Ultrasonic Detection Modeling of the Escherichia coli microbial contamination of UHT Milk packages using Artificial Neural Network

Detecting microbial contamination of milk using novel engineering techniques is very worthy. In current study, microbial contamination of UHT milk packages was detected using ultrasonic sensors. Milk packages artificially were inoculated to E. coli in four dilutions and three replications. Monitoring of ultrasonic properties was performed by measuring amplitude and time delay factors. Artificial neural network designed for predicting total count and pH of milk packages based on ultrasonic properties. Results showed that contamination of milk packages for initial dilution 1000 CFU/ml after 7.5 h is capable to detect, and detection period would be increased in conjunction with initial bacterial dilution decreasing. Trained neural network predicted total count and pH values with the coefficient of determination 0.979 and 0.795 against the experimental values. According to the current project, is resulted that microbial contamination is detectable using ultrasonic technique, and to achieve high accuracies, more researches are needed.