in UPH and over 5x return on investment (ROI).
Many production lines today still rely on manual data collection. Industrial Engineers (IE) record operators’ task times manually with stopwatches, trying to identify the root causes of inefficiencies. However, this manual method is time-consuming, and the data quality is often unreliable.
Not only is it time-consuming, but the quality of manually collected data is also problematic. Manual recording methods struggle to capture a comprehensive and objective record of what happens on the entire production line, leading to biased data analysis.
To overcome this challenge, deploying AI vision technology on the production line offers a more effective solution.
For example, in an electronic component manufacturing plant, IEs initially believed that the cycle time for a specific workstation was the longest. However, after using the AI vision system, it was discovered that operators at that station were not spending as much time as expected on assembly tasks. Supported by accurate data, managers can identify production line issues objectively and establish the most suitable process standards, avoiding ineffective evaluations based on misconceptions.
The AI vision system collects continuous video footage of the entire production line, capturing every activity at each workstation. With complete visual data, AI data analysis can interpret the data and detect abnormal production processes, alerting IEs to help monitor production conditions at every workstation.
Both anticipated and unanticipated downtime can consume up to 10% of production line time.
(source: Forbes)
Production lines face a diverse range of issues daily. Human-related hazards, equipment failures, and material supply problems are three common issues on the production line. Immediate shutdown or suspension of the production line is the first step to prevent problems from escalating. In most cases, Industrial Engineers (IE) can only enter the production line after an anomaly occurs to identify and correct the root cause. However, during the downtime while waiting for IE to address the issue, the factory still bears associated costs, including labor and fixed equipment costs, wasted material expenses, and even penalties for delayed deliveries.
To minimize losses and get the production line back up and running quickly, we are integrating AI visual inspection systems into the production process.
AI visual inspection systems offer real-time alerts, traceability, and remote management capabilities. With 100% data and image support, the efficiency and quality of IE work improve significantly. AI visual systems can instantly report errors at specific workstations, reducing the time IE needs to identify problems. Additionally, real-time analysis of worker movements helps IE identify operational bottlenecks.
For example, in a motorcycle manufacturing plant, despite having a MES system and various connected tools, overall production efficiency and quality remained suboptimal. To obtain a complete production history, the factory introduced an AI visual system. Comparing the AI visual system with the old MES revealed that there was often a few seconds’ discrepancy in the captured production cycle times. AI visual inspection traced this discrepancy to a practice where operators, instead of clicking the system’s completion button after tightening two screws, clicked it after the first screw for convenience, then tightened the second screw.
AI visual systems provide a transparent and comprehensive production line history, optimizing capacity and improving overall efficiency.
The AI vision system ensures a workstation maintains a 95% yield rate and a 97.6% pass-through rate.
In the factory environment, the production line faces different situations daily, with “humans” being the most significant variable. To maintain efficient production line operation, IEs relying on experience and manual records to analyze and detect production line process issues while tracking multiple production lines is an immense challenge.
In precision manufacturing, the value comes from complex processes and fine craftsmanship. Maintaining a high yield is crucial for ensuring high-quality products. However, the detailed steps in the production process increase the potential for errors. Even minor defects can result in a product being deemed unqualified. Any lapse in detail control leads to costs from rework, material waste, and potential damage to brand reputation.
However, without sufficient data support, IEs often lack confidence in validating improvements to the SOP.
The AI vision system can comprehensively record all factors affecting production efficiency. With real-time AI analysis, product managers can objectively review inefficient SOPs. With visual data support, other management levels, such as factory managers, can also monitor production line activities and track the effectiveness of subsequent improvements.
From these four steps to efficient production, it is evident that “humans” are the most significant factor affecting the entire production line process.
Whether it is an operator deviating from the SOP or relying solely on IEs for production line management, “human” uncertainty impacts both execution and management efficiency.
Therefore, utilizing production line management tools can help achieve better results in efficient resource utilization, production efficiency enhancement, and production cost control.
Many factories still rely heavily on experienced workers’ oral instructions and accumulated mistakes to gradually establish SOPs that align with the production line process. After repeated practice...
Read more
Many factories still rely heavily on experienced workers’ oral instructions and accumulated mistakes to gradually establish SOPs that align with the production line process. After repeated practice...
Read more
You’re all set!
Keep an eye on your inbox for more information.
Oops! Something went wrong.
Please try again.
Oops! Something went wrong.
Your subscription wasn’t completed. Please try again.
You’re all set!
Stay tuned for industry insights and the latest from PowerArena.
