Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Pome fruit, as well as all other fruit and vegetables, are a source of vital food constituents such as proteins, vitamins, polysaccharides, phenolics and minerals. In 2018, the trade of apple and pear at the auctions of the Association of Belgian Horticultural Cooperatives (VBT) amounted to 100 000 and 180 000 tons, respectively, corresponding to a collective turnover of 130 million EUR.Since fruit start to deteriorate after harvest due to respiration and associated metabolic processes, they are commonly stored under conditions of low temperature, decreased O2 partial pressure and slightly increased CO2 partial pressure (controlled atmosphere storage, CA) to optimally maintain their quality. When the O2 partial pressure during storage becomes too low, however, the fermentation pathway becomes dominant which leads to off-flavors and the development of storage disorders. For this reason, O2 and CO2 levels during CA are kept at safe and steady, but suboptimal setpoints that may be above or below the critical levels. The application of CA storage, however, causes an important challenge towards fruit batches from different origins and seasons. Respiration rate, fruit size and tissue structure may be different between fruit and batches, affecting internal gas gradients. In general, more ripe, larger and more dense fruit are more sensitive to low O2 partial pressures compared to less ripe, smaller and porous fruit. Computer models are available to study the effect of these parameters on the physiology of fruit stored under CA and to optimize the storage protocol. These models assume material properties such as gas diffusivities that are homogenous throughout the fruit. However, the microstructure of fruit tissue is quite heterogeneous and this may very well affect gas transport considerably. So far, however, no model is available that takes into account the complete heterogeneity of tissue microstructure. The objective of this dissertation was, therefore, to quantify the heterogeneity of apple tissue and to incorporate it into existing models for gas transport in apple.Existing models to describe gas transport in pome fruit use a single phase formulation. This formulation uses an effective diffusivity, either obtained from experiments or from a microscale simulation of gas transfer through cells and pores in small tissue samples. This single phase formulation in essence also assumes equilibrium between pores and cells everywhere in the apple. Still, the rate of gas transport is significantly different in these two phases and, therefore, a two-phase multicomponent multiscale model may be more appropriate, solving gas transport in the pores and cells with separate equations, and interphase transport between them. In a first step towards achieving such a model, a two-phase formulation was proposed in Chapter 3. The two-phase formulation contains two effective diffusivity values per gas species, one for pores and one for cells, and an additional interphase transfer term that is a function of microstructural parameters obtained from micro-CT analysis. It was assumed that the effective diffusivity values, like in the single phase model, can be obtained from 3D microscale simulations on the tissue geometries of the micro-CT scans. This model in principle allows also non-equilibrium conditions between the two phases inside apples. Previously developed models for similar applications showed that, besides the interphase transfer term, additional terms occur in the two-phase model that describe diffusion phenomena at the interface of pores and cells. Based on what was suggested in literature, these additional terms were lumped into the effective diffusivities of the separate phases. The proposed two-phase multiscale model was evaluated for steady-state conditions using in silico experiments and was found sufficiently accurate for the purposes of this dissertation. O2 diffusion in 'Jonagold' apples stored in CA conditions was evaluated with the two-phase model and results showed that equilibrium conditions were satisfied. Afterwards, the sensitivity of the two-phase model was checked towards: (i), open porosity in the cortex samples; (ii), tissue respiration; and (iii), interphase resistance. The former two had significant effects on the O2 distribution inside the apple. Increasing the interphase resistance significantly, on the other hand, had a small effect on the average O2 distribution, but resulted in non-equilibrium O2 concentrations between the pore and cell phase.In Chapter 4, X-ray CT scans of intact 'Braeburn' apples with unprecedented resolution were made to visualize the internal microstructure of an entire apple fruit. A high microstructural variability was observed, both between different apples as well as within a single fruit. To optimize controlled atmosphere storage conditions, the effect of this heterogeneity on transport of metabolic gasses (O2, CO2) needed to be clarified. 'Braeburn' apples, who are highly susceptible to internal browning during storage, were characterized in terms of porosity distribution throughout the whole apple. These scans were used to identify different tissue compartments inside the apples. Based on the 3D connectivity of the pores, the cortex tissue was divided into a region with high porosity (HPC) and low porosity (LPC). The HPC had a porosity of 30.4 ± 2.1 % and featured relatively larger pores compared to the LPC, which had a porosity of 13.2 ± 3.3 %. On the internal boundary of the HPC and LPC, around a relative radius of 0.4, the porosity reached minimal values. A barrier to gas transport was identified at this position, where the exocarp and the main vascular tissue of the apple are situated. Furthermore, results showed that in two out of four tested apples the ovary was connected to the environment due to incomplete growth.Chapter 5 incorporated these compartments in the developed two-phase multiscale model. For each compartment, effective diffusivities were estimated and microstructural parameters were calculated. The relationship between gas diffusivity and microstructural parameters was studied and results showed that the gas diffusivity of a tissue sample did not always scale with the porosity of said sample, but rather with the ratio of the open porosity over the tortuosity. Macroscale simulations in commercial CA storage conditions showed that the O2 concentration profiles were highly variable between and within 'Braeburn' apples. Internal microstructure, and, most importantly, the open porosity, seems to vary between the apples in order to provide sufficient O2 throughout the apple. Furthermore, the results of the macroscale simulations with multiple compartments also showed that minimal O2 concentrations are not necessarily reached in the center of the 'Braeburn' apple, which can potentially have a relationship with the position of internal brown development in the cortex tissue.To further include structural heterogeneity in the modelling of gas transport, an alternative modeling approach to the two phase model was required. Hereto, a network modelling approach was studied in Chapter 6. The network model translated the pore and cell phase of the apple into individual cells and pores, represented as nodes, which made up a nodal network of the apple tissue. Compared to the multicompartment multiscale model, more detailed results towards O2 concentration profiles were found: a larger concentration gradient was found in the radial direction of 'Braeburn' apples when using the network model. Network modelling, therefore, provided a good and computationally efficient alternative to multiscale modelling to further investigate the transport of gasses in heterogeneous porous fruit such as apples.Concerning future prospects, a relatively new approach for storing fruit based on a dynamically controlled atmosphere (DCA) has triggered significant interest from the horticultural sector in Flanders and beyond. Instead of a constant setpoint for O2, the lowest O2 setpoint below which fermentation occurs is continuously searched based on the stress response of the fruit. With DCA, the occurrence of disorders and quality losses are further minimized compared to CA. The dynamic nature of the method, however, means that gas concentrations are time-dependent and can be continuously changing. In this respect, the models presented in this thesis will need to be elaborated and evaluated for transient conditions. Future work should focus on a comparison of the proposed two-phase model with the more elaborate two-phase model formulations in literature, and experimental results obtained with needle oxygen sensors or gas scattering spectroscopy measurements.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Tomatoes are an important source of food and income for the people of Ethiopia. It is a cash-generation crop for small-scale farmers as well as for employers in the production and processing industry. To be able to store tomatoes properly and for a longer time, good postharvest technology and infrastructure are needed. In developing countries like Ethiopia, good techniques and knowledge often lack. The ripening process of tomatoes or its decay can be delayed by storing the tomatoes at lower temperatures. A cold storage room is needed. Conventional refrigerators or storage rooms make use of electricity. In Ethiopia, 85 percent of all farms are located in rural areas where no electricity is available. A valuable solution to tackle this problem is a cooling system powered by solar energy, more specifically a solar powered diffusion absorption cooling system. A diffusion absorption refrigerator (DAR) is a cooling technology that can be driven purely by thermal energy without the need of any mechanical pumps, making them an interesting cooling system in off-grid communities. Therefore, an optimal design of such a system is needed. Optimisation must be performed based on models instead of trial-and-error. In this thesis, it is attempted to make a model for the evaluation of a diffusion absorption refrigerator. A commercial diffusion absorption system was used for experiments. Temperature measurements were performed on this apparatus, so that the measured temperatures could be used as an input in the model. Further, heat flux sensors were used to determine the cooling capacity and efficiency of the refrigerator. The measured cooling capacity for three different power inputs was approximately constant, in a range of 10.3-11.3 W. The cooling capacity measured by the heat fluxes was used for validation. In this thesis, two steady state models were constructed. The first model is based on already existing models and temperatures are used as an input. With this model the heat flows are analyzed. Based on the heat flows, 4 new equations were introduced in the model, so that in a second steady state model only the ambient temperature, working pressure, power supplied to and temperature of the generator could be used as an input. The set of thermodynamic equations was solved using an optimization tool. No unambiguous solution was obtained, and errors present in the first model were mirrored in the second model and might even be enlarged. The measured results were compared with the model predictions. In the first model the highest relative difference obtained was 21%, while in the second model it was 72%. A sequential analysis was performed, which revealed that the equations for the partial pressure and the VII heat losses must be revised. Further, the sequential analysis was solved using a nonlinear solver, which showed promising results to work with in a future research.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Worldwide there is an increasing demand for high quality fruit year-round. To make this possible fruit must be stored for a long time. Unfortunately, internal disorders, such as browning and cavities, can develop under suboptimal storage conditions and often show no external symptoms. During sorting, it is important that pears affected by these internal disorders are removed, because consumers largely base their repurchase decisions of a product on the internal quality. Nowadays, the few inspection methods available for the internal quality of pear fruit are only able to classify pears as healthy or defect. This leads to losses that should be avoided, e.g. some defect pears may have only minor disorders and can thus still be consumable. Therefore, this thesis aimed to go a step further by developing a method to automatically quantify the severity of internal disorders in pear fruit. Hereto, a combination of X-ray computed tomography (CT) and deep learning was used to create a deep neural network that is able to segment regions affected by internal disorders in X-ray CT scans of pear fruit. Additionally, a consumer study was done to gain information about the consumer acceptance of defects in fruit. Hereby, fruit was classified into three categories: “healthy”, “defect but consumable” and “defect and non-consumable”. In this thesis, it was found that the severity of the internal disorders could be accurately predicted from the deep learning based segmentations of CT scans. Moreover, in a first classifier it was shown that fruit could be classified as “consumable” and “non-consumable” with a high accuracy and low false positive and false negative rates of 99.4%, 0.0% and 3.8%, respectively. Pears with a different internal quality could thus be distinguished which provides useful information for decisions on pricing and the discarding of fruit. Using a second classifier, pears could be classified as “healthy”, “defect but consumable” and “defect and non-consumable”. The classification of “defect but consumable” pears was shown to be the most difficult. The developed system has thus several advantages. First, it is financially beneficial for growers since for top label fruit it can be guaranteed that no internal disorders are present. Fruit classified as “defect but consumable” can be marketed at a lower price. Second, it is beneficial for consumers since low false negative rates were obtained. Third, it leads to less food losses and food waste in the supply chain. Lastly, the developed method can facilitate future research into the origin and development of internal disorders in fruit. By combining the quantitative information about the severity of the internal disorders with the results of a consumer study, the sorting process of pear fruit can be optimized by creation of a sorting system based on the severity of a disorder and its acceptance, and by that reducing waste. However, before the system can be used in commercial sorting lines, further developments are needed in both software and hardware. In particular, it must be possible to implement X-ray CT inline. Finally, the generalizability of the classifiers should be tested for other pear cultivars and the system can be extended to other types of fruit, such as apple.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Infectious diseases and sexually transmitted diseases in particular represent a major burden on public health in developing countries, where they have far-reaching social, demographic and economic consequences. Globally, more than 1 million sexually transmitted infections are acquired every day. Early detection of infectious diseases is therefore an important part of prevention of disease outbreaks. Detection has traditionally been conducted in centralised laboratories, where complex tests are more reliable. However, this practice has disadvantages in developing regions of the world. Laboratory services are often limited, unaffordable and geographically unavailable. Also, there is a large amount of time between the test and its results, thus reducing the chance of a timely diagnosis. An emerging approach to this problem is the development of Point-of-Care (POC) tests, which are all diagnostic tests performed near the patient. These devices handle small liquid volume samples from a patient and transport them via small channels towards analysis zones, where the markers for a disease can be detected. Since these POC devices are used near patients, they should be cheap, simple to use, portable and importantly rapid to give results. To create such tests, 3D printing of porous devices can be used. The 3D printer is programmed to assign which parts of the device repel water, and which parts do not. This way, channels can be created. The fluids are transported by capillarity, which is the same mechanism as water uptake in paper. Furthermore, devices fabricated in this way do not require power or electricity to have water be pumped through the channels. Previously, 3D printed devices were created via trial-and-error, which is time-consuming, wasteful and costly. In this thesis, a computer model was developed. The model allows for prediction of how the fluids will be transported in the channels and how dissolved solutes will be mixed. With the model, operations useful for POC tests were simulated. These included improved mixing designs, controlled dilutions and concentration gradients. Ultimately with this modelling approach, future devices for integrated diagnostics can be designed. Furthermore, the model enables design improvement to make devices smaller and thus require less time to give results.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Recently, cabbages enjoy a sharp increase in popularity thanks to their high amount of vitamins, minerals and glucosinolates. Among these in-demand cabbages are white, oxheart cabbages, a small and sweeter subgroup. When compared to other cabbages, these oxheart cabbages are very prone to quality decay during storage. Determining the optimal storage conditions is primordial in order to offer these small cabbages year- round to the consumers. Controlled Atmosphere (CA) and Dynamic Controlled Atmosphere (DCA) are two innovative storage techniques that adapt the gas-mixture inside of the storage spaces. By reducing the O2 concentration and increasing the CO2 concentration the metabolism of the stored products is slowed down. This research assessed if these storage techniques are suited for the storage of oxheart cabbages. A first, large storage experiment was conducted in order to compare traditional low temperature storage with both CA and DCA storage. CA and DCA yielded a better quality preservation. However, this improvement was limited to the outer appearance of the cabbages. Furthermore, trends of the quality parameters indicated that long- term DCA storage would probably produce better results than CA storage. This was further endorsed by respiration measurements that showed a clear time-effect after six months. In addition, it was also evaluated if aroma could be used as an indicator for the physiology of the stored cabbage. This was researched during short jar-experiments in which the relation between quality, respiration and aroma was studied. This showed that aroma is a more sensitive indicator than respiration. Thereby, seven aroma- compounds were singled out that can function as an alarm signal that warns when cabbage-quality is decreasing. Ultimately, a respiration model was obtained. This indicated that cabbages can be stored at very low O2 concentrations.In addition, it was also evaluated if aroma could be used as an indicator for the physiology of the stored cabbage. This was researched during short jar-experiments in which the relation between quality, respiration and aroma was studied. This showed that aroma is a more sensitive indicator than respiration. Thereby, seven aroma- compounds were singled out that can function as an alarm signal that warns when cabbage-quality is decreasing. Ultimately, a respiration model was obtained. This indicated that cabbages can be stored at very low O2 concentrations.