Myriam Chanet

Electric impedance spectrometry for the control of manufacturing process of comminuted meat products

In order to improve the control of the composition of potted minced pork, it is necessary to estimate the water and lipid contents of the product during the manufacturing process. We propose to carry out this assessment using the electric impedance spectrometry. On-site tests show that the PLS model obtained from modulus spectra allows to estimate the water content with a standard deviation of 0.66%, and that the PLS model obtained from phase spectra allows to estimate the lipid content with a standard deviation of 1.08%. The results also suggest that it is feasible to monitor variations in water content to achieve fast corrective actions in the event of a process drift.

Modelling the rheological properties of sludge during anaerobic digestion in a batch reactor by using electrical measurements.

Anaerobic digestion is a significant process leading to biogas production and waste management. Despite this double interest, professionals still face a lack of efficient tools to monitor and manage the whole procedure. This is especially true for rheological properties of the material inside the reactor, which are of major importance for anaerobic digestion management. However, rheological properties can hardly be determined in-situ and it would be very helpful to determine indicators of their evolution. To solve this problem, this paper investigates the evolution of sewage sludge rheological and electrical properties during the anaerobic digestion in a batch reactor. We especially focus on apparent viscosity and complex impedance, measured by electrical impedance spectroscopy. Both of them can be modelled by a linear combination of raw sludge and inoculum properties, weighted by time-dependent coefficients. Thus, by determining digested sludge electrical signature, it is possible to obtain those coefficients and model sludge apparent viscosity. This work offers many theoretical and practical prospects.

Impact of Soil Structure on Microwave Volume Scattering Evaluated by a Two-Dimensional Numerical Model

Soil volumetric structure is an important parameter for tillage operation. The aim of this paper is to assess whether volume characteristics can be inferred from radar measurements. A 2-D numerical model (the 2DSCAT model) was developed based on a numerical solver using a time-domain finite-element method to solve Maxwells equations. Perfectly matched layers were implemented as well as a near- to far-field transformation. A focused incident beam was generated by adapting the boundary conditions. To represent the soil structure, a simulator was developed describing the soil as biphasic media (fine earth and clods). Clods were represented by randomly deformed ellipses, with randomly determined dimensions, locations, and orientations. The model performed successfully, as evaluated against exact analytical solutions available for an infinite perfectly conducting cylinder and the reflection of flat semi-infinite media. The model was then evaluated against measurements made by an X-band FM continuous-wave radar on a box filled with dry clods of different sizes. The effect of the clod size on the backscattering power was very well reproduced, showing the potential of using a 2-D numerical model to understand microwave-backscattering patterns from cloddy soils. Analysis of the volume scattering shows that this phenomenon can be mostly hidden in the scatter diagram by surface scattering when the latter occurred. However, the volume scattering gives a stronger residual signal in time because of propagation through the medium. Thus, time studies of the scattering signal provide further information about volume heterogeneities.

A microwave sensor for agricultural implements

For soil preparation, the concept of precision agriculture requires real time measurements of soil characteristics and cultivator behavior. The microwave sensor presented in this paper overcomes the limitations of sensors based on optical or ultrasound devices towards agricultural environment (dust, rain, etc.). The sensor uses the principle of frequency-modulated continuous wave (FM-CW) radar, low cost technology well intended for short range applications. First laboratory results of two major applications are presented. The first one is the distance measurement applied to the control of working depth. Soil-implement distances are computed with the measurement of a beat frequency: a maximum positioning error of /spl plusmn/5 mm is obtained for distances between 0.7 and 1 m. The second one is the evaluation of soil roughness in order to control seedbed quality. The backscattered coefficient, measured at normal incidence with samples of dry soil, is correlated with the rms height.

Using Proximal Sensors to Continuously Monitor Agricultural Soil Physical Conditions for Tillage Management

Homogeneity of crop establishment, which directly depends on physical conditions within the seedbed, is very important for crop yield. We did a field experiment to test the abilities of various sensors to characterise seedbed physical conditions and the possibility of continuously modifying the intensity of soil tillage with the objective of producing a uniform seedbed. The experiment was done on a silt soil with 19% clay and 74% silt in northern France. We created two initial soil structures (with and without clods >10 cm) and controlled water content (field capacity or less). A special cultivator was developed with a continuous-output GPS and a microwave ground-based radar sensor; it also carried a laser profile meter for soil surface characterisation, a capacitance probe for monitoring soil water content, and a load cell for measuring soil mechanical resistance. Each sensor was able to detect differences in soil physical conditions at the field scale. Because of the simultaneous effect of soil water content and soil structure on the geophysical parameters obtained with the sensors, it was not possible to obtain a continuous characterisation of the soil’s bulk density, water content, clod-size distribution, or surface roughness (although the use of two radar angles of incidence might allow better assessment of soil surface roughness). For tillage control, seedbed conditions depended on initial soil conditions (water content and degree of compaction) and soil tillage tool characteristics (working depth and speed of rotation). Working depth and speed of rotation had opposite effects on the size of clods at the seedbed surface; within the seedbed, they could reduce initial soil variability. Seeding rate could be controlled by the same sensors if they were put in front of the seeder. Results of the experiment are relevant to spatial parameterisation of existing soil models.

Simulation of realistic soils for 3-D computational models

This paper investigates the problem of simulation of realistic soils for electromagnetic 3-D computational models. The major difficulty is to take into account the heterogeneous nature of real-life soils, which is an aggregate of random and regular structures, due to environmental (climatic conditions, etc.) or human (soil tillage) processes. The developed methodology uses 3-D measurements of pertinent soil parameters, able to describe the heterogeneity of real-life soils. It can be broken down into four main tasks: (a) 3-D measurement of soil parameter(s); (b) characterization of the 3-D spatial distribution of soil parameter(s); (c) simulation of 3-D volumes with the same spatial distribution; and (d) simulation of 3-D soil permittivity maps using electromagnetic mixing model. This methodology is illustrated with the example of soil density measurement, which is evaluated with the use of a motor-driven penetrometer.