Posted March 16, 2020 in Blog
We can talk all day about how near infrared (NIR) measurements improve process control and quality control in various industries, such as snack food, but the proof is in the pudding (pun intended).
One American snack food manufacturer has implemented NIR benchtop sensors in an at-line configuration. This company makes a variety of products such as potato chips, corn products, pretzels and popcorn. There was a need to measure as many as 12 products, essentially being produced at the same time. In one facility there was a need to measure moisture and oil in many different products from various lines. Each product needed to be measured at least once every half hour. There would be no practical way to accomplish this by using classical analytical techniques. A conventional moisture balance would require at least five minutes per sample and some level of grinding and sample preparation. The oil analysis would require significantly more time, both in terms of assay time and sample preparation time. In addition, this company did not want to be saddled with use of dangerous and expensive solvents and reagents.
The solution was to use a series of modern benchtop NIR analyzers that can measure moisture and oil on product taken directly off the process line. The whole sample is placed in a stainless-steel bowl, which is then placed onto the NIR system. The sample is automatically rotated at the press of a button, and the system displays the percentage of moisture and oil within 10 seconds. Because the sensor “sees” all the various product surfaces during the sample rotation, and because the sensor is making 30 measurements per second, the answer is both accurate and extremely repeatable. This company is now able to react to out-of-spec conditions more rapidly, greatly reducing the amount of waste and, most importantly, improving the product quality.
Another U.S. snack food producer has chosen to rely more on continuous online measurements. This company has used an old NIR analyzer in their Quality Assurance Lab and was therefore familiar with the basic technology. They have installed new NIR sensors directly over their potato chip, corn chip and tortilla lines. Again, the new style NIR sensors are making 30 measurements per second as the product is conveyed under the sensors. The sensors measure both moisture and oil continuously in real time about 20 feet from the outlet of the fryer. This, of course, completely eliminates the need for any routine sample handling and testing. The moisture and oil content are displayed continuously at the operator stations. Some care had to be taken in the placement of the sensors so that they were able to see a consistent flow of chips.
As this company made the gradual change from an off-line testing routine to total online measurements, there were a series of “people” issues that needed to be addressed. In short, the operators had to buy into the new technology. What would happen if the new NIR sensor’s values did not match up with the Quality Assurance Lab’s values? What answer do you believe? When should the operator consider making an adjustment to residence time, for example? The answers to these questions quickly worked themselves out. The first step was to establish calibrations, comparing the online NIR gauge to the old QA system. This was easily and quickly done as the values for the two devices, for both moisture and oil, matched up nicely (correlation coefficients of .98 and low standard errors).
The next step was to compare a sample taken directly from the process line. The sample was grabbed immediately after having gone under the online NIR sensor and the values recorded. The samples were then taken to the QA Lab and run on the old NIR system. The results were mixed—some samples matched up well; some did not. It turned out that the chips were not very homogenous, especially in oil content. This was a sampling issue.
At this point, one of the operators made a suggestion: Why not take the analog output of the online NIR sensor and run it into a process chart recorder? Then grab samples only when the recorder indicated that the moisture or oil value was trending toward an out-of-specification level; then grab samples and compare the online values to the lab values. This was magic. Once again, the correlation coefficient was at .98 and the standard errors were low. The operators learned to trust the online NIR sensor and to make changes based on their ability to see the trace on the recorder moving toward an out-of-spec condition.
Another interesting application at this company is the automatic control of flavoring addition for both cheese curls and popcorn. A new NIR sensor is installed on a conveyor after a tumbler. The tumbler applies the flavoring in the form of slurry. The NIR sensor actually measures the oil content of the slurry. The amount of oil in the slurry has a direct relationship to the total amount of flavoring being applied. The NIR sensor’s 4-20mA output is run into a process controller that automatically adjusts the amount of flavoring applied by manipulating a pump to a desired set point.
NIR analyzers have been assets to the snack food industry for many years. They have been widely used and are an accepted method of analysis in many QA environments. Our generation of benchtop and online NIR sensors have benefited from greatly improved detector and optical technologies. NIR sensors are now easy to calibrate, simple to use and reliable even in difficult processing locations. They can not only make repeatable measurements but can also be attached to plant control and data acquisition systems that make these measurement devices an even more valuable tool to help in the delivery of a high-quality, consistent snack food product.
Learn more about NIR analysis in the food and snack industry and request a free quote or a demo by clicking on your closest application here.