Background Image

Machine Monitoring :

Why computer vision is the best solution

BY KHIZER HAYAT

4 MINUTE READ

Automation in manufacturing refers to using technology and machines to perform specific tasks without the need for humans to intervene. The goal of automation is to increase efficiency, productivity, and accuracy in the production process, reducing manual labour and minimizing the risk of human error. Automation is most frequently used to automate repetitive or dangerous tasks, which frees up human labour to focus on more highly skilled tasks, thus allowing for a more efficient and cost effective manufacturing process. According to the International Federation of Robotics, there are around 3.5 million industrial robots in operation around the globe.

With the increase in automation it is imperative for machines to be monitored to make sure the efficiency and productivity gains are being realized. Current methods of monitoring machines especially the complex human-machine interactions fail to properly identify the key components of production downtime. i-5O clients use computer vision for a better approach for monitoring machine metrics to provide more granular insights into machine productivity as well as the complex interactions between humans and machines.

Humans still play a critical role

While automation continues to increase in manufacturing as companies look towards the most streamlined and efficient process possible, complete automation is impossible and humans still play an integral part. In fact, 70%+ tasks in manufacturing are still manual. This is primarily because machines can’t replicate human adaptability, cognition and dexterity. Stanley Black and Decker's failed automation experiment shows why humans are still critical in manufacturing.The importance of humans in manufacturing can also be seen from the fact that employment in manufacturing has stayed stable or increased in some of the most automated countries in the world including the US, Japan, Germany, and South Korea over the last 10 years.

Increased human and machine synergies

Not only are humans crucial for certain complex manufacturing processes, but they increasingly interact with machinery by setting up and performing machine changeover within the manufacturing line. For example, machined parts require humans to set them up and perform changeover. While humans are crucial for this, there is also more room for error; improper set-up can cause the part to be defective, resulting in scrap or rework. This causes financial losses for the organization. Long changeover times can impact a company’s throughput which can impact their revenue as lower throughput would mean that they can’t produce enough to meet market demand.

Challenges with current monitoring solutions

Even for organizations with a high degree of automation it is imperative to track their machine productivity in order to make sure that they are operating at the highest level of efficiency. The most common solution currently is for engineers to pull data directly from the PLCs (Programmable Logic Controllers) in the machines and then massage the data to generate insights to monitor the machine productivity as well as the OEE (Overall Equipment Effectiveness).
The process of extracting data is invasive and requires a tremendous amount of engineering time. The raw data itself needs to be thoroughly engineered using a tool like SQL to extract valuable insights and then visualized using a BI tool like Tableau or Power BI. Furthermore, the data doesn’t help identify whether the downtime was due to changeover, unplanned/planned maintenance or lack of parts available for production. This information would be manually gathered via time or motion studies.

Using computer vision for machine monitoring

Using computer vision, machine monitoring can be done in a simple non-invasive fashion by using cameras to monitor the status of the machine. A camera is placed near the machine(s) to be monitored and it measures the productivity of the machine as well as the human-machine interaction to identify sources of downtime in the operation(s). This data can then be sent to a visualization tool for post-analysis and/or to send real-time alerts to shop floor personnel to identify problems immediately.

Why monitoring with computer vision is better

Using this technique i-5O has successfully helped its clients gather data about machine uptime/downtime as well as the reasons behind the downtime e.g. lack of parts, long changeover time, unplanned maintenance etc. One of i-5O’s clients (a global leader on automotive Exhaust Systems, Thermal Insulation and Advanced Materials components) was able to reduce their planned downtime by 50% after using the data from our platform to improve their maintenance operations. Thus, computer vision is a great solution for machine monitoring tasks as it can not only gather machine data but also data around the workers interacting the machines to provide a complete picture of the operation.

Try it for Yourself

Schedule your 1st Free Consultation and Start your AI Journey Today