Industry 4.0 and Lean

Industry 4.0 is a natural extension of lean, While the term "Industry 4.0" is rather new, the exploration of cyber-physical systems has a long history.

Given the large number of articles about Industry 4.0, IIOT (Industrial Internet of Things), Cloud Computing, AI, Machine Learning, Digital Twins, RPA (Robotic Process Automation), etc. and the large number of companies springing up to sell you specialized software or services and equipment, it's easy to be confused over where to start. Individual tools may be tempting to apply. Be careful that they do not become silos of optimization. Speeding up one function of a production process doesn't improve the system output unless it is the bottleneck process.

An easy way to recognize the potential of cyber-physical systems in manufacturing is to look at the transition of the taxi industry. There were a multitude of barriers that we took for granted when you needed a taxi. Some areas of a city would have enough taxis roaming around that you could hail one from the curb. The other option was finding a phone (attached to the wall or phone booth), finding a phone book (those were the paper things you used before Google to find a phone number), looking up in the phone book the number for the taxi service, dialing the number (putting money into the payphone), waiting for an answer, telling the dispatcher where you were and where you wanted to go. The dispatcher would find an open taxi to send to your address (be ready to wait 30 minutes), you would reconfirm the location to go, they would start the meter (the fee was based on miles and time, which ever was more), when you arrived you would pay in cash whatever the meter read. These barriers we just accepted as part of the system.

Compare the old taxi methods (all the steps) to today's options. You take out your smart phone, open ride hailing app, enter location you want to go, and accept price offered. On many apps you can see the taxi's current location and estimated time of arrival.

"Any sufficiently advanced technology is indistinguishable from magic." - Arthur C. Clarke

Now, imagine how you could transform all the necessary communication in your production processes. What could you do if you knew exactly where all the materials were at any moment? What barriers could you remove if tasks could be done automatically in the background? There are a few prerequisites to get started with the transformation to integrate into your existing systems.

Toyota created a "Communication and Information Flow" chart that mapped information between their factory and the suppliers as well as the material movements. This was later adapted into the "Value Stream Map" (VSM) that concentrated on the steps within the factory and shows the links to the suppliers and customers. You may have already identified some of the material and information flows, but the VSM is missing the location data and data formats.

Prerequisites for Digitization (Industry 4.0)

1. Define Locations

The largest impacts in early Industry 4.0 applications have been in logistics. While optimization and scheduling algorithms have been around for years, significant savings could be realized when they were coupled with geo-location. Tracking packages as they pass transfer points has been around for years. You probably have used it when waiting on an important delivery. You entered a unique tracking number and the system could tell you the last place it had been seen... not yet real time, but better than not knowing.

Many of the logistics solutions use GEOLOCATION (GPS - Global Positioning Satellites) to define the location within about three meters (about 10 feet). While this is suitable for moving cargo, GPS is not accurate enough for use in internal location tracking of a production area. If you have done some lean, you probably have done some 5S and painted out locations for materials and machines. Advanced use will have 'defined parking' designated. This might be unique coding for the location.

In the manufacturing operation, defining where all your materials are, where the work is done, and where the finished goods are stored is the basic starting point. Most factories use a grid format for the manufacturing floor (plus floor levels where needed) and additional shelf locations in the warehousing. Knowing WHERE things are is a fundamental necessity to begin digitization. Start with a drawing of the facility, then overlay the work locations.

Loading and unloading locations (docks) are numbered and assigned to specific orders or shipments. To reduce time at the dock, one option is to have defined queuing locations to load or unload. Production picks up or delivers to these queuing locations when you have a functioning just-in-time system.

A leading edge example of using defined locations would be the advanced Amazon warehouses. Their systems have product locations, shipping material locations, packing locations, etc. The transport robots go to a specific point, pick up a bin, take it to the packing location, and manage to avoid running into other robots because of their location tracking software. It even monitors its battery level and knows where an open charger is available when it needs a recharge.

2. Routing - Sequence of work

Now that you have defined where specific work activities are done, materials are stored, etc. It is time to map the sequence of the work for each SKU (Stock Keeping Unit - you should have a unique number for each product variant you produce). With the routing defined, you know the path of the materials from start to finish.

Routing is an extension of knowing WHERE, at a level of where the work is done, where the materials came from, where the finished goods go (or sub-assemblies). Each product can have a unique sequence of where the work is done (routing). Many will have common sequences and we can group these and label them as 'families'. This starts the knowledge base of the materials flows that you connect to locations.

Lean recognizes that similar routings can be combined to simplify layouts.

3. Define where information is used, measured, or created

This can be part numbers, part quantities, performance values, machine performance, etc. The first step is to define where USEFUL information is created or used, then where it is needed. Don't limit this to only what the managers and support staff need, be sure you understand what the people in production need to make good decisions. Use your routing map to define where information is needed, used, and created. Identify where data is not available but is needed. Identify the data format used or created.

Once you can define information source and needs, overlay these onto your routing map. You now know where you need 'inputs/outputs' in your digital system.

Minimum information levels would be part consumption, finished goods created, and critical quality data (with location). Planning and sales can tell you what information would be useful for them to have for decision making. Quality can tell you of the critical processes to monitor and the pass/fail values. Production can tell you the rate of finished goods (or sub-assemblies) needed to make scheduled shipments.

Define the minimum information you need at each step to make factual decisions. This includes the support functions for production. Know what information is needed and where it can be created is a first step. You can then identify the missing links and take steps to close the gap.

The initial goal would be to track in real-time the production processes. You would know materials on-hand, WIP, and finished goods inventories. With the facts you can begin to make to order, assigning shipping labels to generic products as they are packed in the final step. Each production area would need some data functions in the beginning and end of the process sequence.

4. Data Standards

The typical facility will have many different digital systems, from their ERP/MRP/MES, etc. to the PLC's running automation, to generic sensors. The high level barrier is communication between different digital formats. The standard solution is to find or create an API (application programming interface) to share data between different formats. Standardizing the data generation side can reduce the number of API's needed to interact with the primary system. Your choices may be limited as you work to link your legacy systems together.

The basic level barrier is choosing how you will capture the data. Some machines already have sensors that have data you can feed to your primary system. Other points may require you to be creative on how the data is captured and transmitted. For instance, when do you count a completed product? Do you count it at the packaging machine in the individual containers, or at the carton level when you put them in multiples for shipping?

Define who needs what information and how you will supply it to them in real time. On the low tech side, some enterprising companies are using smart phones, watches and tablets for data capture and WI-FI to send it to a central program.

The value of Industry 4.0 is having factual information available in real time to make decisions. You know what you have, where it is, when work can be done, how it was done, who did it, etc.

Do not rush to automate

The computer industry had a saying "Garbage In, Garbage Out"... rushing to automate processes without improving them only cements the poor processes by hiding them from sight.

The primary goal of Taiichi Ohno's Toyota Production System was to create flow. He recognized that you must have robust processes and design that can ensure 100% good work at every step. Kanban Rule #1: Do not send defective product to the next process. Standard Work was the structure to build in quality. Defining your expectations for each task with job instruction lays the foundations for Standard Work.

The goal of the Kanban System was to create a virtual conveyor linking disconnected processes together. Digitization can facilitate the elimination of the physical cards to notify the upstream processes to produce. For more information on Kanban Systems, read the TPS manuals. (Available on request.)

One of the outcomes of observing each task and defining the key points required to achieve the desired outcome of zero defects, is to see many opportunities to improve the process. Only when you truly understand the work content can you effectively improve the process.

Toyota had a set of guidelines defining use of automation... after you had done the basics and mastered the task manually.

• Step 1: Manual operation to power drive - exchanging human labor for mechanical power.
• Step 2: Manual feed to power feed - this is more applicable to lathes and milling machines but is also seen in the stamping presses.
• Step 3: Automatic stop and return to start - this is where you have the machine do a single cycle and return to its starting point to get ready for the next part.
• Step 4: Automatic unloading - ejecting or removing a part once the cycle is complete is the easiest automation to add to stock machines.
• Step 5: Automatic loading - this is the first higher level of complexity... requiring the orienting of the part to be fed to the machine and the proper placement into the jig, fixture, chuck, or other part holding device with enough precision to have a process that is capable of meeting production demands (speed) as well quality.
• Step 6: Automatic problem detection - there are actually two levels to this; detection after a defect is produced and detecting conditions that have a high probability of creating defects. There is an unwritten expectation that the machine will be stopped if defective parts or conditions exist. This is the step designed to eliminate babysitting of the machine. (This step is part of the Jidoka or Autonomation pillar of the Toyota House.)
• Step 7: Automatic transportation - this should be the final step; connecting series of automated machines together where the production sequence is completed through multiple machines without human intervention. There is a tendency to connect series of machines before they have the capability to detect their own problems or produce at a high level of process capability.

Autonomation is a step in Factory Digitization

Autonomation is "Adding a brain and a heart to automation." Developing sensing capabilities in the equipment.

If you already have a machine to do the work but must have an operator to watch (babysitting) to make sure it works properly and does not break (itself or the work).

• First step in autonomation is to add safety function that can detect the same failures that an operator would and stop the machine.
• Second step would be to understand why and how each breakage occurs, then identify nonstandard factors that lead to breakage and develop sensors to stop machine before it happens or remove the contributing material so the machine can continue. This is embedding the brains into the machine.
• Third step is giving it a heart, or sensing capability, this becomes a dynamic 100% inspection process of the work (shortest feedback loop will be measuring as the work is in process).

Having the sensing capability can create data. Your choice is whether the information will be useful for subsequent steps or later decision making, and in what form the data will be useful.

Part of a lean transformation is to have everything in its place and a place for everything to avoid wasting time wandering around looking for things or information. Like the ride hailing apps, digitization can reduce the barriers of knowing where things are and how many you have in a production environment. Approach it as an additional tool to help you make good decisions with facts instead of having to guess or go looking for materials or information.

Doing these basics will uncover many problems (opportunities) that need to be solved to gain the full value from your digitization efforts.

This is not meant to be a detailed road map to implement Industry 4.0, its intention is to point out how you can build the basic structure to begin to leverage digital tools.

Did I miss something needed to get started? Let me know - [email protected]