Seed Cotton Mass Flow Measurement in the Gin

Authors: Hardin Iv, Robert

Submitted to: Applied Engineering in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/28/2018
Publication Date: 6/19/2018
Citation: Hardin IV, R.G. 2018. Seed Cotton Mass Flow Measurement in the Gin. Applied Engineering in Agriculture. 34(3): 535-541. https://doi:10.13031/aea.12647.
DOI: https://doi.org/10.13031/aea.12647

Interpretive Summary: An accurate measurement of seed cotton mass flow rate is needed for improved process control systems in cotton gins. Previous research developed a mass flow rate measurement system for seed cotton based on the energy required to accelerate the material when introduced into the conveying air stream, but improvements were needed for use in commercial gins. The energy of the conveying air stream needed to accelerate material can be determined by measuring the static pressure difference in the conveying system over the distance the material is accelerated from rest to its maximum velocity. This pressure difference should be proportional to the mass flow rate of material and the air velocity. The feeding system for seed cotton in gins is well-suited to making the necessary measurements. Previous testing identified that air leakage at the blowbox rotary valve, where seed cotton is fed into the conveying air, and rapid heat exchange between the air and seed cotton were significant sources of error in the original sensor design. An estimate of air leakage and a measurement of the temperature in the blowbox were incorporated into the model used by the measurement system. Using the improved system, an experiment was conducted in the commercial size gin at the Cotton Ginning Research Unit to test the effect of cultivar, air velocity, seed cotton mass flow rate, and dryer temperature on accuracy of the mass flow measurement system. Mean absolute error for the improved system in predicting the conveyed seed cotton mass was 3.9% for the first stage conveying system and 2.9% for the second stage system; however, dryer temperature had a significant effect on the model parameters. The temperature measurement location in the blowbox was likely not optimal; therefore, another regression parameter was added to the model to better estimate the average air density in the blowbox. This improvement reduced the mean absolute percentage error to 2.5% for both systems and eliminated the effect of dryer temperature on the model parameters. This level of accuracy would be useful for gin process control.

Technical Abstract: Seed cotton mass flow measurement is necessary for the development of improved gin process control systems that can increase gin efficiency and improve fiber quality. Previous studies led to the development of a seed cotton mass flow rate sensor based on the static pressure drop across the blowbox, primarily due to acceleration of the seed cotton. The initial sensor did not perform satisfactorily in a gin, and modifications were made to account for air leakage through the rotary valve at the blowbox and the temperature drop occurring due to heat exchange between the seed cotton and air. Mass flow rate was predicted based on the static pressure differences across the blowbox and rotary valve, the air velocity and density at the blowbox inlet, the air density in the blowbox, and the ambient air density. The first- and second-stage seed cotton cleaning and drying systems of the commercial-scale gin at the Cotton Ginning Research Unit were instrumented to test the improved model. Air velocity, cultivar, dryer temperature, and seed cotton feed rate were varied to determine their effects on model accuracy. The mean absolute percentage errors in predicting mass flow rate were 3.89% and 2.85% for the first- and second-stage systems, respectively; however, dryer temperature had a significant effect on the regression coefficients. An additional regression parameter was added to the model to better estimate the average blowbox density, reducing the mean absolute percentage error to 2.5% for both systems and eliminating the effect of dryer temperature on the regression coefficients.