- Email: [email protected]
Optical online measurement technique used for process control of the drying step during pasta production
Available online at www.sciencedirect.com
Available online at www.sciencedirect.com
Procedia Food Science
Procedia – Food Science 00 (2011) 000–000
Procedia Food Science 1 (2011) 1301 – 1308 www.elsevier.com/locate/procedia
11th International Congress on Engineering and Food (ICEF11)
Optical online measurement technique used for process control of the drying step during pasta production Frauke Großa, Rainer Benninga, Ute Bindrichb, Knut Frankeb, Volker Heinzb, Antonio Delgadoa a* a
Institute of Fluid Mechanics, Technical Faculty, Friedrich-Alexander University Erlangen-Nuremberg, Cauerstr. 4, D-91058 Erlangen, Germany ([email protected], [email protected], [email protected]) b German Institute of Food Technologies, Professor-von-Klitzing-Str. 7, D-49610 Quakenbrück, Germany ([email protected], [email protected], [email protected])
Abstract The presented paper concentrates on the development of a fuzzy system to control the drying process of pasta plates. The system adjusts drying temperature as well as humidity by including knowledge of experts or from research in setting up models of the actual process respectively product state. As a basis the research led to the development of an appropriate online measurement setup that was implemented in a climate test chamber. Here relevant product information like – in this case – moisture content, colour and surface temperature of the pasta right during the running process are obtained. These three selected parameters are measured by an optical non-invasive approach with fast data acquisition. In addition to gravimetric measurements for referencing the spectroscopic approach resulting in values of the water content of pasta, magnetic resonance imaging is used to gain information of the moisture distribution within the cross section of a pasta plate. In combination with further product characteristics the quality of the product is evaluated and used in a fuzzy logic system that determines a situative and adjusted control strategy by correlating the product state with the suitable drying parameters temperature and moisture in the dryer. Adjusting the drying process with respect to the current process respectively product state realises an individual optimisation and leads to a high product quality, less product losses (due to e.g. internal thermal stresses or unwanted colour changes) and an improved energy usage. © Published by Elsevier B.V. Selection peer-review underpeer-review responsibility ofunder 11th International Congress ©2011 2011 Published by Elsevier Ltd.and/or Selection and/or responsibility of ICEF11 on Engineering and Food (ICEF 11) Executive Committee. Executive Committee Members Keywords: process optimisation; control; fuzzy logic; pasta; drying
* Corresponding author. Tel.: +499-131-852-9506; fax: +499-131-852-9503. E-mail address: [email protected].
2211–601X © 2011 Published by Elsevier B.V. Selection and/or peer-review under responsibility of 11th International Congress on Engineering and Food (ICEF 11) Executive Committee. doi:10.1016/j.profoo.2011.09.193
1302 2
Fraukename Groß/ et al. / Procedia Science 1 (2011) 1301 – 1308 Author Procedia – Food Food Science 00 (2011) 000–000
1. Introduction Automation of processes in food technology is a complex technological field due to a lack of appropriate methods to gain information from the process. Incomplete or even missing knowledge information results in the inability to define appropriate process strategies or to ensure the desired progress of a process. These characteristics can inhibit a high level of automation, which is an important tool to save not only time but also energy and to secure a high and defined quality of the product. As a consequence of the mentioned facts the presented research concentrates on the development of a fuzzy system to control the drying process of pasta plates in order to increase process efficiency. Basically, it is important to use as well offline as online analytics to define characteristic product and process properties in order to guarantee high product quality. Another major objective is to increase process efficiency by saving energy. In this context, the legally defined maximum of 13 % water content in dried pasta is defining a concrete stop criterion. The usage of this border, instead of drying to mach lower levels, is realised by a novel measurement approach and the processing of obtained data by a fuzzy logic control system. 2. Materials and Methods A laboratory optical online measurement setup is developed acquiring colour (colorimeter PR0075-A, PREMOSYS GmbH with integrated electronics, MAZeT GmbH) and surface temperature values (series BS-05T, ASM GmbH) as well as near infrared (NIR) spectroscopic data containing moisture information inside the product (PSS1720-GS1, Polytec GmbH). These three selected parameters are measured by an optical non-invasive approach with fast data acquisition. The surface temperature is detected with a pyrometer working in the infrared region and the colour with a colorimeter using white light LED (see Fig. 1).
Fig. 1. Scheme of the optical online measurement approach
All the parameters acquired online (NIR spectra, surface temperature and colour of the pasta) are implemented in a single MATLAB environment, enabling data processing (e.g. modelling of the moisture content) as well as fuzzy control and, therefore, creating and, finally, setting of output parameters, e.g. drying temperature and relative humidity in the dryer (see Fig. 2).
Frauke Großname et al./ /Procedia Procedia– Food 1 (2011) – 1308 Author Food Science Science 00 (2011)1301 000–000
Fig. 2. Scheme of the data acquisition and management in MATLAB
Furthermore, a broad offline and system adapted analytical data generation is established with respect to raw material (e.g. gluten content and particle size distribution), the dough (e.g. water activity and immobilisation, structure and strength) and the dried pasta (e.g. glass transition temperature and breaking strength). In combination with the online data a pool of characteristic properties is developed, which is processed by a fuzzy logic classifier. Thus, the description of the product state results in corresponding drying parameters in the dryer. Adjusting the drying process with respect to the current state realises an individual optimisation and leads to a high product quality, less product losses and an improved energy usage. The “diffuse” classifying by fuzzy logic enables dealing with data sets, which can not be described by conventional mathematics. Input variables are connected via operators like “and”, “or”, “not”. These aspects are then linguistically associated via “if… then”-rules to realise an adaptation of the drying parameters (air temperature, air moisture), which is based on the actual values of measured parameters. The basic structure of the fuzzy control system is shown in Fig. 3. Initially, all possible input data are classified by considering the linguistic categories "good", "medium" and "bad". Temperature and humidity of the climate test chamber and – if necessary and applicable – the speed of the drying air are defined as output parameters. They are adjusted to the values "smaller", "the same" or "higher" according to the evaluation of the rules.
Fig. 3. Principle structure of the fuzzy control system
3 1303
1304 4
Frauke Groß /etProcedia al. / Procedia Science 1 (2011) 1301 – 1308 Author name – FoodFood Science 00 (2011) 000–000
In this way a control system is built up being responsive not only to varying properties of the raw materials or the pasta dough characteristics, but also to the progress of the drying process and to desired product features. 3. Results and Discussion To measure the moisture content, the principle of diffuse reflectance spectroscopy in the near infrared region of light is used. Light in the near infrared spectrum (wavelengths: 1100-2100 nm) is focussed upon the analysed pasta plate and partially interacts with the material, which results in oscillations of dipoles within the material. The depth of the entering light depends on the wavelength of the light, the distance between projection and detection area ∆x and the angle α between light source and detecting sensor [1] (see Fig. 4, a). The resulting interactions cause different absorption respectively reflectance characteristics, where overtone oscillations of water are found as characteristic NIR absorption bands in the region of 1400 and 1900 nm [2]. The resulting profiles are analysed by a polychromator that generates absorption spectra, which is the absorption as a function of wavelength. A higher water content in the pasta plate results in higher corresponding absorption bands. With increasing duration of the drying process the peaks of the absorption bands decrease (see Fig. 4, b).
Fig. 4. (a) NIR spectroscopic setup; (b) absorption spectra of pasta plates during a drying process
Optimal signals considering the signal to noise ratio are found with an angle α of 50° or 70° between light source and detector (sensor head). To quantify the spectroscopic measurements and, consequently, the moisture content of the analysed pasta, gravimetrical measurements are used as reference values. A calibration of the laboratory measurements is realised based on the Partial Least Squares (PLS) method. Multivariate data processing suits perfectly to involve also “hidden” information of the spectra like the influence of temperature on the absorption characteristics. However, to realise a reference system not only to determine only the moisture content but also its profile, i.e. information about the depths of the measurement, it is of essential importance to set up an additional referencing technique complementing the gravimetric one. Therefore, magnetic resonance imaging (MRI) is adopted to obtain information concerning the moisture profile. These data are processed as references for further use in the optical NIR online setup. MRI is a non-invasive imaging technique commonly used to measure water distributions. In a strong magnetic field, which is combined with an alternating high frequency field (radio frequency), protons show magnetic resonance. The magnetic moment of the nuclei causes an alignment of the proton in the magnetic field. The nuclei precess if they are disturbed from equilibrium and thus induce a voltage in the measuring tuned coil, which can be
FraukeAuthor Groß et al. // Procedia (2011) – 1308 name Procedia –Food FoodScience Science100 (2011)1301 000–000
detected. The relaxation time, respectively realignment, is characteristic for the detected element and bonding [3]. The obtained signal intensity correlates to the lightness respectively light intensity in the MR images. At the Department of Experimental and Clinical Pharmacology and Toxicology, Institute for Pharmacology and Toxicology in Erlangen, MRI measurements are realised in a 4.7 T tomograph “Biospec” (Bruker BioSpin GmbH; see Fig. 5, a). The measurement of pasta plates requires the use of a coil being suitable and appropriate for this particular application. As in the said tomograph usually small animals are under investigation, a special method of the measurement and the related settings had to be developed and established adapting to the shape of thin dough, which thickness of only one millimetre should be resolved properly. A mechanical setup (see Fig. 5, b) was developed for NIR measurements directly in the tomograph to realise both MRI and NIR measurements at the same time.
Fig. 5. (a) Bruker BioSpec; (b) device for NIR measurements in the tomograph
Magnetic resonance imaging visualises the decrease of the water content within the analysed pasta plates during the drying process starting at the surface (see Fig. 6).
Fig. 6. Combined NIR and MRI measurements to visualize the drying process of pasta; 10 minutes duration of the measurements (1); 220 minutes duration of the measurements (2); 500 minutes duration of the measurements (3)
The laboratory setup is transferred to a climate test chamber in order to investigate the optical measurement technique under defined temperature and humidity profiles. Mainly two industrially applied high temperature drying profiles are performed in this drying chamber, which are much more efficient than classic low temperature drying profiles lasting twice as long (see Fig. 7).
51305
Frauke Groß et al. / Procedia Food Science 1 (2011) 1301 – 1308 Author name / Procedia – Food Science 00 (2011) 000–000
6
100 temperature [°C] / relative humidity [%]
1306
90 80 70
temperature LT temperature HT temperature reduced HT relative humidity LT relative humidity HT relative humidity reduced HT
60 50 40 30 20 10 0 0
200
400
600
time [min]
Fig. 7. (a) Drying profiles for low (LT), high (HT) and reduced high temperature drying; (b) climate test chamber and optical setup in the climate test chamber
During one drying experiment the colour values L* (lightness), a* (red/green component), b*(yellow/blue component) and the surface temperature of the pasta are registered as well as the relative humidity and temperature in the dryer and the modelled moisture content derived from the NIR spectra (see Fig. 8). During the first hour of drying, the drying temperature is increased to 80 °C in order to remove the free water in the pasta dough structure. Then the temperature is kept constant until the moisture content decreased to a level between 12 and 10 % (legal maximum 13 %). The relative drying humidity is only decreased slightly from 80 % at the beginning to 77 %. This sets the thermodynamically optimal conditions considering the avoidance of thermal stresses and breaks in the pasta. The colour values stay more or less the same during the whole drying process. Most important is to observe the a*-value because an increase indicates a higher reddish component due to Maillard reactions.
Fig. 8. Characteristic profiles of the online measured values (colour, L*, a*, b*; surface temperature, drying temperature and relative humidity; moisture content) during a reduced high temperature drying process in the climate test chamber
1307 7
Frauke Groß etname al. / Procedia Science 1 (2011) 1301 – 1308 Author / ProcediaFood – Food Science 00 (2011) 000–000
The intrinsic knowledge of drying, given by sophisticated drying profiles, and extended offline experiments are a part of the basis for creating the fuzzy system, which is completed by the described online measurement methods. The elaborated control is able to react to changes, e.g. in the input conditions, like an increase in the initial water content of pasta (see Fig. 9).
L* [-] a* [-] b* [-] surface temperatur [°C] relative humidity [%] drying temperature [°C] moisture content [%]
110 105 100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 -5 2 -10
22
42
62
82
102
122
142
162
182
202
time [min]
Fig. 9. Characteristic profiles of the online measured values (colour, L*, a*, b*; surface temperature, drying temperature and relative humidity; moisture content) during fuzzy controlled drying experiment in the climate test chamber
Due to the higher water content (35 % instead of 30 %) the fuzzy control system sets a much lower and softer temperature gradient and reaches the defined maximum of 90 °C after only two hours of drying when an uncritical amount of water between 18 and 15 % (where most of the free water has been removed) is definitively reached. In order to save energy and time it is of great interest to stay close to the legal maximum moisture content of 13 %. As soon as a stable moisture value of approx. 13 % is reached, the drying process is stopped by the fuzzy system. 4. Conclusion An optical online measurement setup is developed, which acquires, besides colour and surface temperature, near infrared (NIR) spectroscopic data containing moisture information from inside the product (moisture profile). This setup is adjusted and recalibrated in a climate test chamber under defined temperature and moisture conditions. Based on this online data combined with the offline analytics, e.g. concerning product properties, and knowledge of pasta experts a fuzzy system has been developed to control the drying process of pasta plates taking into account product quality and economy respectively ecology. Acknowledgements The authors would like to acknowledge the Department of Experimental and Clinical Pharmacology and Toxicology, Institute for Pharmacology and Toxicology for the allowance to use the equipment to
1308
Frauke Groß et al. / Procedia Food Science 1 (2011) 1301 – 1308 Author name / Procedia – Food Science 00 (2011) 000–000
8
realise magnetic resonance imaging and Dr. Andreas Hess and Dr. Lubos Budinsky for their great support in implementing and performing all the measurements. This research project was supported by the FEI (Forschungskreis der Ernährungsindustrie e.V., Bonn), the AiF and the Ministry of Economics and Labour. Project No: 284 ZN. References [1] Okada E., Firbank M., Delpy DT. The effect of overlying tissue on the spatial sensitivity profile of near-infrared spectroscopy, Phys. Med. Biol.; 1995, Vol. 40, pp. 2093–2108 [2] Ellis JW, Bath J. Modifications in the Near Infra-Red Absorption Spectra of Protein and of Light and Heavy Water Molecules When Water is Bound to Gelatin, Journal of chemical physics; 1938, Vol. 6, pp. 723 [3] Richardson JC, Bowtell W, Mäder K., Melia CD. Pharmaceutical applications of magnetic resonance imaging (MRI), Advanced Drug Delivery Reviews; 2005, Vol. 57, pp.1191–1209
Presented at ICEF11 (May 22-26, 2011 – Athens, Greece) as paper AFT855.
Available online at www.sciencedirect.com
Procedia Food Science
Procedia – Food Science 00 (2011) 000–000
Procedia Food Science 1 (2011) 1301 – 1308 www.elsevier.com/locate/procedia
11th International Congress on Engineering and Food (ICEF11)
Optical online measurement technique used for process control of the drying step during pasta production Frauke Großa, Rainer Benninga, Ute Bindrichb, Knut Frankeb, Volker Heinzb, Antonio Delgadoa a* a
Institute of Fluid Mechanics, Technical Faculty, Friedrich-Alexander University Erlangen-Nuremberg, Cauerstr. 4, D-91058 Erlangen, Germany ([email protected], [email protected], [email protected]) b German Institute of Food Technologies, Professor-von-Klitzing-Str. 7, D-49610 Quakenbrück, Germany ([email protected], [email protected], [email protected])
Abstract The presented paper concentrates on the development of a fuzzy system to control the drying process of pasta plates. The system adjusts drying temperature as well as humidity by including knowledge of experts or from research in setting up models of the actual process respectively product state. As a basis the research led to the development of an appropriate online measurement setup that was implemented in a climate test chamber. Here relevant product information like – in this case – moisture content, colour and surface temperature of the pasta right during the running process are obtained. These three selected parameters are measured by an optical non-invasive approach with fast data acquisition. In addition to gravimetric measurements for referencing the spectroscopic approach resulting in values of the water content of pasta, magnetic resonance imaging is used to gain information of the moisture distribution within the cross section of a pasta plate. In combination with further product characteristics the quality of the product is evaluated and used in a fuzzy logic system that determines a situative and adjusted control strategy by correlating the product state with the suitable drying parameters temperature and moisture in the dryer. Adjusting the drying process with respect to the current process respectively product state realises an individual optimisation and leads to a high product quality, less product losses (due to e.g. internal thermal stresses or unwanted colour changes) and an improved energy usage. © Published by Elsevier B.V. Selection peer-review underpeer-review responsibility ofunder 11th International Congress ©2011 2011 Published by Elsevier Ltd.and/or Selection and/or responsibility of ICEF11 on Engineering and Food (ICEF 11) Executive Committee. Executive Committee Members Keywords: process optimisation; control; fuzzy logic; pasta; drying
* Corresponding author. Tel.: +499-131-852-9506; fax: +499-131-852-9503. E-mail address: [email protected].
2211–601X © 2011 Published by Elsevier B.V. Selection and/or peer-review under responsibility of 11th International Congress on Engineering and Food (ICEF 11) Executive Committee. doi:10.1016/j.profoo.2011.09.193
1302 2
Fraukename Groß/ et al. / Procedia Science 1 (2011) 1301 – 1308 Author Procedia – Food Food Science 00 (2011) 000–000
1. Introduction Automation of processes in food technology is a complex technological field due to a lack of appropriate methods to gain information from the process. Incomplete or even missing knowledge information results in the inability to define appropriate process strategies or to ensure the desired progress of a process. These characteristics can inhibit a high level of automation, which is an important tool to save not only time but also energy and to secure a high and defined quality of the product. As a consequence of the mentioned facts the presented research concentrates on the development of a fuzzy system to control the drying process of pasta plates in order to increase process efficiency. Basically, it is important to use as well offline as online analytics to define characteristic product and process properties in order to guarantee high product quality. Another major objective is to increase process efficiency by saving energy. In this context, the legally defined maximum of 13 % water content in dried pasta is defining a concrete stop criterion. The usage of this border, instead of drying to mach lower levels, is realised by a novel measurement approach and the processing of obtained data by a fuzzy logic control system. 2. Materials and Methods A laboratory optical online measurement setup is developed acquiring colour (colorimeter PR0075-A, PREMOSYS GmbH with integrated electronics, MAZeT GmbH) and surface temperature values (series BS-05T, ASM GmbH) as well as near infrared (NIR) spectroscopic data containing moisture information inside the product (PSS1720-GS1, Polytec GmbH). These three selected parameters are measured by an optical non-invasive approach with fast data acquisition. The surface temperature is detected with a pyrometer working in the infrared region and the colour with a colorimeter using white light LED (see Fig. 1).
Fig. 1. Scheme of the optical online measurement approach
All the parameters acquired online (NIR spectra, surface temperature and colour of the pasta) are implemented in a single MATLAB environment, enabling data processing (e.g. modelling of the moisture content) as well as fuzzy control and, therefore, creating and, finally, setting of output parameters, e.g. drying temperature and relative humidity in the dryer (see Fig. 2).
Frauke Großname et al./ /Procedia Procedia– Food 1 (2011) – 1308 Author Food Science Science 00 (2011)1301 000–000
Fig. 2. Scheme of the data acquisition and management in MATLAB
Furthermore, a broad offline and system adapted analytical data generation is established with respect to raw material (e.g. gluten content and particle size distribution), the dough (e.g. water activity and immobilisation, structure and strength) and the dried pasta (e.g. glass transition temperature and breaking strength). In combination with the online data a pool of characteristic properties is developed, which is processed by a fuzzy logic classifier. Thus, the description of the product state results in corresponding drying parameters in the dryer. Adjusting the drying process with respect to the current state realises an individual optimisation and leads to a high product quality, less product losses and an improved energy usage. The “diffuse” classifying by fuzzy logic enables dealing with data sets, which can not be described by conventional mathematics. Input variables are connected via operators like “and”, “or”, “not”. These aspects are then linguistically associated via “if… then”-rules to realise an adaptation of the drying parameters (air temperature, air moisture), which is based on the actual values of measured parameters. The basic structure of the fuzzy control system is shown in Fig. 3. Initially, all possible input data are classified by considering the linguistic categories "good", "medium" and "bad". Temperature and humidity of the climate test chamber and – if necessary and applicable – the speed of the drying air are defined as output parameters. They are adjusted to the values "smaller", "the same" or "higher" according to the evaluation of the rules.
Fig. 3. Principle structure of the fuzzy control system
3 1303
1304 4
Frauke Groß /etProcedia al. / Procedia Science 1 (2011) 1301 – 1308 Author name – FoodFood Science 00 (2011) 000–000
In this way a control system is built up being responsive not only to varying properties of the raw materials or the pasta dough characteristics, but also to the progress of the drying process and to desired product features. 3. Results and Discussion To measure the moisture content, the principle of diffuse reflectance spectroscopy in the near infrared region of light is used. Light in the near infrared spectrum (wavelengths: 1100-2100 nm) is focussed upon the analysed pasta plate and partially interacts with the material, which results in oscillations of dipoles within the material. The depth of the entering light depends on the wavelength of the light, the distance between projection and detection area ∆x and the angle α between light source and detecting sensor [1] (see Fig. 4, a). The resulting interactions cause different absorption respectively reflectance characteristics, where overtone oscillations of water are found as characteristic NIR absorption bands in the region of 1400 and 1900 nm [2]. The resulting profiles are analysed by a polychromator that generates absorption spectra, which is the absorption as a function of wavelength. A higher water content in the pasta plate results in higher corresponding absorption bands. With increasing duration of the drying process the peaks of the absorption bands decrease (see Fig. 4, b).
Fig. 4. (a) NIR spectroscopic setup; (b) absorption spectra of pasta plates during a drying process
Optimal signals considering the signal to noise ratio are found with an angle α of 50° or 70° between light source and detector (sensor head). To quantify the spectroscopic measurements and, consequently, the moisture content of the analysed pasta, gravimetrical measurements are used as reference values. A calibration of the laboratory measurements is realised based on the Partial Least Squares (PLS) method. Multivariate data processing suits perfectly to involve also “hidden” information of the spectra like the influence of temperature on the absorption characteristics. However, to realise a reference system not only to determine only the moisture content but also its profile, i.e. information about the depths of the measurement, it is of essential importance to set up an additional referencing technique complementing the gravimetric one. Therefore, magnetic resonance imaging (MRI) is adopted to obtain information concerning the moisture profile. These data are processed as references for further use in the optical NIR online setup. MRI is a non-invasive imaging technique commonly used to measure water distributions. In a strong magnetic field, which is combined with an alternating high frequency field (radio frequency), protons show magnetic resonance. The magnetic moment of the nuclei causes an alignment of the proton in the magnetic field. The nuclei precess if they are disturbed from equilibrium and thus induce a voltage in the measuring tuned coil, which can be
FraukeAuthor Groß et al. // Procedia (2011) – 1308 name Procedia –Food FoodScience Science100 (2011)1301 000–000
detected. The relaxation time, respectively realignment, is characteristic for the detected element and bonding [3]. The obtained signal intensity correlates to the lightness respectively light intensity in the MR images. At the Department of Experimental and Clinical Pharmacology and Toxicology, Institute for Pharmacology and Toxicology in Erlangen, MRI measurements are realised in a 4.7 T tomograph “Biospec” (Bruker BioSpin GmbH; see Fig. 5, a). The measurement of pasta plates requires the use of a coil being suitable and appropriate for this particular application. As in the said tomograph usually small animals are under investigation, a special method of the measurement and the related settings had to be developed and established adapting to the shape of thin dough, which thickness of only one millimetre should be resolved properly. A mechanical setup (see Fig. 5, b) was developed for NIR measurements directly in the tomograph to realise both MRI and NIR measurements at the same time.
Fig. 5. (a) Bruker BioSpec; (b) device for NIR measurements in the tomograph
Magnetic resonance imaging visualises the decrease of the water content within the analysed pasta plates during the drying process starting at the surface (see Fig. 6).
Fig. 6. Combined NIR and MRI measurements to visualize the drying process of pasta; 10 minutes duration of the measurements (1); 220 minutes duration of the measurements (2); 500 minutes duration of the measurements (3)
The laboratory setup is transferred to a climate test chamber in order to investigate the optical measurement technique under defined temperature and humidity profiles. Mainly two industrially applied high temperature drying profiles are performed in this drying chamber, which are much more efficient than classic low temperature drying profiles lasting twice as long (see Fig. 7).
51305
Frauke Groß et al. / Procedia Food Science 1 (2011) 1301 – 1308 Author name / Procedia – Food Science 00 (2011) 000–000
6
100 temperature [°C] / relative humidity [%]
1306
90 80 70
temperature LT temperature HT temperature reduced HT relative humidity LT relative humidity HT relative humidity reduced HT
60 50 40 30 20 10 0 0
200
400
600
time [min]
Fig. 7. (a) Drying profiles for low (LT), high (HT) and reduced high temperature drying; (b) climate test chamber and optical setup in the climate test chamber
During one drying experiment the colour values L* (lightness), a* (red/green component), b*(yellow/blue component) and the surface temperature of the pasta are registered as well as the relative humidity and temperature in the dryer and the modelled moisture content derived from the NIR spectra (see Fig. 8). During the first hour of drying, the drying temperature is increased to 80 °C in order to remove the free water in the pasta dough structure. Then the temperature is kept constant until the moisture content decreased to a level between 12 and 10 % (legal maximum 13 %). The relative drying humidity is only decreased slightly from 80 % at the beginning to 77 %. This sets the thermodynamically optimal conditions considering the avoidance of thermal stresses and breaks in the pasta. The colour values stay more or less the same during the whole drying process. Most important is to observe the a*-value because an increase indicates a higher reddish component due to Maillard reactions.
Fig. 8. Characteristic profiles of the online measured values (colour, L*, a*, b*; surface temperature, drying temperature and relative humidity; moisture content) during a reduced high temperature drying process in the climate test chamber
1307 7
Frauke Groß etname al. / Procedia Science 1 (2011) 1301 – 1308 Author / ProcediaFood – Food Science 00 (2011) 000–000
The intrinsic knowledge of drying, given by sophisticated drying profiles, and extended offline experiments are a part of the basis for creating the fuzzy system, which is completed by the described online measurement methods. The elaborated control is able to react to changes, e.g. in the input conditions, like an increase in the initial water content of pasta (see Fig. 9).
L* [-] a* [-] b* [-] surface temperatur [°C] relative humidity [%] drying temperature [°C] moisture content [%]
110 105 100 95 90 85 80 75 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 -5 2 -10
22
42
62
82
102
122
142
162
182
202
time [min]
Fig. 9. Characteristic profiles of the online measured values (colour, L*, a*, b*; surface temperature, drying temperature and relative humidity; moisture content) during fuzzy controlled drying experiment in the climate test chamber
Due to the higher water content (35 % instead of 30 %) the fuzzy control system sets a much lower and softer temperature gradient and reaches the defined maximum of 90 °C after only two hours of drying when an uncritical amount of water between 18 and 15 % (where most of the free water has been removed) is definitively reached. In order to save energy and time it is of great interest to stay close to the legal maximum moisture content of 13 %. As soon as a stable moisture value of approx. 13 % is reached, the drying process is stopped by the fuzzy system. 4. Conclusion An optical online measurement setup is developed, which acquires, besides colour and surface temperature, near infrared (NIR) spectroscopic data containing moisture information from inside the product (moisture profile). This setup is adjusted and recalibrated in a climate test chamber under defined temperature and moisture conditions. Based on this online data combined with the offline analytics, e.g. concerning product properties, and knowledge of pasta experts a fuzzy system has been developed to control the drying process of pasta plates taking into account product quality and economy respectively ecology. Acknowledgements The authors would like to acknowledge the Department of Experimental and Clinical Pharmacology and Toxicology, Institute for Pharmacology and Toxicology for the allowance to use the equipment to
1308
Frauke Groß et al. / Procedia Food Science 1 (2011) 1301 – 1308 Author name / Procedia – Food Science 00 (2011) 000–000
8
realise magnetic resonance imaging and Dr. Andreas Hess and Dr. Lubos Budinsky for their great support in implementing and performing all the measurements. This research project was supported by the FEI (Forschungskreis der Ernährungsindustrie e.V., Bonn), the AiF and the Ministry of Economics and Labour. Project No: 284 ZN. References [1] Okada E., Firbank M., Delpy DT. The effect of overlying tissue on the spatial sensitivity profile of near-infrared spectroscopy, Phys. Med. Biol.; 1995, Vol. 40, pp. 2093–2108 [2] Ellis JW, Bath J. Modifications in the Near Infra-Red Absorption Spectra of Protein and of Light and Heavy Water Molecules When Water is Bound to Gelatin, Journal of chemical physics; 1938, Vol. 6, pp. 723 [3] Richardson JC, Bowtell W, Mäder K., Melia CD. Pharmaceutical applications of magnetic resonance imaging (MRI), Advanced Drug Delivery Reviews; 2005, Vol. 57, pp.1191–1209
Presented at ICEF11 (May 22-26, 2011 – Athens, Greece) as paper AFT855.