Visual Smart Factories with MES Systems

The systemic approach that enables automation systems to move to the next stage can be characterized as a Smart Factory.

It can be characterized as a set of systemic approaches covering production, social business network, integrated product development, data-driven control, supply chain management, digital factory logistics, cloud services, factory design in a digital environment.

Some of the concrete activities to exemplify can be summarized as follows;

- It becomes possible to achieve synchronization with suppliers through integrated collaboration platforms.

- Decision mechanisms can be supported from the very beginning by calculating production costs at the time of design.

- It becomes possible to ensure bidirectional communication between the production phase and the realizations in the design phase.

- Deviations of project predictions in production phases can be monitored and evaluated instantly.

- Optimizations are made using data analytics.

With the developing technology, newer visual applications have started to appear in the market in all areas, applications that aim to make accurate determinations that tell everyone the same thing, without being stuck in one person's point of view.

Visual Factory simply means making factory data easily understandable and always traceable. ANDON screens, management monitoring screens, mobile applications make it easier to follow the work in this area and ensure that the problem can be solved without causing different problems than the first time the problem occurred. Automatic bottleneck detection is one of the main ones. In addition to these benefits, it makes a great contribution to optimizing production processes. Visual management realized with Visual Factory applications makes anomalies visible and aims to create a "talking factory" by enabling the entire team to understand at first glance.

In short, the Visual Factory consists of the whole of the new generation MES applications where factory data is presented in a more visual way by increasing user interaction, that is, the factory almost talks to the user thanks to software and hardware.

Benefits of the Visual Factory to businesses;

Increases team communication.

Increases product quality.

Response times to customer orders are reduced.

Capacity is maximized.

Production time is reduced as losses in processes are identified with their causes.

Unit cost decreases with improvements in operation times.

Corporate workflow processes become easier.

Set-up times are reduced.

All processes that unnecessarily prolong the operation are seen and removed.

With increased communication, the entire team is notified of emergency situations, making emergency intervention possible in the field of occupational safety.

What tools are needed on the way to a Smart and Visual Factory?

First of all, we need to listen to our machines. We can do this with I/O cards or at the highest level with the OPC protocol. Various dashboards, reporting tools and their mobile versions can be used to visualize this data and the analysis made from that data. It is important to design the screens we call ANDON in factories.

In the menu on the left side, there are visual tools you want to see in andon. Chart types, gauge, progress bar etc. In the Queries section, sql queries of the data you will use in the visuals are created. Then you can visualize it by connecting that query to this chart and making other settings from the settings section seen on the upper left side of the bar chart on the screen.

With the reporting modules, the user can create reports as he/she wishes.

In addition, visualizations of your analysis can be as follows.

The trend of the factory, work center, workstation and line-based OEE analysis can be displayed on a monthly, 3-month and 6-month basis.

The results of the analysis can also be seen here. Separate graphs of the components of the A, P and Q OEE analysis can also be displayed.

On the mobile side, sample applications are as follows;

In this way, you can follow every moment of production from anywhere. It can be monitored both from mobile, web and through andons installed in the factory and reports can be generated as desired.

Within the scope of MES, the focal points of smart factories can be expressed as data, action, status and learning.

Data The parameters (tension, temperature, speed, pressure, energy consumption) of the tools (machine, robot, human, tool) used in production should be instantly accessible and process data should be kept in the statistical database.

Action: Establish systems that make decisions or give recommendations by performing data analytics of process data and output data produced. For example, the production plan should be optimized so that energy-intensive production tools (dyeing machine, drying) are used at times when energy is cheap.

In another example, systems that minimize set-up times when moving from product to product should be installed.

Status: Instant warnings and notifications, self-adjustment of machines in case of a work order, stoppages, scrap and the direction of various KPIs, bottleneck detection in line operations, etc.

It is important that the current status can be tracked online;

Monitoring equipment status,

Monitoring of performance indicators,

Monitoring of previous job, next job, current job performances,

Monitoring of work instructions,

Monitoring Production Technical Drawings etc.

Learning: Establish prediction systems by creating scenarios and models with production data. Trend analyzes that think instead of engineering office workers, instant status control of machines should be done through sensors, predicting the failure by predicting with the information from these sensors, and planning should be done by monitoring and predicting machine maintenance plans.

How is automatic bottleneck detection provided in Smart Factories?

A bottleneck is a situation where the entire workflow in production slows down at some point and causes a slowdown in subsequent operations. To give an example, while production is progressing at a speed of 26 units/hour on operations 10 and 20, it drops to 1 unit/hour at station 30, creating a bottleneck. For this reason, it has limited its speed to 1 unit/hour in its next operations. In such a situation, after determining the root cause of the bottleneck on operation 30, the workflow should be relieved by improving the process or eliminating the malfunction.

For example, in a production plan where the workflow should progress at 1000 units/hour in daily life, the occasional operation progressing at a speed of 900 units/hour is also considered a bottleneck.

Automatic bottleneck analysis is a process in which the root cause of the bottleneck is simultaneously communicated to the user. The user is informed instantly and mobile without wasting time by searching in many places. Considering that time is the most valuable resource in production and life, a tremendous gain is achieved.

We have a patent application for this application consisting of software and hardware integration.

How do you think developments in machine learning and artificial intelligence will affect this process?

Factories will face Big Data in the process of becoming a Smart Factory. Because when you follow your production and your machines, your data will expand both on a row and column basis. There will be a need to extract the data that is meaningful for you from this Big Data and to make analyzes that will increase your productivity in your factory from that data. At this stage, machine learning and deep learning techniques will come into play.

So, on which issues may factories need analysis?

One of them is anomaly detection. Anomaly detection is essentially a classification problem. In other words, it is to separate abnormal data from normal data. By performing anomaly detection on the sensor data on the machines, it can be observed whether the machines are working properly. When an anomaly occurs in your machine, which you monitor through sensors while the operator is working, you can intervene before it is too late, that is, before the machine breaks down.

Another is predictive maintenance. In order for production to continue without interruption, our machines must never stop or break down. For this, it is important that their maintenance is carried out on time. For this reason, the maintenance time must be estimated correctly. Performing maintenance ahead of time increases the maintenance costs of the company. If maintenance is done late, it can cause the machine to break down and cause direct loss of production.

Analyses become easier to perform: An analysis can be done in a simple way, perhaps without writing any code, simply by making selections from an interface.

Automating data processing: Before we can do an analysis, the data needs to be preprocessed. For example, finding outliers, eliminating missing values in the data set.

Easier determination of the analysis method: When you take a problem and try to identify a method of analysis, it may not be obvious. In fact, it is often necessary to try several methods and choose the one that gives the best results. Making this easy to do means saving time.

The analysis makes few or no incorrect predictions: When using deep learning and machine learning methods, they need to go through a learning phase. During this learning phase, the model may make incorrect predictions and this creates a financial liability. Developments in this area will reduce such costs.

What kind of data is visualized in Visual Factories? How does it affect capacity, does it increase or decrease? What can be said?

First of all, companies that have just started working with us discover their real capacities and are very surprised to see how much below the real capacity they are actually working by following the losses instantly. Rather than theoretical excel calculations, thanks to the analysis we offer through instant data and real values, we provide businesses with a check-up, so to speak, and since we reveal the problems and improvement points, we ensure that a root cause analysis is made more consciously, and as a result, the place to be improved is improved at once by making a point shot. Thanks to our detailed KPI analyses, other analyses and applications, it becomes possible to digitize and visualize this even in the field of WCM (World Class Manufacturing).

The future is moving towards fully autonomous production with the development of artificial intelligence AI applications. For example, in the factories we have applied so far, rather than reducing the human labor force, it has ensured that it is used in the right place in the right way, and even increased. This will continue to be the case in autonomous systems in the future. Because as the need for physical labor decreases, human beings will continue to exist at more efficient points in production. We are experiencing the transition phase of a very big revolution. As systems evolve, human beings will also evolve and continue to develop for the next revolution. In the near future, production processes and even factories will be animated in 3D, all variations will be simulated and appropriate solutions will be produced, some of the time-wasting indicators will be eliminated, and autonomous systems with VR will appear in interaction with the whole team. With less energy, less time, less raw materials and less carbon emissions, we are in a period where benefits will be experienced even as we transform in many areas.

For example, it will improve the ergonomics of the working environment with robots working together with the personnel.

Newly hired personnel will be trained in virtual reality.

  Another is predictive maintenance.

   In order for production to continue without interruption, our machines must never stop or break down.

   For this, it is important that their maintenance is carried out on time.

   For this reason, the maintenance time must be estimated correctly.

   Performing maintenance ahead of time increases the maintenance costs of the company.

   If maintenance is done late, it can cause the machine to break down and cause direct loss of production.

   Analyses become easier to perform:

   An analysis can be done in a simple way, perhaps without writing any code, simply by making selections from an interface.

   Automating data processing:

   Before we can do an analysis, the data needs to be preprocessed.