It Does Everything on Its Own!

LS Mtron delivers an optimization of injection molding system by interfacing a smart AI technology!

Plastic is sometimes called a ‘gift from God’ not only because it is light but also because it can be fabricated in any shape when being heated.

However, it was difficult to achieve the "autonomy of process" because the quality uniformity of plastic products varies sensitively depending on process conditions. In addition, it takes a long time for changing the once configured process, and it was also deemed that real-time optimization is impossible because of occurrence in actual results and errors.

By the way, there seems to be a solution to that. It is because, last year, Smart Weight Control(M-GA) and AI Molding Assistant(M-QI), which integrated artificial neural network that is the outcome of joint research among LS Mtron Co., Ltd.(hereinafter, called LS Mtron), Korea Institute of Industrial Technology, VM Tech, and POSCO, were presented.

We can meet products from LS Mtron in which offer an optimized injection molding system that allows our customers to enjoy convenience and product uniformity simultaneously.

■ LS Mtron in 2020 has developed an artificial neural network system that recommends the plastic molding processes through joint research -> Smart Injection Products was Presented in 2021

In the last year 2020, the research team of POSTECH(consisting of Professor Roh Joon-seok and a Ph.D. candidate Lee Chi-heon of Department of Mechanical Engineering and Department of Chemical Engineering, Professor Lee Seung-chul and Integrated Course Na Joo-won of Department of Mechanical Engineering, and Professor Park Sung-jin) had been highlighted by developing a system that can recommend injection molding process conditions through the integration of an artificial neural network and random searches together with LS Mtron, Korea Institute of Industrial Technology, and VM Tech as a joint research.

The research team conducted a study that can first learn the relationship between process conditions and final products by employing artificial intelligence and finally find process conditions to satisfy the required quality. First of all, as a result of the learning by obtaining 3,600 simulation data and 476 experimental data from 36 different molds, the research team has developed a weight prediction model, which has the 15 shapes and 5 processes as input values as well as the weight of the final product as output values.

In addition, based on the weight prediction model learned by adopting the transfer learning, a recommendation system, which finds out the optimal process conditions by random search, was developed, thereby achieving an average relative error of 0.66%, as a result of verifying the process conditions recommended by the artificial intelligence model.

Finally, a graphic user interface(GUI) was developed in order to apply it in actual injection machines. This allows an even non-expert in injection molding to configure the process conditions showing an error within 1% of the desired product weight by inputting a piece of shape information for a certain product based on the relevant system.

Meanwhile, LS Mtron anticipates that the relevant industries can further enjoy the convenience since it is able to provide customers with a smart injection molding system by incorporating this technology. M-GA(Smart Weight Control) and M-QI(AI Molding Assistant) from LS Mtron are what those are.

■ Significant reduction in the initial process stabilization time, M-GA(AI Molding Assistant) that exists only in South Korea

If employing the artificial neural network system developed last year, it was expected that even a non-injection expert would be able to obtain uniform results just by inputting the shape and weight of the desired final product.

Through this system, it is anticipated that "unmanned smart factories" can be available in various manufacturing domains that have been difficult to accomplish so far such as plastic injection processes, cutting, 3D printers, and casting.

The importance of an artificial intelligence technology, which can reduce the deviation in production efficiency according to the proficiency of professional molding engineers, has been significantly emphasized when setting the initial molding process for a new molding product.

In this regard, LS Mtron presented M-GA(Smart Weight Control), which dramatically reduced the time required for initial process stabilization by learning and imitating the behaviors of highly skilled molding experts through artificial intelligence technology.

In order to derive the conditions of mass-production products, although a considerable amount of time needs to be invested for skilled workers so far, such a time can be reduced if this product is used.

Currently, LS Mtron is the only company in South Korea that integrated such technologies into injection machines.

An official from LS Mtron stated, “Since M-GA(AI Molding Assistant), which learns and imitates the behaviors of highly skilled molding experts, can reduce the time to reach the good product conditions and recommends the optimal process conditions by supporting the injection molding at the level of molding experts, it is possible to shorten the time required to build a process environment in the early stage, thereby improving productivity”, and added, “A necessity for overseas business travel of high-skilled personnel is reduced because of a decrease in dependency on workers especially in case of relocating the overseas factories”.

It is expected that the product launch of LS Mtron this time can further improve the convenience and accessibility of customers.

■ Smart Weight Control via Real-time Automatic Alteration of Process Conditions, M-QI(Smart Weight Control)

In this regard, LS Mtron presented M-GA(Smart Weight Control), which detects the weight changes of products and automatically alters the conditions to normal weight by using the software of the injection machine. This product, which maintains a constant weight regardless of the state of the crushing material in the injection environment that uses a mixture of grinding materials and others, compensates for a weight change caused by the external environmental changes(temperature and humidity, non-operation time, etc.).

An official from LS Mtron explained, "Smart weight control automatically alters the process conditions in real-time under the identical process in case that weight fluctuations of a product occur due to various factors," and added, "Customers will be able to use it more conveniently."

Meanwhile, LS Mtron is also preparing other things to provide customers with even a smarter injection molding system, which is expected to increase customer convenience and efficiency in the future. If customers meet M-QCF(Smart Clamp Control) and M-QA(AI Weight Control) that are currently being prepared, it is expected that the era of smart injection molding will be expanded fuether.

■ Inquiry: LS Mtron"s Molding Machine Division’s Website www.lsinjection.com