Intelligent Sensors for Sustainable Food and Drink Manufacturing

Abstract

Food and drink is the largest manufacturing sector worldwide and has significant environmental impact in terms of resource use, emissions, and waste. However, food and drink manufacturers are restricted in addressing these issues due to the tight profit margins they operate within. The advances of two industrial digital technologies, sensors and machine learning, present manufacturers with affordable methods to collect and analyse manufacturing data and enable enhanced, evidence-based decision making. These technologies will enable manufacturers to reduce their environmental impact by making processes more flexible and efficient in terms of how they manage their resources. In this article, a methodology is proposed that combines online sensors and machine learning to provide a unified framework for the development of intelligent sensors that work to improve food and drink manufacturers' resource efficiency problems. The methodology is then applied to four food and drink manufacturing case studies to demonstrate its capabilities for a diverse range of applications within the sector. The case studies included the monitoring of mixing, cleaning and fermentation processes in addition to predicting key quality parameter of crops. For all case studies, the methodology was successfully applied and predictive models with accuracies ranging from 95 to 100% were achieved. The case studies also highlight challenges and considerations which still remain when applying the methodology, including efficient data acquisition and labelling, feature engineering, and model selection. This paper concludes by discussing the future work necessary around the topics of new online sensors, infrastructure, data acquisition and trust to enable the widespread adoption of intelligent sensors within the food and drink sector.

A Review of Multisensor Data Fusion Solutions in Smart Manufacturing: Systems and Trends

“…In the context of Industry 4.0, the use of sensors has been widely extended, as the capacity to store and use the acquired data has been enhanced [69]. The heterogeneity of the machinery in the manufacturing plants and the specific needs of each sector generate the need to use different kinds of measurements to monitor the machines, the manufacturing processes, and the parts that are produced.…”

“…Thus, the extraction of relevant features for the specific machine or process and the fact of generating features with reduced missing data are important for the accuracy of the models that can be trained afterwards [108]. It should be highlighted that some advanced artificial intelligence models do not need this feature extraction step to be performed previously, as they operate directly with the acquired time series, but these techniques usually require very large data volumes [69]. The feature extraction techniques can be categorised depending on the domain in which the features are extracted [109].…”

“…In that sense, techniques can be categorised as supervised or unsupervised, depending on the usage of the labels of the target variables [69]. In supervised techniques, the performance of adding or removing a variable is assessed for predicting the target variable, whereas unsupervised techniques use statistical tests to determine if the new features are similar (redundant) or bring new important information.…”

Where to Place Monitoring Sensors for Improving Complex Manufacturing Systems? Discussing a Real Case in the Food Industry

“…In this context, an important issue refers to the compatibility of sensors with different types of food and drink. Watson et al [ 56 ] noted that the type of sensor used can depend on a variety of factors, such as the physical and chemical properties of the food, the desired level of accuracy, and the environmental conditions of the manufacturing process. For instance, some sensors may not be suitable for use with acidic or corrosive foods, while others may be affected by high temperatures or humidity.…”

Combination of machine learning and intelligent sensors in real-time quality control of alcoholic beverages

“…Quality control of alcoholic beverage heavily depends on the sensory evaluation of trained expert; however, human panels are usually expensive, subjective, and it is difficult to establish a corresponding mathematical model (Aleixandre et al, 2018). ML-enabled intelligent sensory systems provides a real-time quality control technique that is nondestructive, high-speed, good repeatability, reliable results, no sensory fatigue, and no complex sample preprocessing process (Watson et al, 2021). This review firstly summarizes the novel intelligent sensors that is suitable for sensory evaluation and the advanced ML algorithms used to process complex sensory data.…”