Machine learning algorithms to estimate drying characteristics of apples slices dried with different methods

Implementation of fuzzy logic control algorithm for temperature control in robusta rotary dryer coffee bean dryer

BACKGROUND

Indonesia is one of the coffee producers ranked third in the world in the supply of coffee beans. To maintain competitiveness international market, it is necessary to maintain and improve the quality of coffee beans.

OBJECTIVE

One crucial aspect of maintaining the quality of coffee beans is maintaining the moisture content of green coffee beans. One of the water content settings is using the drying method. While traditional drying methods often experience weather and long-time constraints.

RESULTS

This study designed an innovative coffee bean dryer based on fuzzy logic to overcome the problem. This system uses temperature control with Mamdani's fuzzy logic control interference algorithm, input and delta errors, and output percentage valve opening. This method achieved a moisture content following SNI standards of 12% and a 0.00015% / s drying rate for each coffee bean mass increased by 1kg. This method is also more efficient and stable in maintaining the temperature at a value of 50°C.

METHODS

The drying equipment also estimates the drying time by considering variations in the mass of coffee beans. This dryer can provide an effective solution to maintain optimal coffee bean quality.

CONCLUSION

The second semi-wash method of drying coffee beans using a fuzzy logic-based coffee bean drier has proven successful for drying coffee beans to a moisture content of 12% in a period of 90 min to 195.65 min with a drying capacity of 1 kilogram to 10kg at 50°C.•The coffee beans utilized in the studies are robusta coffee beans from plantations on Mount Kawi's slopes in East Java, Indonesia.•The trial sample was 1 kilogram of green coffee beans removed from the horn skin.•According to SNI standards, the drying performed is the second in the postharvest semi-wash procedure to achieve a moisture content of 12%.

Effects of ultrasound pretreatment assisted drying methods on drying characteristics, physical and bioactive properties of windfall apples

BACKGROUND

Drying is one of the most important post-harvest processes in order to ensure that the products harvested in certain periods are consumed throughout the year. In this study, Golden Delicious, Oregon Spur and Granny Smith apple cultivars (fallen apples) were dried in open-sun, a controlled glasshouse, microwave oven (200 W), hybrid system (100 W + 60 °C), convective greenhouse (70 °C) and freeze-dryer (-55 °C) with and without ultrasound pretreatments.

RESULTS

Present findings revealed that the longest drying duration was observed in freeze-drying and the shortest in microwave drying. The greatest color change (ΔE) was observed in ultrasound pretreated Oregon Spur apples dried in a controlled glasshouse. The greatest total phenolics (40.08 mg GAE g^-1 ), DPPH (2,2-diphenyl-1-picrylhydrazyl) inhibition activity (46.96%), FRAP (ferric reducing ability of plasma) antioxidant activity (68.01 μg Trolox g^-1 ) and ascorbic acid content (1.52 mg AAE g^-1 ) were observed in open sun-dried Granny Smith apples without ultrasound pretreatment.

CONCLUSIONS

Generally, microwave and hybrid drying yielded better outcomes for bioactive, HMF (5-hydroxymethyl-2-furfural) and drying characteristics of windfall apples. © 2022 Society of Chemical Industry.

Soil Moisture a Posteriori Measurements Enhancement Using Ensemble Learning

This work aimed to assess the recalibration and accurate characterization of commonly used smart soil-moisture sensors using computational methods. The paper describes an ensemble learning algorithm that boosts the performance of potato root moisture estimation and increases the simple moisture sensors' performance. It was prepared using several month-long everyday actual outdoor data and validated on the separated part of that dataset. To obtain conclusive results, two different potato varieties were grown on 24 separate plots on two distinct soil profiles and, besides natural precipitation, several different watering strategies were applied, and the experiment was monitored during the whole season. The acquisitions on every plot were performed using simple moisture sensors and were supplemented with reference manual gravimetric measurements and meteorological data. Next, a group of machine learning algorithms was tested to extract the information from this measurements dataset. The study showed the possibility of decreasing the median moisture estimation error from 2.035% for the baseline model to 0.808%, which was achieved using the Extra Trees algorithm.