Abnormal Conditions

Imagine you are at home, and you hear a funny sound coming from the washing machine, you smell a slight odor of smoke coming from the kitchen, or you see a water mark on the ceiling. Each of these things indicates that something just isn’t right. They are abnormal conditions.

Sometimes, they tell you that a problem already occurred, but they can also forecast a pending problem. They let you know that you need to fix something that is about to get worse.

Abnormal conditions exist at work, just like at home. On the job, though, abnormal conditions link tightly to the processes that you do. Each process typically has a range of conditions under which it operates. When something gets out of whack, an abnormal condition results. When these abnormal conditions exist, it is hard to get consistent outputs.

An abnormal condition is when something is happening outside of the defined or typical parameters of a process.

In many cases, your ability to control the output of your processes rests upon your ability to keep conditions (inputs) constant. Obviously, the first step is to identify an abnormal condition. The more visual you make processes, the easier they are to fix.

For example, when a temperature gets too high on a machine, an andon (warning) light should flash. Or when a queue of items is too long, it should sit on a red portion of a conveyor. Visual controls should make abnormal conditions not just stand out but jump out at you.

While making abnormal conditions stand out is important, an even better approach is to look for ways to avoid abnormal conditions in the first place. For example, make settings permanent, or at least hard to change. Don’t allow adjustments to things that don’t need them. You’ve probably seen the locking plastic covers over thermostats in common areas to prevent them from being changed. These methods of mistake-proofing (often called “poka yoke” in Lean) keep you from spending your time identifying and correcting abnormal conditions.

Unfortunately, it sometimes takes detective work to even understand what constitutes an abnormal condition. Do you know offhand the temperature at which your car overheats? Fortunately, the engineer who designed it had to figure that out and set the sensor to go off whenever the engine runs hotter than the specified temperature. If he didn’t define that, you’d just be guessing whether you were facing a potentially costly failure in your car’s cooling system.

It’s similar with your processes. You might not know specifically, quantifiably when you cross from normal to abnormal conditions. So, in your Lean office, or on the shop floor, spend some time thinking about abnormal conditions for your processes.

Some are obvious—fluid on the floor, for example. Others are less so. How long is a checkout line before it can be called an abnormal condition? (Some grocery stores actually have a set process in place to respond whenever a line gets over a set number of customers.)

Of course, identifying the abnormal condition is only part of the battle. Once you recognize it, you have to put countermeasures in place to keep it from recurring.