Prediction of Wort Density with LSTM Network
This work addresses a domain-specific problem for brewers by providing an automated solution to improve measurement accuracy in beer production, but it is incremental as it applies an existing LSTM method to a new application.
The paper tackled the problem of error-prone and expensive manual measurement of wort density in beer production by developing a system that calculates density from inexpensive sensor data using an LSTM neural network, resulting in a method to reduce errors in data collection.
Many physical target values in technical processes are error-prone, cumbersome, or expensive to measure automatically. One example of a physical target value is the wort density, which is an important value needed for beer production. This article introduces a system that helps the brewer measure wort density through sensors in order to reduce errors in manual data collection. Instead of a direct measurement of wort density, a method is developed that calculates the density from measured values acquired by inexpensive standard sensors such as pressure or temperature. The model behind the calculation is a neural network, known as LSTM.