For The Learning Organization
From here on in, we promise to abandon talk about collaborative robotics, Industrial Internet of Things, Industry 4.0, and even Industry 5.0. The 5C approach focuses on seizing everyday opportunities, solving practical problems, and uncovering the unknowns. The 5C approach can be viewed as a digital transformation maturity model and learning framework. It is extremely important that we start exactly where you are now and unlock changes at a fundamental level to realize value based on facts and simplification of processes. We accept your current organizational, cultural, and technical capabilities, because that is the reality.
Before we begin, you are encouraged to review below research on automation potential in the United States. We will focus on the manufacturing sector’s data collection, data processing, and predictable physical work components. As we walk through the five principles together, keep some of your own opportunities in mind.
The Five C’s
Your system is not broken, but it is complex and even brittle at times. Over time, changes introduced across areas of your business affect the system in one way or another. Those changes are difficult to measure or visualize at a macro level, but certainly have an intended impact where the change has been made. Often, changes made in one area or work stream lead to unintended consequences in other areas. Think of an engineer who makes changes to tooling at an upstream operation to satisfy your customer’s critical part dimensions. Did the engineer consider the impact on process capabilities in downstream operations? Another great example is how production is controlled on the shop floor. Does management hold daily meetings with supervisors and execute production from a redlined spreadsheet? Do those changes ever make it back to the master planning system to allow for a more effective schedule the next time around? Change is inevitable and cannot be imposed. The driving forces of change must be unlocked in order to promote a healthier system. The goals of the five C’s are to uncover the measurable patterns in your order to cash cycles by utilizing already existing data streams and transactions through connectivity and integration.
Connecting equipment to a network has been done for decades. However, with the advancements in technology this task has become a lot more cost effective. There are several aspects to consider when assessing connectivity readiness. One fact remains constant, involvement of a network specialist is necessary. There is nothing worse than bringing down a production enterprise network or expanding its attack surface unknowingly.
Equipment is fitted with sensors and a controller. The benefit is that connecting to a network is as simple as adding networking hardware and running a cable to the controller. The drawback is evident in the data collection step where system ownership and modifications are under scrutiny.
Equipment is electro-mechanically operated. Older equipment that is relay driven is not necessarily a bad place to start. Many inexpensive retrofit options exist that allow us to depart from traditional control schemes and embrace modern technology that simplify the overall design.
Manual production process. Lots of design decisions have to be made before instrumentation can take place. The catch is that temporary hardware can actually assist with your decision making, leading to a permanent installation.
Data collection directly from the source, or in-band, is a lot different than what many of us are used to. Out-of-band and after the fact data collection to satisfy enterprise systems is not what we are talking about here. The moment a manufacturing order or shop traveller is separated from the product, errors and omissions can and will occur. In-band data collection inherently reveals patterns of operation and human behavior. Simply put, the goal is to completely remove human intervention from the data collection process. The reason is not because we distrust the operator, but because we require access to accurate and unfiltered machine signals to support the maturity of our analyses. Let us assume we are trying to build an hour by hour dashboard of productive time and quantities produced. At the same time we want to build the top ten reasons of process failure. Consider the timeline below of a process switching between productive and non-productive states.
On custom built equipment what you will likely find is that the data points required for monitoring purposes are non-existent. Your first instinct might be to interpret multiple signals and extrapolate the data you require. We would recommend this be your last option. Modifying the controller logic or adding instrumentation to support monitoring data points should be your first choice. Over time you will find the investment worthwhile as you create a standard data set which works for you and apply it to other like processes.
Going back to uncovering data about the six big productivity losses. Traditionally we would ask the operator or team lead to track events as they occur. This is impractical and does not allow us to easily construct a process baseline to improve upon. Once properly instrumented, the only piece of data entered by a human should be the reason code for failure. In the beginning, we recommend limiting reason codes to ‘man’, ‘machine’, and ‘other’. This will allow you to focus on automating data collection efforts without additional noise that could lead to premature corollaries. Expanding reason code selection will be done at a later time as part of an escalation model.
Unprocessed raw data and facts have very little value because they lack meaning and interpretation. At this point our hour by hour dashboard is merely a list of productive times and quantities. It does not answer questions like what job ran? How long did it run for? Who ran it? Why was this job running? Combining machine with enterprise data begins to raise eyebrows through the context it generates. Take the time to validate the accuracy of your findings. Sometimes you will find that the enterprise data your business has been relying on is very far from the truth. This uncovered fact is your first opportunity to ask ‘Why?’ Remember that the 5C’s are meant to be iterative and will reveal many patterns and opportunities along the way. To put this into perspective, we often hear upper management’s request to visualize OEE. This is impossible to do in the beginning phases, but what we do offer them are gemba truths that start much more valuable conversations.
We present the data we uncover in many contexts. This allows us to convey meanings across functional boundaries and requires computation. A form of computation that always gets attention because of its familiarity is time and dollar aggregation. However, we are not generating the same values that already exist in black and white for comparison’s sake. We present values that were not quantified before. We have just unlocked new abilities for the business — we unleashed their data.