Manufacturing is all about process improvement and quality control. For those who are unfamiliar with the manufacturing industry, here is a little bit about how it works:
When you want to manufacture something the first step is to set up the manufacturing “line” made famous by 20th century industrialist Henry Ford. Every line is different and the particulars are affected as much by the maturity of the company as by the particular product being developed. More mature companies are likely to rely more on automation while those developing the line and/or those short on financing may rely more on less capital intensive investments like manual labor. As you might expect, when a manufacturer is creating a new product there are likely to be some unforeseen areas for future improvement as the product development process starts to churn out finished items. To minimize this kind of error, manufacturers go through a three phase process to set up their operation:
- Initial Qualification
- Operational Qualification
- Performance Qualification
Initial Qualification (IQ):
IQ involves the setup, calibration and initial testing of equipment used on the line. Ask yourself, “is the equipment installed correctly?” Normally, equipment is delivered from the original equipment manufacturer (OEM) with known tolerances and specifications. That said, manufacturers often quality check installation and initial operation to ensure that the equipment does infact function as expected/advertised. Additional IQ steps include setting up safety checks, configuring preventive maintenance and cleaning schedules, and cataloguing documentation.
Operational Qualification (OQ):
OQ involves testing the entire coordinated effort (the line involving both manufacturing and assembly of parts aimed at a final product). Small discrepancies on one part of the line can lead to large errors farther down in the process. Preventing the “snowball” effect in errors normally means shrinking error tolerances for different sub-segments of the process around targets that minimize errors in the final product. OQ also involves determining worst case specifications for various steps in the process to act as limits that help quality assurance personnel produce acceptable results. Considerations typically include developing raw material specifications, process control limits, personnel training and operating procedures.
Performance Qualification (PQ):
PQ involves destructive improvement and maintenance of the process to ensure reliable results (process stability) over time. Quantitative statistical techniques like sampling and Design of Experiments (DOM) and are used throughout all three processes but are particularly relevant in the PQ stage of the process because the major project management and setup has already been accomplished. PQ is all about small iterations aimed at perfection (commonly termed “Six Sigma” in the manufacturing world) with the overall goal of continuously controlling the process. Considerations include verifying the acceptability of the end product, challenging the process to establish potential deviations, and analysis to establish normal variation within the system.
Three primary questions are normally asked when setting up a manufacturing environment and determining whether or not it meets the desired outcome(s):
- Is a process output verifiable?
- Is the cost of verification reasonable?
- Does verification suffice to ensure the desired results?
If a process is NOT verifiable then manufacturers will aim to “validate” it in lieu of verification. The difference between verification and validation is as follows: items that can be verified are typically quantifiable in terms of size or some other measurable quantity. Validation is more of a “yes” or “no” type scenario in that you are establishing whether or not the process can consistently produce a result that conforms to standards. For example, things that are typically verified include things like manual cutting and pH readings. Both of these can be specified as a target number plus or minus a tolerable range. Items that are more suited to validation include sterilization processes, freeze-drying, and establishing environmental conditions for clean rooms. Manufacturers are ultimately looking to setup a process, verify or validate the process, and then control the process to ensure desired results within specifications.
For all three phases, manufacturers should generally be thinking about the following items:
- What can we verify and/or measure?
- How can we verify and/or measure it?
- When should we verify and/or measure it?
- How many measurements are statistically significant?
- What are the acceptance and/or rejection criteria?
- What kind of documentation is required to catalogue the process and the results?
Typically the kind of product characteristics that get measured include visual defects, weight, pressure, temperature, and/or length (thickness, width etc.). Recorded values are generally compared against things like national and international standards as well as specific product requirements. Different measurement techniques like sampling, DOE, or Taguchi methods are commonly used methodologies for testing and results are often measured with control charts involving both upper and lower significance lines. These tools generally indicate the need to make adjustments to the manufacturing process. Process control procedures are typically drawn up to ensure acceptable results and it is important that they are accessible in order to make procedural modifications when out-of tolerance results necessitate adjustments.
Evaluating the Quality of a Process:
One of the principal tools available to quality assurance or operational excellence personnel in a manufacturing environment is the capability study. Capability studies measure the ability of a process to identify potential errors, reduce variation, and prevent deviations from manifesting as unacceptable results. Conducting a capability study typically involves taking samples over a given period of time and plotting the results in a control chart.
Control charts are a visual representation of process outputs as they relate to target values, upper and lower specification limits, worst case values, and “action values.” Samples above or below action levels or frequently repeated samples above or below the target (e.g. 7 in a row above the target level) typically dictate stopping the line to determine what is going wrong.
In summary, manufacturing environments work on very regimented schedules with extremely high attention to detail. Careless design and/or a fundamental misunderstanding of general manufacturing principles and the way a line works can lead to problems when you bring your product to mass production. Taking careful note to simplify both manufacturing processes and assembly will save you money in the long run because it means less working hours for the manufacturer tasked with producing your product.
To learn more about the manufacturing process click here