Changes In Process Parameters Research Articles

Numerical Simulation and Experimental Study the Effects of Process Parameters on Filament Morphology and Mechanical Properties of FDM 3D Printed PLA/GNPs Nanocomposite.

The selection of optimal process parameters has a decisive effect on the quality of 3D printing. In this work, the numerical and experimental methods were employed to investigate the FDM printing deposition process of PLA/GNPs nanocomposite. The effect of process parameters on cross-sectional morphology and dimension of the deposited filament, as well as the mechanical property of the FDM printed specimens were studied. The extrusion and the deposition process of the molten PLA/GNPs nanocomposite was simulated as a fluid flow by the paradigm of CFD, the effects of printing temperature and shear rate on thermal-physical properties, such as viscosity and surface tension, were considered in models. Under the assumptions of non-Newtonian fluid and creep laminar flow, the deposition flow was controlled by two key parameters: the nozzle temperature and the nozzle velocity. The numerical model was verified by experiments from four aspects of thickness, width, area, and compactness of the deposited PLA/GNPs nanocomposite filament cross-section. Both the numerical simulation and experiment results show that with the increase of nozzle temperature and nozzle velocity, the thickness, area, and compactness of the deposited filament decreases. While the width of deposited filament increased with the increase of nozzle temperature and decrease of nozzle velocity. The decrease in thickness and the increase in width caused by the change of process parameters reached 10.5% and 24.7%, respectively. The tensile strength of the printed PLA/GNPs specimen was about 61.8 MPa under the higher nozzle temperatures and velocity condition, an improvement of 18.6% compared to specimen with the tensile strength of 52.1 MPa under the lower nozzle temperatures and velocity condition. In addition, the experimental results indicated that under the low nozzle velocity and nozzle temperature condition, dimensional standard deviation of the printed specimens decreased by 52.2%, 62.7%, and 68.3% in X, Y, and Z direction, respectively.

Improving Tapping Experience—and Why to Choose a Comprehensive Approach

Tapping is an essential task in almost all furnaces in the metal-producing industry. The layout of the tap-hole and tapping process needs to be optimized for the individual process and furnace to ensure safe and efficient furnace operations. Various factors need to be considered for tap-hole and tapping design, in addition to the actual tapping procedure. Generally, tap-hole and tapping optimization is not a one-time task, but an ongoing improvement cycle, comprising evaluation and adaptation of geometries, materials, data and technologies, considering changing process parameters and use of new technologies. The present article describes methods and technologies to optimize tap-hole performance, tapping safety and efficiency.

Investigation of Process Parameters for Optimum Thermal Energy Distribution in Spark Erosion Process

In the Spark Erosion machining process, the electrical energy in the inter-electrode gap gets converted into thermal energy. Whatever thermal energy generated in inter electrode gap gets distributed among anode, cathode and the dielectric fluid. The workpiece in spark erosion process is kept as anode and the fraction of the total energy which is going to the anode (workpiece) should be maximum. As this thermal energy is responsible for the melting of material on the workpiece. This fraction of energy is the important parameter used to predict the effective use of total heat energy generated between tool and workpiece. The fraction of energy that is going to workpiece changes with change in machine parameter. It has been observed from the literature that the researcher simply considered a variation of spark voltage in analysis as a significant parameter in deciding material removal rate. In this work is variation of fraction of energy is experimentally determined with process parameters by using heat transfer equations during the spark erosion process of EN-31. The results showed that the ratio of amount of energy going to workpiece and total energy generated at inter electrode gap ranges from 9.56% to 18.97%. So, fraction of energy which is going to workpiece at different machining process parameter changes significantly, which proves that the it has significant impact on material removal rate and should be considered in study of spark erosion process.

The use of the Cpm index to monitor processes

AbstractThe Cpm index was originally created to measure the ability of the process to produce products meeting specifications but, more recently, the Cpm has also been used to control processes. In this new application, the charting points of the control chart are the estimated values of the Cpm index obtained with the mean and the variance of the samples. The performance of the Cpm chart is not affect by the values of the specification limits, but it is highly affected by d, the distance between the middle point of the specifications and the in‐control mean of the X distribution. Because of that, a search is conducted to find d that optimizes the overall performance of the Cpm chart in signaling the three usual types of changes in process parameters: isolated mean shifts, isolated variance increases, and the two together. Comparing the joint and S2 charts, the Cpk chart and the Cpm chart, the last one is always the most sensitive in detecting all three types of process changes, except when the variance remains unaltered and the mean increases/decreases a lot (one and half standard deviations); in this case, the joint and S2 charts are slightly faster than the Cpm chart.

Adaptive Practical Nonlinear Model Predictive Control for Echo State Network Models.

This article proposes an adaptive practical nonlinear model predictive (NMPC) control algorithm which uses an echo state network (ESN) estimated online as a process model. In the proposed control algorithm, the ESN readout parameters are estimated online using a recursive least-squares method that considers an adaptive directional forgetting factor. The ESN model is used to obtain online a nonlinear prediction of the system free response, and a linearized version of the neural model is obtained at each sampling time to get a local approximation of the system step response, which is used to build the dynamic matrix of the system. The proposed controller was evaluated in a benchmark conical tank level control problem, and the results were compared with three baseline controllers. The proposed approach achieved similar results as the ones obtained by its nonadaptive baseline version in a scenario with the process operating with the nominal parameters, and outperformed all baseline algorithms in a scenario with process parameter changes. Additionally, the computational time required by the proposed algorithm was one-tenth of that required by the baseline NMPC, which shows that the proposed algorithm is suitable to implement state-of-the-art adaptive NMPC in a computationally affordable manner.

Evaluation of robotic fiber placement effect on process‐induced residual stresses using incremental hole‐drilling method

AbstractThe robotic fiber placement (RFP) process depends on continuous fusion bonding and solidifying of prepreg tows onto a substrate by subjecting them to a compression force and heat flux. However the advantages provided when using the RFP to fabricate the composite structures, it has adverse effects due to the induced residual stresses (RSs). The current work presents the effect of the RFP on the induced RSs in thermoset composites. RFP is utilized to fabricate specimens from carbon fiber reinforced polymers. Effects of placement speed, compression force, heating temperature, heating gas flowrate, and the location of the heat flux outlet nozzle are all included in the present study. A total of 69 flat panels are manufactured under different process conditions. Subsequently, the in‐depth RSs are measured using the incremental hole‐drilling method. Besides, the microstructures of the samples are also analyzed. The results show that the change in process parameters could significantly affect the induced RSs of the thermoset composites. Subjecting the samples to higher thermal and pressure loads would induce more RSs, which are attributed to the occurrence of partial curing during the manufacturing process along with the change in laminate geometry and porosity.

Performance of the CUSUM Control Chart Using Approximation to ARL for Long- Memory Fractionally Integrated Autoregressive Process with Exogenous Variable

The cumulative sum (CUSUM) control chart is a well-known process monitoring tool that is sensitive to small-to-moderate changes in process parameters. In this paper, we propose an approximated average run length (ARL) method based on the numerical integral equation (NIE) method for monitoring the mean of a long-memory autoregressive fractionally integrated process with an exogenous variable (ARFIX) running on a CUSUM control chart. The approximated ARL based on the NIE method is realized by solving a system of linear equations and integration based on the partitioning and summation of the area under the curve of a function derived by using the Gauss-Legendre quadrature. In a comparative study with the ARL based on analytical formulas, the proposed approximated ARL method could detect shifts of various sizes in the process mean of an ARFIX process running on a CUSUM control chart. In addition, the proposed method was compared with their analytical formulas in terms of the relative percentage change (r%) to verify the accuracy of the ARL results. The results revealed that the ARL results obtained from the NIE method are an approach to analytical formulas with r% of less than 0.25. Hence, the NIE method is very accurate and in excellent agreement with the analytical formulas approach. Apparently, the NIE method is an alternative as efficiently as the analytical method for this situation. It also performed well in comparison with the approximated ARL for the same process running on an exponentially weighted moving average control chart. In addition, real datasets are also used to demonstrate the efficacy of the proposed method.

Effect of processing parameters on the microstructure and mechanical properties of TiAl/Ti2AlNb laminated composites

In order to improve the intrinsic brittleness of TiAl alloys, Ti2AlNb alloys with outstanding ductility and toughness at room temperature, and good high-temperature performance are competitive candidates in constructing the TiAl-based laminated composites. In this work, TiAl/Ti2AlNb laminated composites are successfully synthesized by vacuum hot pressing combined with the foil-foil (sheet) metallurgy. Under the pressure of 65 MPa, different holding time and temperature of hot pressing are tried and the optimized fabrication parameter is acquired as 1050 °C/120 min/65 MPa. Along with the changes of processing parameters, the defect, microstructure, interface, phase transformation and the corresponding mechanical properties are detailly discussed. The results show that the TiAl/Ti2AlNb laminated composite fabricated at 1050 °C for 2 h achieves a good metallurgical interface bonding. The corresponding interface microstructure is composed of region I and region II. The region I consists of O, α2 and B2/β phase, and region II is made up of α2. Subsequently, the tensile tests indicate that the composite synthesized at 1050 °C for 2 h possesses a maximum strength of 812 MPa and a total elongation of 1.31% at room temperature, and a strength of 539.71 MPa and the highest total elongation of 10.34% at 750 °C. The well synergistic deformation ability between the interface and the two base alloys endows the composite an excellent tensile performance. Moreover, the composite processed at 1050 °C for 2 h behaves the best fracture toughness in both arrester orientation and divider orientation with the value of 32.6 MPa.m1/2 and 30.1 MPa.m1/2, respectively. The Ti2AlNb alloy in the laminated structure effectively release the stress around the crack tip and plays a role in toughening. Further, crack deflection, crack bridging, crack blunting and fragmentation also make contributions to enhance the fracture toughness of the laminated composites.

Flow boiling critical heat flux enhancement in ZrSi2 accident-tolerant fuel cladding with porous structures

In this study, critical heat flux (CHF) of Zirconium Silicide (ZrSi2) ATF coating for the zircaloy-4 (Zry-4) substrate has been assessed in flow boiling and enhanced by homogeneous porous surfaces fabricated by electrophoretic deposition (EPD). Three different porous structures were fabricated on physical vapor deposition (PVD) ZrSi2 coating using EPD. The particle size distribution, thickness, and porosity were adjusted by the changing process parameters of EPD to produce different porosity levels. Mean particle size and porosity of the EPD porous layer were increased by 35% and 21% respectively when the applied voltage was increased from 100 to 400 V. The CHF of the bare Zry-4, ZrSi-Zry, and ZrSi-Zrys modified by EPD (EPD ZrSi-Zrys) was evaluated in flow boiling at a mass flux of 300 kg/m2s. The optimized ZrSi2 coating modified by EPD enhanced the CHF of ZrSi-Zrys up to 146% when deposited at 400 V. Flow boiling enhancements of CHF of EPD ZrSi-Zrys were analyzed by evaluating liquid inflow rates of the surfaces.

Melt pool feature analysis using a high-speed coaxial monitoring system for laser powder bed fusion of Ti-6Al-4 V grade 23

In laser powder bed fusion (LPBF), defects such as pores or cracks can seriously affect the final part quality and lifetime. Keyhole porosity, being one type of porosity defects in LPBF, results from excessive energy density which may be due to changes in process parameters (namely, laser power and scan speed) and/or result from the part’s geometry and/or hatching strategies. To study the possible occurrence of keyhole pores, experimental work and simulations were carried out for optimum and high volumetric energy density conditions in Ti-6Al-4 V grade 23. By decreasing the scanning speed from 1000 to 500 mm/s for a fixed laser power of 170 W, keyhole porosities are formed and later observed by X-ray computed tomography. Melt pool images are recorded in real-time during the LPBF process by using a high-speed coaxial Near-Infrared (NIR) camera monitoring system. The recorded images are then pre-processed using a set of image processing steps to generate binary images. From the binary images, geometrical features of the melt pool and features that characterize the spatter particle formation and ejection from the melt pool are calculated. The experimental data clearly show spatter patterns in case of keyhole porosity formation at low scan speed. A correlation between number of pores and amount of spatter is observed. Besides the experimental work, a previously developed high fidelity finite volume numerical model was used to simulate the melt pool dynamics with similar process parameters as used during the experiments. Simulation results illustrate and confirm the keyhole porosity formation by decreasing laser scan speed.

Residual Stresses Measurements in Laser Powder Bed Fusion Using Barkhausen Noise Analysis.

In recent years, the advancement of technology brought the laser powder bed fusion process to its industrialisation step. Despite all the advancements in process repeatability and general quality control, many challenges remain unsolved due to the intrinsic difficulties of the process, notably the residual stresses issue. This work aimed to assess the usability of Barkhausen noise analysis (BNA) for the residual stress in situ monitoring of laser powder bed fusion on Maraging steel 300 (18Ni-300/1.2709). After measuring the evolution of grain size distribution over process parameter changes, two series of experiments were designed. First, a setup with an external force allows to validate the working principle of BNA on the chosen material processed using LPBF. The second experiment uses on-plates samples with different residual stress states. The results show a good stability in microstructure, a prerequisite for BNA. In addition, the external load setup acknowledges that signal variation correlates with the induced stress state. Finally, the on-plate measurement shows a similar signal variation to what has been observed in the literature for residual stress variation. It is shown that BNA is a suitable method for qualitative residual stresses variation monitoring developed during the LPBF process and underlines that BNA is a promising candidate as an in situ measurement method.

Kinematic characteristics of ultrasonic surface treatment

We investigate the kinematic characteristics of ultrasonic surface-plastic deformation when the oscillations are introduced in the direction tangential to the treated surface in order to assess the potential of this method for treating parts made of metals and alloys of various hardness. The kinematic parameters (trajectory, velocity, acceleration) of ultrasonic surface-plastic deformation were calculated in a rectangular coordinate system. Analytical dependencies were obtained for the trajectory length, resulting velocity, and acceleration as functions of the constituent types of motion (rotational, translational, oscillatory). The obtained equations and their graphical solutions allowed us to establish that, under ultrasonic surface-plastic deformation, the indenter moves along a complex trajectory with a variable velocity and acceleration. It is shown that the translation of ultrasonic oscillations to the indenter determines the complex nature of its movement, which differs significantly from that in the diamond burnishing method. In this case, the process parameters (velocity and acceleration) change according to a cyclic (sinusoidal) law, the periodicity of which depends on the frequency of ultrasonic oscillations and determines the final state of the treated surface. According the obtained results, the change in the input direction of ultrasonic oscillations (from normal to tangential) ensures the possibility of changing the contact condition from cyclic to constant, thereby affecting the deforming force during processing. It was found that the angle of oscillation input relative to the main velocity vector is a technological parameter significantly affecting the kinematic characteristics. An assumption is made about the possibility of using the tangential pattern of ultrasonic surfaceplastic deformation for parts made of metals and alloys of various hardness.

Three‐dimensional forming of plastic‐coated fibre‐based materials using a thermoforming process

AbstractThree‐dimensional (3D) forming of fibre‐based materials has been a topic of growing interest over recent years and 3D forming processes using hydroforming, press forming and deep drawing processes have been widely explored. Thermoforming as a potential alternative method for forming these materials remains, however, relatively understudied. This research attempts to provide a fundamental understanding of the thermoforming limitations of plastic‐coated paperboards. In the work, a variety of commercial paperboards are subjected to experimental tests with different forming parameters and moulding depths. Shape accuracy, maximum acquired depth, thickness distribution behaviour and damage mechanisms are used to evaluate thermoformability, and the results linked to the material properties and forming conditions. The research findings indicate that the plastic‐coated paperboards studied are thermoformable but only in simple geometric shapes and with low mould depths. Unlike plastic, thermoforming can result in thickness increase in plastic‐coated paperboards, which is thought to be a result of out‐of‐plane auxetic behaviour of paperboards. Paperboard thermoforming was also found to be hindered by rupture, blistering and curling defects. Tensile strain at break is the key factor determining thermoformability. Additionally, the density of the paperboard can impact the heating step and the rate at which the moisture content of the material changes during the forming process. Furthermore, it was observed that changes in process parameters affected materials differently, with the direction and rate of change differing based on the material being used.

Mechanical properties and corrosion behavior of a friction stir processed magnesium alloy composite AZ31B–SiC

Abstract The optimization of friction stir processing (FSP) parameters of magnesium alloy composite (AZ31B–SiC) based on orthogonal test was researched. The results show that the distribution of silicon carbide (SiC) particles, microhardness, tensile property, and fracture mode are greatly affected by the change in process parameters. The results show that the composite was made with a rotating speed of 750 rev·min−1, a traversing speed of 30 mm·min−1, and a processing time of three; the distribution of SiC particles is even, the microhardness difference of composite is small, the tensile property is better, and the ductile fracture is the main fracture mode. Besides, the existence of SiC and the number of FSP have a certain influence on the corrosion performance of the magnesium alloy composite (AZ31B–SiC), and the corrosion resistance of the FSP sample is obviously better than that of the AZ31B magnesium alloy.

Monitoring of resistance spot welding expulsion based on machine learning

There is still a lack of effective monitoring methods for resistance spot welding expulsion in manufacturing sites. In this study, more than 63,766 dynamic resistance curves of welding points were collected in manufacturing sites, and the machine learning method is applied to monitor the resistance spot welding expulsion. Generally, the model’s generalisation ability is poor due to the complex conditions of the manufacturing site. This study solves this problem by improving the data preprocessing and model selection method. The experimental results show that when the process parameters do not change, the accuracy of expulsion recognition can easily reach 95%. When the process parameters change, it can reach 90% by selecting the appropriate data preprocessing method and model selection.

Online reconstruction and diagnosibility analysis of multiplicative fault models for process-related faults

Multiplicative faults generally refer to the change of process parameters or structures which are well-suited to represent the process-related anomalies. Unlike sensor faults and external disturbances that are added into process observations and independent with process states, process-related faults directly influence process states such that it is more challenge to reconstruct and diagnose. To address the process-related fault diagnosis, an online fault reconstruction method based on the multiplicative fault model is proposed with the commonly used multivariate statistical process monitoring framework. The fault reconstruction strategy based on the multiplicative fault representation is given by minimizing reconstruction errors. The diagnosability of the proposed reconstruction method is guaranteed for the change of a single parameter, also known as a unidimensional fault. Moreover, the reconstruction-based contribution is derived for providing heuristic references when diagnosing multidimensional faults. Experiments on a numerical example and a simulated continuous stirred tank heater process benchmark are carried out to investigate the effectiveness of the proposed method. The results show that this method can accurately diagnose the faulty variable or loop and further reconstruct the faulty samples into normal ones.

Long-Term, Simultaneous Impact of Antimicrobials on the Efficiency of Anaerobic Digestion of Sewage Sludge and Changes in the Microbial Community

The aim of this study was to evaluate the influence of simultaneous, long-term exposure to increasing concentrations of three classes of antimicrobials (β-lactams, fluoroquinolones and nitroimidazoles) on: (1) the efficiency of anaerobic digestion of sewage sludge, (2) qualitative and quantitative changes in microbial consortia that participate in methane fermentation, and (3) fate of antibiotic resistance genes (ARGs). Long-term supplementation of sewage sludge with a combination of metronidazole, amoxicillin and ciprofloxacin applied at different doses did not induce significant changes in process parameters, including the concentrations of volatile fatty acids (VFAs), or the total abundance of ARGs. Exposure to antibiotics significantly decreased methane production and modified microbial composition. The sequencing analysis revealed that the abundance of OTUs characteristic of Archaea was not correlated with the biogas production efficiency. The study also demonstrated that the hydrogen-dependent pathway of methylotrophic methanogenesis could significantly contribute to the stability of anaerobic digestion in the presence of antimicrobials. The greatest changes in microbial biodiversity were noted in substrate samples exposed to the highest dose of the tested antibiotics, relative to control. The widespread use of antimicrobials increases antibiotic concentrations in sewage sludge, which may decrease the efficiency of anaerobic digestion, and contribute to the spread of antibiotic resistance (AR).

Sulfide leaching of high-grade arsenic copper concentrates

Such factors as lower-grade copper ores, more demanding mining and technical conditions, and higher implementation costs for new development projects are typical for the Ural region in general and for the Uchalinsky Mining and Processing Plant (Uchalinsky GOK), developing tennantite-containing copper-sulfide ore reserves, in particular. Depending on the process flow design and the mineral composition of the ore, 10–70 % of arsenic is transferred into copper concentrates. The mass fraction of arsenic in the resulting copper concentrates may reach 2 %, significantly reducing the concentrate grade. Sulfide-alkaline leaching has been studied as a promising conditioning method for flotation copper concentrate. It ensures more effective arsenic removal and compliance of the final product with the requirements for further pyrometallurgical processing. The studies were carried out using a low-grade high-arsenic concentrate sample from the Uchalinsky GOK containing 16.04 wt. % Cu, 5.30 wt. % Zn, 1.36 wt. % As, and 0.19 wt. % S. It has been established that iron is predominantly occurring in the form of pyrite, with chalcopyrite and sphalerite unevenly distributed in its mass. The concentrate had high concentrations of fahlore minerals (tennantite, etc.). Studies were conducted to assess the effects of changes in the sulfide leaching process parameters (temperature, raw material size, NaOH and Na2S concentrations, solid-liquid ratio, duration) on the transfer of arsenic into the alkaline solution. The specific conditions have been established (t = 95 °C; Cm(NaOH) = 3.5 M, Cm(Na2S) = 1.5 M; solidliquid ratio = 5; duration of three hours or more) that ensure almost complete removal of arsenic from the flotation concentrate to obtain a conditioned copper product meeting the requirements of GOST R 52998-2008.

Optimization of resource-intensive dynamic systems with a continuous supply of raw materials according to the criterion of minimum use of reserves

Optimization of production processes has always been one of the cornerstones for industrial enterprises seeking to improve productivity while minimizing the costs involved. A particularly difficult situation is when it is necessary to manage the process of the entire production chain with a continuous supply of raw materials. It is necessary to keep under control the actual production data, current production requirements, and adhere to the international strategy of energy saving. This paper reports a devised optimal dynamic system with a continuous supply of raw materials, which automatically changes the control trajectory in order to reduce the amount of resources used. The theoretical scientific component is represented in the form of an interface model of the system, and the research results are represented in the form of time diagrams that show the verification of the proposed model. The model provides for the interconnection of the chain of such developed dynamical systems, in which the continuity of the process is ensured by buffering systems, and the optimality of operation is enabled by adaptation mechanisms. The time diagrams can demonstrate the interaction of systems and mechanisms that generate information signals through the port sections. At each subsequent control action, the process parameter changes were made within a set range. As a result of a targeted search for permissible controls, the system, driven by the adaptation mechanism, enabled a gradual reduction in the consumption of the energy product and stabilized the intensity of the target product being processed, which made it possible to subsequently avoid shutdowns and restarts of the production line and reduce overall production costs.

JUSTIFICATION OF A RATIONAL METHOD AND EQUIPMENT FOR DRYING HIGH-MOISTURE SEEDS OF MELONS

The main reason holding back the wide distribution of pumpkin, like some other crops, is the lack of quality seeds. The receipt and storage of seed is largely dependent on the methods and means of drying. When choosing a technology and means for drying pumpkin seeds, one must take into account their adhesion properties, that is, the tendency to form conglomerates under certain conditions, sticking to working bodies, etc. Drying equipment should ensure uniform drying throughout the entire volume of products with different moisture content, as well as a change in the main process parameters – temperature and drying agent feed rate, which will allow the use of differential drying modes. But in drying installations used for grain processing, it is practically impossible to organize a high-quality process due to the impossibility of using differentiated modes of drying seeds, as well as increased cohesive-adhesive properties of the pumpkin seed layer. A much greater effect is achieved by devices with active hydrodynamic modes, in particular, in a fluidized bed with its modifications: an aerial boiling, vibroboiling and aerovibroboiling bed. In order to improve the flowability of a layer of materials prone to sticking, sticking, the formation of channels in the material layer, agglomeration, at the initial stage of drying, along with vibration and aeration mixing, it is advisable to use additional mechanical mixing devices. The article proposes a fundamentally new concept of energy-saving drying of high-moisture seed materials, in particular pumpkin seeds, the drying process of which through specific morphological, thermophysical, increased cohesive-adhesive properties, the presence of residues of the surface film and fruit in the seed layer is the most energy-intensive. At the first stage in the process of filtration drying, in which the heated drying agent under the influence of pressure drops moves through the cellular structure of the gas-permeable material, the drying of the material also occurs due to a change in the state of aggregation of the available, mainly surface, moisture. At the second stage, when the seeds acquire discrete properties, drying occurs in the aerovibroboiling layer due to the developed phase contact and an increase in the convective diffusion rate, which contributes to an increase in the drying rate in general and allows the process to be carried out at the maximum permissible values of seed heating. To carry out research to determine the rational technological parameters of the process, a structural and technological scheme of a vibration unit has been developed and a prototype has been created, which sequentially implements the filtration and convective stages of drying pumpkin seeds.