By Chris Seider
As a proponent of lean, TPM, and six sigma, I’ve always advocated implementing on the most critical to quality (CTQ) inputs and the outputs that showed the least process capability. However, with today’s technology, we need to reconsider how to deploy SPC as one tool to maintain quality.
First, if the relationships between inputs and outputs are well established, then the relative requirements of the inputs are well known. Today’s use of feedback controllers and PLC’s and distributed control systems allow for people to not have to use SPC charts to maintain many X’s. This assumes proper calibrations and proven measurement systems of the X’s are maintained. Think of driving a car in the 1980’s when speed control required adjustment of the driver. Now, one can set a set point on cruise control and have confidence the speed is well maintained. This can be done for many machines including temperature, pressure, and torque controls.
Second, the cost of automation has dropped drastically compared to PID type feedback controllers from the 1970’s. The investment in the controllers along with devices that allow valves to be controlled with actuators or pistons, etc. are much lower than in the past. Today’s refineries don’t leave any temperature, pressure, or level to be uncontrolled if there’s a corresponding way to control the X. Why don’t they allow an uncontrolled furnace or distillation column or even octane output of the gasoline blending operations? Basically, they want to reduce waste and allow optimization to occur to determine optimum setpoints based on constraints and assumptions given to the financial models.
Third, often quality characteristics of the output are measured offline because of the difficulty to sample online or even measure online—expense or technology doesn’t exist. Even well-controlled polymer extrusion of laminators is often sampled offline to ensure measurements aren’t drifting of the online measurements. Someone controlling the process can and should use SPC, statistical process control, of many outputs. However, the ability to create “wallpaper” where no one acts to out of control conditions is worthless. Why have someone mark on a chart the lab results if trends or clearly other out-of-control conditions aren’t acted to address the situation?
Fourth, with today’s advances, great software tools such as MINITAB can now do SPC easily on data generated by online or offline measurement systems. However, this leads to the ability to generate even more “wallpaper” which is a euphemism for SPC charts where someone doesn’t respond to out of control conditions. What’s worse—maybe someone in an office or a home-based work situation may think it’s valuable to review the out-of-control conditions from the previous day/shift’s results. SPC without action as defined by OCAP, Out of Control Action Plans, is just doing the work without the intent of maintaining quality for customers or even maintaining the most profitable production results while maintaining customer requirements.
Maybe one should consider that if a supervisor has difficulty when supervising more than 6 – 10 folks, then maybe one should not have too many SPC charts for anyone operating a process.
