EXTENSION OF THE NEURAL NETWORKS OPERATING RANGE BY THE APPLICATION OF DIMENSIONLESS NUMBERS IN PREDICTION OF HEAT TRANSFER COEFFICIENTS

ABSTRACT

The paper presents a study aimed at extending the neural network mapping ability. In traditional modelling, operational process parameters (gas/material temperature, air velocity, etc.) are the inputs and outputs to and from the network. In this approach dimensionless numbers (Re, Ar, H/d) were used as inputs to predict the heat transfer coefficient in a fluidised bed drying process. To produce the data set necessary to train the networks, drying trials of different materials in a fluidised bed were carried out.

A series of simulations were performed and several neural networks structures were tested to find an optimal topology of the network. Training data set contained information only about two materials. The networks were tested using data obtained for the third product.

Performance of the network was satisfactory, however further improvement of mapping ability may be expected after filtration of the testing data.     

HYBRID NEURAL MODELLING OF FLUIDISED BED DRYING PROCESS

The aim of this study was to investigate the applicability of hybrid neural models in modelling of drying process. A study aimed at extending a neural network mapping was also carried out. In this approach dimensionless numbers (Re, Ar, H/d) were used as inputs to predict the heat transfer coefficient in a fluidised bed drying process. To produce a data set necessary to train the networks, trials of drying different materials in a fluidised bed were carried out. On the basis of this network, a hybrid model describing the process of drying in a fluidised bed dryer was built. Results obtained were compared not only with available experimental data but also with results obtained using other types of models: a pseudo-dynamic neural model and a classical mathematical model. The analysis of results leads to a conclusion that hybrid models constitute a solid alternative method of process modelling.     

HYBRID NEURAL MODELLING OF FLUIDISED BED DRYING PROCESS

The aim of this study was to investigate the applicability of hybrid neural models in modelling of drying process. A study aimed at extending a neural network mapping was also carried out. In this approach dimensionless numbers (Re, Ar, H/d) were used as inputs to predict the heat transfer coefficient in a fluidised bed drying process. To produce a data set necessary to train the networks, trials of drying different materials in a fluidised bed were carried out. On the basis of this network, a hybrid model describing the process of drying in a fluidised bed dryer was built. Results obtained were compared not only with available experimental data but also with results obtained using other types of models: a pseudo-dynamic neural model and a classical mathematical model. The analysis of results leads to a conclusion that hybrid models constitute a solid alternative method of process modelling.     

PROCESSES RELATED TO DRYING: PART II USE OF THE SAME MODEL TO SOLVE TRANSFERS BOTH IN SATURATED AND UNSATURATED POROUS MEDIA

From the mathematical formulation presented in part I, a numerical code is developed to simulate heat and: mass transfers in porous media. The aim of this· tool is to understand and to improve each process related to drying. The association of a comprehensive set of equations with a efficient 2-D computer code allows us to predict the comportment of several porous media even if submitted to severe drying conditions. A few runs have been selected with special attention paid to the effect of internal gaseous pressure:

Convective drying of softwood at high temperature illustrate the typical two-dimensional transfers that occur in an anisotropic medium.

Microwave drying of light concrete pinpoints liquid expulsion of water which is driven by the pressure due to internal heating.

Finally, appropriate physical behaviours of a bed of glass spheres allows one to deal with simple processes for which full saturation occurs.     

DRYING OF SUSPENSIONS IN A MODIFIED SPOUTED BED DRIER WITH AN INERT PACKING

Authors summarize results of drying experiments carried out in aim to investigate effects of operational and process-conditions on the drying output and the product quality. The studies were performed in a laboratory-size spouted bed drier of a new construction with tangentional air inlet and with an inner conveyor screw maintaining the spouted bed recirculation of the inert particles

Effects of parameters assuring intensive gas-solid contact were studied, such as the inlet air-velocity, the recirculation of inert particles as well as chemical and physical properties of the inert material.

Various chemical suspensions (e.g. fine grain inorganic-salt pulps, waste products), biologically active materials (e.g, suspension of yeast) food products were used in preliminary experiments which showed effectiveness of this dryin5 method. On this basis two rode1 materials (cobalt-carbonate suspension and pulp of zinc--carbonate) were chosen for detailed experiments.

As a result cf these studies main optimal operational and process pszramters.have been obtained.     

KINETICS OF THE CLAY ROOFING TILE CONVECTION DRYING

Kinetics of the convection drying process of flat tile has been investigated experimentally in an industrial tunnel dryer. Several velocities of wet tile movement through the dryer were tested to obtain maximum allowable drying rate curve. As there are various models to describe the kinetics of convection drying, finding a model that would fairly well approximate the kinetics of the whole drying process was part of the research. Especially the polynomial and exponential models were tested.

It was found that exponential model of the type: B(t) = (a-B_e. EXP(-bt²) +B_et (-dB(t)/dt) = 2bt(B(t)-B_e significantly correlates the kinetics of the whole tile drying process.

Applying the maximum allowable drying rate curve obtained for flat tile in the first period of drying, a grapho-analytic model for the optimal conducting of the process has been developed.     

KINETICS OF DRYING IN CYCLICALLY SHIFTED SPOUTED BED

This paper presents kinetics of batchwise drying in a spouted bed apparatus with cyclically shifted gas stream, i.e. drying medium flowrate is changed cyclically along inlet cross-section of a bed during operation time.

Such an apparatus provides advantages of both spouted and oscillating beds which include: good hydrodynamic conditions within the bed and energy savings. Two types of particulate materials, i.e. beet and wheat seeds, which differ largely in physical properties were used in experiments.

Based on the results of this investigation formulae for reduced moisture content (X^red) in dependence on drying time and air temperature were developed.     

TWO STAGE DRYING OF SYNTHETIC DETERGENTS

A key ingredient in granular synthetic detergents is Sodium Tripolyphosphate (STP). The hexahydrate of STP is unstable during the spray drying process. The decomposition (reversion) is shown to be solely a function of particle temperature. The reversion products interact with calcium ions in the wash water and form insoluble species. This results in undesirable deposition onto the fabrics. The goal of this study is to develop a drying process with lower temperatures to reduce the decomposition. This, in turn, will improve product quality by reducing undesirable fabric deposits.

A two stage drying process is used to reduce the drying temperatures. Experimental evidence is presented to demonstrate the near-elimination of STP decomposition.

The design of a production scale fluid bed dryer uses a ‘normalization'procedure to calculate the rate constants at constant bed temperature from the data taken at constant inlet air temperature. The rate constants are corrleated to bed temperature using an Arrhenius-like equation.     

Intelligent Computation of Moisture Content in Shrinkable Biomaterials

A technique of intelligent computation of moisture content in shrinkable biomaterials from multiple predictors was developed. All measurable predictors were structured in three sets: biomaterial properties (volume, density, porosity, diffusivity); drying conditions (time, air temperature, humidity, velocity, pressure); and drying technologies. Two typical drying models were considered: time-dependent (thermodynamical) and time-independent (relational). The relationship between predictors and moisture content was established on the basis of multi-factorial linear regression (MLR) and neural networks (NN). Accuracy of statistical approximation was strongly dependent on drying model and chosen set of predictors. Time-independent models demonstrated better accuracy (MSE = 0.214) than time-dependent models (MSE = 0.254). Redundant predictors did not affect the accuracy and generalization ability of statistical models.

Results of NN training and testing showed superior accuracy with respect to statistical models. NN worked perfectly well for any combination of non-correlated predictors, improving accuracy to MSE = 0.01. Elimination of redundant predictors further improved accuracy and generalization ability of NN models.

The performance of both models was tested for drying of ginseng roots in the range of temperatures from 38 to 50°C, sizes from 10 to 32 mm, and relative humidity from 12 to 40%. Due to the high accuracy and computational efficiency, NN can be used as online estimator of moisture content in drying process.     

NEURAL-NETWORK-AIDED DESIGN OF AUTOMOBILE EXHAUST CATALYSTS

A priori design of catalysts is not yet possible. Such task would demand unavailable scientific knowledge of the correlations among synthesis parameters and resulting solid state and surface structures, on the one hand, and among those atomic-level structural details and their catalytic functions, on the other hand. To avoid testing every possible combination, therefore, the applied chemist or chemical engineer must identify empirical correlations underlying the existing experimental data base.

The ability of artificial neural networks to identify complex correlations and to predict the result of experiments has recently generated considerable interest in various areas of science and engineering. In this paper, neural networks are used to identify composition-performance relationships in automobile exhaust catalysts.

This work employs an artificial neural network technique to do a sensitivity analysis of the conversions of pollutant gases as a function of the catalyst composition and the operating conditions. This approach converges on the optimum catalyst composition and operating condition in order to produce specified conversions of carbon monoxide, hydrocarbons and nitrogen oxides, to carbon dioxide, water and di-nitrogen respectively.     

UNSTEADY, TWO-DIMENSIONAL HEAT AND MASS TRANSFER IN A PACKED BED OF FINE PARTICLES WITH AN ENDOTHERMIC PROCESS

The basic differential equations controlling the temperature and concentration field in a single packed bed of fine particles were derived and solved for the general case in which unsteady, two-dimensional heat and mass transfer lakes place with an endothermic process.

The time-change of particle- and fluid-temperature and concentration of water vapor (humidity) were calculated by a numerical method which assumed that the rate of the endothermic process can be expressed by a first-order rate equation and that the fluid flowing through the bed is of the piston flow type.

The experiments were conducted for the drying of silica-gel and the two-stage dehydration reaction of natural gypsum to demonstrate the applicability of the present theoretical analysis.

It has been found that the calculated results show satisfactory agreement with the measured data within the range of the experimental conditions employed.     

A STUDY OF THE PRIMARY PHASE OF FOOD FREEZE-DRYING IN VACUO

In the development of a novel freeze-drying technique in a fluidized bed at atmospheric pressure, a parallel study was undertaken using a conventional vacuum equipment. Two kinetically distinct phases were observed during freeze drying of representative Pood samples:

(1) a period during which the rate of drying was constant and (2) a second period during which the drying rate sufferedcontinual reduction. This paper focused attention on the primary drying period which corresponded with the kinetics of sublimation of pure     

THE INFLUENCE OF PRESSURE AND TEMPERATURE ON THE KINETICS OF VACUUM DRYING OF KETOPROFEN

This paper explores the influence of temperature and pressure on drying kinetics of 2-(3-benzoylphenil propionic acid) ketoprofen, in a vacuum dryer on laboratory scale, Experimentally determined relations between moisture content and drying rate vs time, were approximated with an exponential model. Model parameters were correlated with drying conditions (temperature, pressure) and defined by functions of their potentions.

From an energy balance of the process, a mathematical model for simulating dependence of sample temperature vs drying time, and moisture content of material, has been developed.

Simulation of the drying kinetics and sample temperature, by use of those functional dependencies shows good agreement with experimental results.     

INFERENTIAL SENSORS FOR ON-LINE MONITORING OF TISSUE MACHINE QUALITY PROPERTIES

There are many processes in a pulp and paper mill where an on-line parameter analyzer cannot be used due to several reasons

•The analyzer is very expensive

•It cannot survive in the environment we want to use it

•It is not operational due to hardware problems, maintenance etc

•Such an analyzer does not exist in the market

In all these situations it would be great for the mill to have an alternative way of measuring those parameters in real-time. Neural network models can serve as virtual sensors that infer process parameters from other variables, which can be measured on-line. One excellent application of inferential sensors in the pulp and paper industry is the on-line prediction of paper properties, like tensile, stretch, brightness, opacity etc. In tissue machines, the most important quality parameter is softness, which is usually measured in a very subjective manner by the touch of a human finger. In this work we examine how neural networks can be deployed in order to build online virtual sensors for softness and other tissue quality properties. The results are promising and show that neural network technology can improve productivity and minimize out of specs production in a tissue machine, by providing accurate real time monitoring ofquality parameters.     

SOLAR DRYING OF GRAPES

Mass losses and low quality are the most serious disadvantages of traditional grape drying methods. For the production of high quality raisins an increase in the drying rate is required and the grapes should be protected from rain, dust and insects during drying.

Under the terms of a joint German-Greek research program low-cost solar grape dryers were developed in the Institute of Agricultural Engineering of Hohenheim University and were tested in Greece in cooperation with the Crete Agricultural Research Center.

The required data basis for optimizing solar grape dryers was obtained by additional laboratory tests measuring the influence of various drying parameters on drying rate and quality.

Tests with the solar dryers have shown that it is possible to reduce the drying time and improve the quality significantly compared to the traditional drying methods. Also mass losses due to rain can be prevented.     

OVERVIEW AND SYSTEM ANALYSIS OF VARIOUS SEWAGE SLUDGE DRYING PROCESSES

An incineration process is routinely used in Japan to treat nearly all the generated sewage sludge. The drying process now is recognized to play an important role as a pretreatment process of a incineration process.

This paper provides a brief introduction to the conventionally utilized drying methods/equipment (i.e, hot gas drying, fluidized bed drying, pneumatic conveyor drying, and steam drying( for treating dewatered sludge cakes.     

DEVELOPMENT AND VALIDATION OF A PROCESS SIMULATOR FOR DRYING SUBBITUMINOUS COAL

Drying subbituminous coal has never been practiced commercially. The commercial dryers built to date have been designed for drying surface moisture in conjunction with upstream coal preparation facilities. This type of drying is mainly controlled by input energy and the basis of the design is an energy balance. In drying inherent moisture from subbituminous coal, the thermal conductivity of the coal and the diffusion of molecular water within coal particles impose limitations on the process conditions. Energy input and solids residence time in the dryer have to be controlled properly for simultaneously balancing the heat and mass transfer within the coal particles. Improper control of either parameter can cause fires and explosions during the key steps of the drying process—drying and cooling

In parallel to the Anaconda coal drying pilot plant program, a cross-flow, fluid-bed coal drying/cooling process simulator was developed for: (1) understanding the drying phenomena on an individual particle basis; (2) analyzing potential risks and safety limits, and (3) designing the Anaconda pilot plant program

The development of the process simulator was based on both first principles and laboratory data and can be divided into two phases:

1 Development of a semi-mechanistic drying model for Powder River Basin subbituminous coal employing an analytical solution of the diffusion equation

2.Formulation of a fluid-bed cross-bed cross-flow dryer/cooler simulator employing simultaneous heat and mass transfer

This model was validated against process variables data taken on a 4 tph pilot plant. An operable range, or process envelope, has been developed through the pilot plant experience and the process simulation study. Based on the model predictions, an uncertainly range was defined in the design recommendations of a pioneer coal drying plant in scale-up.     

STEAM DRYLNG OF WOOD FUELS

The objective of this work is to investigate some of the important aspects in the design of a steam band dryer for wood fuels. For this purpose the drying of the material in a bench-scale fixed bed dryer has been studied.

Drying times and thermal efficiencies for experiments conducted under different conditions are compared. The investigated materials are soft-and hardwood chips and softwood bark.

The thermal efficiency, the part of the sensible heat which is used during one passage of the steam through the bed, increases with increasing mass load (mass of dry matter per unit area) and with decreasing steam mass flux. At a mass load of 30 kg/m the thermal efficiency is about 0.85 even at steam mass fluxes as high as 0.6 kg/m² s (1.2 m/s). The thermal efficiency proves to be almost independent of pressure and temperature of the steam.

Due to the very inhomogeneous materials the steam mass flow distribution was uneven. This causes a decreased thermal efficiency. When bark     

A STUDY OF CLOSED CYCLE DRYING FOR THE COATED FILM WITH ORGANIC SOLVENT ON THE CONTINUOUS SHEET MATERIALS

The conventional drying of the coated film with organic solvent on continuous sheet materials is usually performed in open or one pass drying system using air as the drying medium. The concentration of evaporated organic solvent in the drying system must be low enough to prevent explosion and large volume of off gas from dryer to solvent recovery system is required, resulting in poor heat economy for all the plant

To improve the heat economy of the plant it had been proposed to use a closed cycle drying system, where an inert gas, e.g. nitrogen, is used as the drying medium. High concentration of the organic solvent in the recycle gas mixture may then be used, which results in smaller volume of recycle gas and possibility of applying an inexpensive solvent recovery system of dehumidification. This investigation includes two important problems to realize the closed cycle drying

(1)Measurement of solvent (toluene) evaporation rate from coated film in the gas mixture of toluene andnitrogen over a wide range of toluene/nitrogen ratios (0-1.0) and drying temperature using bench scaleapparatus

(2)Development of a contactless sealing method, using an inert gas, for closed cycle drying of continuoussheet materials. The results of preliminary tests demonstrate its feasibility.     

RECIRCULATING REACTOR FOR THE STUDY OF MULTIPHASE KINETICS

A new laboratory reactor was set up to measure kinetic coefficients in a solid (catalyst)-liquid-gas reacting system.

The reactor consists of two parts: an absorber, where the liquid is partially saturated by the gas reactant and a reacting zone, where the liquid alone, containing the dissolved gas, flows through a fixed bed of catalyst.

The ricircle of the liquid in the absorber maintains a high concentration of the gas reactant in the liquid also in the zone of reaction, allowing the use of a high mass of catalyst (significative from a statistical point of view) and the achievement of sufficiently high conversion.

The tested reaction is the catalysed hydrogenation of ∝-metylstyrene: in order to consider a drastic situation and to verify the results with the literature data, the experimental conditions examined corresponded to very high chemical reaction rate (instantaneous reaction) at the surface of the pellets.

The tests were carried out with the reactor working both in batchwise and in continuous operative mode (steady state); the results show the reliability of the new reactor above all when the steady state operation is considered. For the use of the reactor in batchwise condition, the accumulation of the product inside the catalyst particles must be considered for an accurate measurement of the kinetic parameters,