How to Automate Critical Product Cleaning – Whether or not you ever automate

If you plan to automate the cleaning process, explore some factors that are key to successful automation. Your reward will be economical automation that does what you need it to do. If you never, ever expect to automate the cleaning process, you, too, will achieve more rugged, reliable manufacturing by understanding the non-machine factors involved in automation. The secret to successful automated cleaning (and to successful non-automated cleaning) is truly understanding how the cleaning process works and completely understanding what that cleaning process must accomplish. 

 Many, perhaps most, manufacturers will change their cleaning processes within the next 3 to 5 years. These new processes will involve new cleaning agents and/or cleaning agents that are used in cleaning equipment, equipment that takes cleaning out of the hands of technicians and assemblers. Moving from manual cleaning with an aqueous cleaning agent to a simple aqueous cleaning system (one wash tank, two rinses, a drier, for example) involves automation. Moving from an open-top degreaser with a chlorinated solvent to an airless/airtight system involves automation. Automation starts in the cleaning equipment; and it involves sample handling and cleaning recipes.

Ideal automation
Once upon a time, to learn about a topic, people looked to a printed encyclopedia, like the Encyclopaedia Britannica. The on-line version, “Britannica,” contains a general discourse about automation. An intuitive concept of “automation” evokes a manufacturing facility devoid of people, one where raw materials become components and components are assembled into a final product. Everything is handled reliably by hoists, conveyor belts, perhaps robotic assistants. With artificial intelligence (AI), no human intervention is needed; and engineers spend their days pondering the next generation of product.

“In general usage, automation can be defined as a technology concerned with performing a process by means of programmed commands combined with automatic feedback control to ensure proper execution of the instructions. The resulting system is capable of operating without human intervention.” (3)

While automation is discussed in many applications, the presumption is that in manufacturing, automation will be used in fabrication, assembly, surface coating, and inspection. However, the author recognizes we are far from the ideal. Even with systems that have extensive data and that include feedback and software controls, the reality is that automated systems must interact with humans in some way; for example, “to manage the system, load and unload parts, change tools, and repair the equipment.” 

Sustainable automation
Automating critical product cleaning is complex and specialized. It requires knowledge, understanding, and flexibility. Substantiable automation requires an intelligent, alert, and responsive workforce. 

Back in 2007, Ed and I discussed steps to automation (1). We took the “Stages of Knowledge” outlined by Bohn (2) and then interpreted them to the challenges of critical cleaning. 

The eight stages are:

1 – Complete ignorance
2 – Awareness of parameters, ignorance of relative importance
3 – Understand parameters, can’t control them
4 – Control of the mean; uncontrolled parameters
5 – Process capability
6 – Process characterization
7 – Understanding “why” the process works
8 – Complete process understanding

While these steps evoke a quest for the meaning of life, perhaps ending with a session with a seer on the top of a mountain far from civilization, they reflect reality. Automating involves observation, introspection, and (gasp!) reasoning. Who has to do all this? Maybe you, maybe you and your team – sometimes we can help. For example, process capability involves having defined cleaning recipes. Asserting that the cleaning recipe is one that the FDA accepted as part of a submission or that the cleaning technique is a requirement from a key client is a bit of a cop-out. Process characterization involves understanding how the process works and why it works. We pointed out that “for aerospace, pharmaceutical, medical, and microelectronics applications, performance requirements may shift and the desired surface quality may be even more elusive.(3)” 

Cleaning machines and hoists are not enough
As soon as you change a cleaning process from having workers hand-clean the product in a “bucket” or small tank or with a spray bottle to one using a cleaning machine, you are automating the process. In response to a reprint of our recent article asking if manual or automated cleaning is better (4), suppliers of automated equipment and automation specialists made comments offering their services. Quality, well-designed automated equipment is a must, and quality, well-designed cleaning equipment is a must. Purchasing equipment does not guarantee a successful automated cleaning process. 

The success of this automation depends on your knowledge and your understanding. Bohn (1) poses a number of questions to gauge how much you know; and, perhaps more important, how much you think you know but don’t. He describes false knowledge as “one of the most painful forms of ignorance.” For example, suppose you are certain you are at stage six or higher knowledge about a process variable, for example, the length of time the product needs in the wash tank. Suppose that knowledge is based on historical experience, but it is not correct for the present process because a supplier changed a machining fluid to one that is more adherent and takes longer to clean. Automation based on the historical suppositions is likely to be semi-successful at best. 

Interactive automation
To be sustainable, automation cannot be based on a static model but rather on “a dynamic, interactive, knowledge-based model.” The only constant is change. Fabrication, assembly, and cleaning processes change. Metalworking fluids, coolants, substrates, and product designs change. Customer expectations change. Regulatory requirements continue to evolve. To understand your own cleaning process, you have to understand the impact of changes in the supply chain. Some companies constantly change suppliers, using the lowest bidder. Where critical cleaning is ultimately required, this can lead to so-called “unexpected” product failure. You ought to expect failure with such an approach! Even if your company works with a defined, constant set of suppliers, if your supplier or a captive shop changes a process fluid or cleaning process, your rationally designed cleaning process used with sophisticated carefully-designed cleaning equipment, can fail. 

An observant cleaning technician performing a manual cleaning process might notice a change in the soil level and adapt the cleaning technique, perhaps without documenting the change. This adaptation may work in the short run, but it won’t necessarily be repeatable over time; and the process may fail as soon as a different technician is in charge. Further, without documentation, the cleaning process is not defendable. 

Data versus information versus knowledge versus standards
Data alone is not the same as information; the data has to be organized to provide useful information, for example, on a control chart (1). Building on that, records or tests showing conformance to one or more standards is not the same as process knowledge or understanding. Ask questions! Is the requirement to meet the standard pertinent? Is the standard relevant? Was it ever relevant? Is the standard wrong? What additional factors ought to be considered in determining appropriate surface cleanliness or surface quality? 

Sometimes, manufacturers are mystified as to why the cleaning process conforms to a standard or specification but that they have issues with reject rate or product performance. Something is missing! What’s missing is process knowledge and understanding. 

The “I” in Artificial Intelligence is key. The more complexity, the more unknowns, the less we understand about the variables in a process, the more we need human intelligence. Human intelligence includes not only knowledge of the cleaning process but also process and product understanding. We have to know what happens upstream (what our suppliers are actually doing versus what we require) and what happens downstream (what our customers, internal and external, require). This knowledge encompasses many facts, many factors. It takes reasoning and interpretation of the facts to move from knowledge to understanding. 

Understanding is a key to successful automation, including automating the cleaning process. We know you can achieve process understanding; it’s a matter of cleaning “street smarts.” And, of course, we are always happy to help you speed up the road to understanding. 


  1. B. Kanegsberg and E. Kanegsberg, “The Winding Road to Automation,” Controlled Environments Magazine, April, 2007. (note: please contact us to obtain a copy of the article)
  2. Bohn, R.E. “Measuring and managing technological knowledge,” Sloan Management Review, 36(1) 61-73 (1994)
  3. Groover, Mikell P.. “Automation”. Encyclopedia Britannica, 22 Oct. 2020, . Accessed 17 January 2022.
  4. B. Kanegsberg and E. Kanegsberg, “Manual Cleaning or Automated Cleaning: Which Is Better?,” reprinted from Clean Source September, 2021, by Tim Pennington in, 
Back To Newsletter Archive


Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.