LIUC Corso di Gestione della Produzione e della Logistica 2009-2010.
Docente : Prof. Claudio Sella
Un caso di controllo statistico del processo.
In manufacturing the results of each process is vital to the company. At Honda of America they rely on statistical process control to continually control and monitor their processes.
Sam Kennedy- Associate Chief Inspecting Engineer
“Statistical process control is simply monitoring the consistency and repeatability of a specific process. Its not a complicated thing, really its quite a simple approach to understanding: What is the reliability of my process? An example of that might be torquing a bolt. In torquing a bolt there are a lot of variable that can go wrong to cause an over-torque or under-torque situation. In that example, if we apply statistics and understand the actual capability of the process, we can predict the failure rate within that process ad making running adjustments so that we don’t have actual problems occur.”
Another example of Honda’s use of data collection to understand its processes is the Door-Closing Speed Test, this test measures how fast a door needs to be moving in order to close, but without slamming it.
“Connie is preparing to measure the door-closing speed on a 4 door civic. The purpose of this test is to measure the speed that it takes to close the door based on a customers’ perception. The first thing she’ll do is she’ll scan the VIN number (Vehicle Identification Number) in to the DataMite, the DataMite is simply a handheld computer. She prepares the car by making sure it is completely sealed, and then she attaches the sensor to the rear door then an interrupter to the left side of the driver’s door. She will repeatedly take readings of the door closing versus the door not closing until she gets within 0.1km/hr. The gage reads in km/hr. So when she reaches a point where she’s within 0.1km/hr of the door closing versus the door not closing, that’s the reading she reports. It may take her about 10-15 tries to get an accurate reading. She found her reading and now she’ll punch the numbers into the DataMite. The door closing speed data is input into the computer where a statistical software package is used to analyse the data. We use SPC control charts, to analyse the data. At the top is an X-chart (individuals chart). As you can see the data is consistent and stable. There are no assignable causes present, only common cause variation. In addition to the SPC chart we use a histogram in order to do a process capability on this particular characteristic. The histogram shows a distribution of data based on about the last 2 weeks of data. The CPK has a capability index that tells us how well we are meeting our specifications. Industry standards push for a CPK of 1.3. Here we are currently meeting that.”
“We have a number of different ways that we collect data when we do a statistical control and monitoring. Thos range from equipment that automatically collects the data, to a manual process where the associate is actually taking a reading and plotting it manually on a chart to a process where we may be doing some offline inspection and actually collecting the data with some hand gages. This data is processed in a number of different ways but the most effective is by associate understanding: What do they do when they are out of control? When they find a point or measurement that is not where it is supposed to be and then hey take action on it, on an ongoing basis. Waiting and analysing the result after is not always that effective. For continuous improvement, doing it as you go along is what’s important. Waiting and analysing the result after is not always that effective but for continuous improvement, doing it as you go is what’s important.”
During the door closing speed test result came back that extended outside the upper control limit on the control chart. This indicated that for one door on one certain car it took an unusually high amount of speed to close the door without slamming it.
“This is an individuals/X chart, showing a plot of data of door closing speed over a certain period of time. On the chart you’ll notice one out of control condition at this point. Meaning there is an assignable cause potentially present. This assignable cause was investigated and it was found to be a problem with the fixture used to install the striker. The striker was installed slightly offset due to a broken fixture. When the striker is slightly offset it will cause the door closing speed to be higher than normal. The fixture was replaced and the data fell back into control and it was consistent and predictable once again.”
By teaching their employees to understand statistical process controls and how to immediately counteract out of control situations. Honda prevents problems fro recurring during manufacturing and is able to maintain an outstanding level of quality control.