P. Cano Marchal

On-line system based on hyperspectral information to estimate acidity, moisture and peroxides in olive oil samples

We propose an on-line system to be applied in olive oil industry.We employ hyperspectral images for the prediction of olive oil parameters.Optimal wavelengths have been selected with different methods.We show the regression lines and the validation results. The analysis of the quality of virgin olive oil involves the determination of a series of properties, such as chemical indexes and organoleptic characteristics. In addition, the determination of these properties in real-time could be useful in order to improve the olive oil extraction process since the process parameters could be regulated with the real-time moisture information.In this paper, the feasibility of using a non-invasive hyperspectral device, in order to determine on-line three parameters of the olive oil (free acidity, peroxide index and moisture) is studied. In order to study the correlation between these parameters and the information obtained by the hyperspectral sensor (absorption level), three different methods were applied: genetic algorithms (GA), least absolute shrinkage and selection operator (LASSO), and successive projection algorithm (SPA). From the experimental results, reduced values in cross validation were obtained and the optimal wavelengths were pointed out.

Optimal Production Planning for the Virgin Olive Oil Elaboration Process

Abstract The quality and obtained quantity of Virgin Olive Oil is bounded by the characteristics of the olives to be processed, and further determined by the influence of the process variables during the actual elaboration. Since the quality of the olives evolves during the harvesting season, it is relevant to consider when to harvest the olives in order to maximize the profit over the whole season. This work proposes a method to determine an optimal production plan for the whole harvesting season and presents the results obtained in its application to four different scenarios.

Decision Support System Based on Fuzzy Cognitive Maps and Run-to-Run Control for Global Set-Point Determination

The selection of the set points of the process variables plays a fundamental role in the final output of modern processes, as the quality characteristics of the final product and the performance metrics of the process are heavily influenced by these decisions. This paper presents a decision support system to determine these set points based on three components: a model of the system of interest based on the Fuzzy Cognitive Maps methodology, an optimization problem that provides the most appropriate actions to perform in each particular situation, and an observer that augments the optimization problem and allows to include feedback to the system, practically implementing a run-to-run control approach. A simple application example is included to illustrate the proposed approach.

Fuzzy Decision Support System for the Determination of the Set Points of Relevant Variables in the Virgin Olive Oil Elaboration Process

The virgin olive oil elaboration process (VOOEP) is an industrial process whose control relies heavily in the knowledge of experts. In this paper, the basis of a fuzzy expert system for the determination of the set points of the VOOEP which uses inputs from the expert operator is presented, and a system is constructed and validated based on the knowledge of expert operators of the process. The expert system is built based on the decomposition of the original problem into three sub-problems, deriving a sub-system for each of them. Besides, a notion of maximum attainable quality estimate and a paste preparation variable are introduced.

Control System of the Malaxing State for the Olive Paste Based on Computer Vision and Fuzzy Logic

The olive paste preparation is one of the most important task within the olive oil elaboration process. It greatly determines the quality of the obtained oil, and imposes an upper bound on the achievable extraction yield. Currently, the malaxing state control of the olive paste is performed by the master miller, who observes periodically the paste inside the thermo mixer and acts on variables manually. In this paper, the implementation of a new methodology to control the malaxing state of the olive paste is presented. The methodology is based on a software sensor that obtains some olive paste features from the acquired images inside the thermo mixer, such as granularity, presence of floating oil and viscosity. This software sensor is formed by two camera devices, and it is coded by image processing algorithms from textural and color information. Concretely, the grey level co-ocurrence matrix and the segmentation method based on threshold, were employed. Also, a fuzzy controller was designed to act automatically on the thermo mixer variables. This controller is based on the master miller expert knowledge, and it allows to adjust the malaxing time needed to prepare the olive paste to the minimum value. This fact supposes a significative improvement in the energy consumption of the plant.

Modelling uncertainty in production processes using non-singleton fuzzification and fuzzy cognitive maps - a virgin olive oil case study

Decision support systems (DSSs) are a convenient tool to aid plant operators in the selection of process set points. Inputs to these systems for variables that are not easily measured online often come from assessments made by experts, with an associated degree of uncertainty. The application of fuzzy sets and systems as part of DSSs provides a systematic approach to addressing the uncertainty in its variables. This paper builds on prior work on DSSs utilising fuzzy cognitive maps and introduces a non-singleton fuzzification stage which directly addresses uncertainty in system inputs. The motivation of the proposed system is grounded in the real world challenges of producing high-quality olive oil and the paper provides promising application and analysis results as part of the Virgin Olive Oil Production Process.

Extractability control for the virgin olive oil elaboration process

The olive paste preparation is one of the most important task within the olive oil elaboration process. It greatly determines the quality of the obtained oil, and imposes an upper bound on the achievable extraction yield. Currently, the master miller is the person in the plant who adjusts the productive process parameters. The decisions are based on his knowledge and it depends on the proposed elaboration objective. In this work, this knowledge has been employed to design an expert system based on fuzzy logic that permits the maximization of olive oil that is extracted from the olive paste using extractability models of the olive in each maturity index. For this purpose, the system by means of computer vision estimates the olive oil amount that is susceptible of being extracted from the continuous supervision of floating olive oil on the paste, while this paste is inside the thermo-mixer, during the malaxing phase. The main advantages of this methodology are the optimization of the supplied heat to the process and the malaxing time, in order to guarantee the final product quality and quantity whereas the production cost is reduced such as the energy consumption.