Stages Of Degradation Research Articles

Turnover of non-polymeric leaf lipids in a loamy grassland soil

BackgroundLeaves constitute a large input of lipids to soil, yet comparatively little is known about the fate of leaf lipids in soil. Our aim was to explore the initial stages of degradation of leaf lipids, both the loss of intact lipid and subsequent mineralisation. We focussed on intracellular lipids – triacylglycerols implicated in storage, membrane lipids such as phospholipids and galactolipids, and pigments – because they collectively constitute more than 1% of leaf mass.MethodsA mixture of U-13C lipids was extracted from leaves of wheat grown with 13CO2. The lipid mixture included the range of plant lipids soluble in organic solvent (e.g. free fatty acids, acylglycerols, pigments) but not polymeric lipids such as cutin and suberin. Mineralisation was deduced from 13CO2 efflux, while LC–MS examined degradation of intact 13C lipids.ResultsThere was no delay before lipids were mineralised. Instead, mineralisation was significant within minutes and reached a maximum within three hours. There was rapid loss (i.e. degradation) of a broad range of intact lipids including phospholipids, galactolipids, pigments (chlorophylls), and triacylglycerols. Around two-thirds of added lipid-C was respired over the course of 15 days, with one-third of lipid-C persisting in soil.ConclusionsOur study indicates that non-polymeric leaf lipids degrade quickly in soil, yet a fraction of lipid-C likely persisted in degradation products and/or microbial biomass. Persistence of lipid-C probably also reflected the presence of lipids that are more resistant to degradation (e.g. phaeophytins), and a fraction of added lipid being protected (e.g. by interaction with clays).Graphical abstract

Effects of organically-modified montmorillonite and alkalinization on physical, mechanical, chemical, morphological, and thermal properties of wheat straw/recycled polypropylene nanocomposites

Valorization of wheat straw fiber/recycled polypropylene nanocomposites was investigated through studying the individual and combined effects of organically modified montmorillonite (OMMT) and alkali treatments on the given composites. FTIR spectra indicated the reduction of wheat straw fibers' hydroxyl groups due to hemicellulose removal resulting from alkali treatment. There were also trace diminutions in C-O peak intensities due to the degradation of little amounts of lignin. X-ray analysis revealed the intercalation of nanoclay in the polymer matrix. The improved interaction of fiber-polymer interfaces brought about by alkali treatment was also confirmed by scanning electron microscopy. The water absorption and thickness swelling properties of the nanocomposites were improved due to the barrier properties of nanoclay against water ingress. The flexural strength and modulus of elasticity increased by the individual, and likewise combined treatments. Thermogravimetric analysis demonstrated that the temperatures of first and second stages of composites' thermal degradation increased via the formation of a carbonized char layer thermally insulating the deeper composite layers. Differential scanning calorimetry showed some slight increases in the melting temperature, melting enthalpy, crystallisation temperature, crystallisation enthalpy, and crystallinity index of the treated composites due to the nucleating effect of nanofillers. The overall results showed that the combined treatment of OMMT and NaOH (vs. the individual ones) could significantly improve the overall properties of the studied composites. This was due to some interesting synergistic effects of the given treatments converting the wheat straw fiber/recycled polypropylene nanocomposites to high performance materials of choice for industry.

Roles of HNO x and Carboxylic Acids in the Thermal Stability of Nitroplasticizer.

In the thermal aging of nitroplasticizer (NP), the produced nitrous acid (HONO) can decompose into reactive nitro-oxide species and nitric acid (HNO3). These volatile species are prone to cause cascaded deterioration of NP and give rise to various acidic constituents. To gain insight on the early stage of NP degradation, an adequate method for measuring changes in the concentrations of HONO, HNO3, and related acidic species is imperative. The typical assessment of acidity in nonaqueous solutions (i.e., acid number) cannot differentiate acidic species and thus presents difficulty in the measurement of HONO and HNO3 at a micromolar concentration level. Using liquid-liquid extraction and ion chromatography (IC), we developed a fast and unambiguous analytical method to accurately determine the concentration of HONO, HNO3, acetic/formic acids, and oxalic acid in aged NP samples. Given by the overlay analysis results of liquid chromatography coupled with quadrupole time-of-flight mass spectrometry and IC, the prominent increase of produced HONO after the depletion of antioxidants is the primary cause of HNO3 formation in the late stage of NP degradation, which results in the acid-catalyzed hydrolysis of NP into 2,2-dinitropropanol and acetic/formic acids. Our study has demonstrated that the aging temperature plays a crucial role in accelerating the formation and decomposition of HONO, which consequently increases the acidity of aged NP samples and hence accelerates the hydrolyzation of NP. Therefore, to prevent NP from undergoing rapid degradation, we suggest that the concentration of HNO3 should be maintained below 1.35 mM and the temperature under 38 °C.

Magnesium Ortho-Vanadate/Magnesium Oxide/Graphene Oxide Embedded through Cellulose Acetate-Based Films for Wound Healing Applications.

A multifunctional nano-films of cellulose acetate (CA)/magnesium ortho-vanadate (MOV)/magnesium oxide/graphene oxide wound coverage was fabricated. Through fabrication, different weights of the previously mentioned ingredients were selected to receive a certain morphological appearance. The composition was confirmed by XRD, FTIR, and EDX techniques. SEM micrograph of Mg3(VO4)2/MgO/GO@CA film depicted that there was a porous surface with flattened rounded MgO grains with an average size of 0.31 µm was observed. Regarding wettability, the binary composition of Mg3(VO4)2@CA occupied the lowest contact angle of 30.15 ± 0.8o, while pure CA represents the highest one at 47.35 ± 0.4°. The cell viability % amongst the usage of 4.9 µg/mL of Mg3(VO4)2/MgO/GO@CA is 95.77 ± 3.2%, while 2.4 µg/mL showed 101.54 ± 2.9%. The higher concentration of 5000 µg/mL exhibited a viability of 19.23%. According to optical results, the refractive index jumped from 1.73 for CA to 1.81 for Mg3(VO4)2/MgO/GO@CA film. The thermogravimetric analysis showed three main stages of degradation. The initial temperature started from room temperature to 289 °C with a weight loss of 13%. On the other hand, the second stage started from the final temperature of the first stage and end at 375 °C with a weight loss of 52%. Finally, the last stage was from 375 to 472 °C with 19% weight loss. The obtained results, such as high hydrophilic behavior, high cell viability, surface roughness, and porosity due to the addition of nanoparticles to the CA membrane, all played a significant role in enhancing the biocompatibility and biological activity of the CA membrane. The enhancements in the CA membrane suggest that it can be utilized in drug delivery and wound healing applications.

Study on high moisture extruded pea protein isolate based on acid-induced process: Physicochemical properties, conformational changes and fibrous structure mechanism

The present study aims to gain a better understanding of the physicochemical properties and quality attribute changes, and fibrous structure formation mechanisms of pea protein analogs with different glucan-δ-lactone (GDL) additions under high moisture extrusion. The results showed that the pH value of the extrudates gradually approached the isoelectric point and the zeta potential increased with the increase in GDL, which reduced the water holding capacity, nitrogen solubility index and browning index values, while increasing the hardness, chewiness and fibrous degree of the extrudates. The comparison revealed the maximum zeta potential value and total SH, and minimum free SH of the raw material, suggesting the unfolding and cross-linking of proteins under the triple action of the extruder. The relative content of β-sheet as the predominant component of protein extrudate decreased, while α-helix increased, and λmax was bathochromiclly shifted through FTIR and fluorescence spectroscopy. However, negative trends were observed when the GDL concentration reached 1%, due to the excessive concentration of the GDL accelerating the aggregation and sedimentation of proteins, and intensified the phase separation degree in the cooling die during the extrusion process, thus promoting the change in the velocity gradient. In addition, the thermal stability of the extrudate first increased and then decreased as the GDL concentration increased, and showed a three-step thermal degradation stage. The pores and highly irregular fibrous filaments transformed into a highly compact and coarse matrix with an excessive GDL concentration, as observed by SEM. These findings validate the potential of GDL to improve the textural attributes and structure formation of meat analogs, and provide a reference for the development of acid-induced proteins under high moisture extrusion.

E-RULENet: remaining useful life estimation with end-to-end learning from long run-to-failure data

This paper presents the e-RULENet, which is a novel framework to train a data-driven model for remaining useful life estimation from long run-to-failure data with an end-to-end manner. In order to enable end-to-end learning, a change point from the healthy stage to the degradation stage is estimated for each instance from measurements. The change point estimation in the previous framework is computationally expensive since it utilizes entire measurements over the lifetime for each instance. The e-RULENet solves this issue with estimating the change points with segments sampled from run-to-failure data. It is evaluated on three datasets for RUL estimation including two real-world datasets, and on a dataset for tension estimation as another use case. It outperforms an approach without end-to-end learning on all the three datasets. The results indicates that the e-RULENet works not only for cases with multiple health stages but also for a single degradation stage, which does not have any changes points.

A piecewise method for bearing remaining useful life estimation using temporal convolutional networks

Most existing research on the prediction of bearing remaining useful life (RUL) focus on building a whole-lifecycle degradation model, which may affect prediction performance due to different states during the bearing degradation process. In addition, traditional network models such as CNN and RNN have limitations in directly dealing with time series problems. This paper proposes an adaptive degradation stage division strategy and a temporal convolutional network (TCN)-based RUL piecewise estimation method, called ADSD-TCNPE. This method mainly includes three steps. (1) Extract features from different domains and select the features that are highly correlated with bearing degradation. (2) Adaptively divide the whole lifecycle of bearing into different degradation stages. (3) Establish a TCN-based piecewise degradation model for different degradation stages to accurately predict bearing RUL. Finally, experimental verification and result analysis demonstrate the effectiveness and superiority of the proposed method.

Data-driven battery health prognosis with partial-discharge information

The unpredictability of battery degradation behavior is a challenging issue impeding the development of battery applications, due to the complexity of the degradation and the limitation of state measurement methods. Nowadays, with accessible battery aging datasets and machine learning algorithms, there are opportunities for data-driven battery health prognosis. However, most of the previous work is restricted in the scope of full-discharge capacity records extrapolation, which has insufficient prospects in real-life applications. In this work, we propose using partial discharge information for degradation estimation and prediction. Our Gaussian process regression model achieves good performance by limited partial discharge information without requirements of feature selection. The accurate battery health prognosis in 300 cycles can be carried out by one partial-discharge cycle at any degradation stage. The capacity estimation gives around 1 % root mean square error (RMSE) when using 30 % information on the discharge process. As full-cycle discharge is not required, the proposed model can diagnose the battery state of health (SOH) with a limited portion of battery operation information extracted during the discharge process and reduce the effort of capacity tests. Further development of this method brings opportunities for battery state evaluation and prediction in real applications with better applicability and accuracy.

Potential‐free sodium‐induced degradation of silicon heterojunction solar cells

AbstractSilicon‐based photovoltaic (PV) modules suffer from potential‐induced degradation (PID) caused by sodium (Na) permeation, which is present in large quantities in soda‐lime glass. Here, we report that Na atoms can decrease the performance of amorphous/crystalline silicon heterojunction (SHJ) solar cells without the help of a voltage bias. The three degradation stages are investigated in this work. First, H2O molecules open channels for Na transport in the transparent conductive oxide (TCO), while the device performance remains almost unchanged. Next, when Na atoms reach the boron‐doped hydrogenated amorphous silicon (p‐a‐Si:H), the field passivation is poisoned, leading to a great decline in the fill factor (FF), whereas the open‐circuit voltage (Voc) only slightly declines. Finally, Na atoms further diffuse into the intrinsic a‐Si:H layer and c‐Si surface, resulting in a substantial decrease in Voc. These findings have important implications for the installation of SHJ solar modules in Na‐abundant environments. As a feasible solution, we demonstrate that a compact SiO2 thin film can efficiently prevent H2O molecules from penetrating into the TCO layer and therefore guarantee a long‐term stable operation of SHJ solar cells.

The Importance of Soil Elevation and Hydroperiods in Salt Marsh Vegetation Zonation: A Case Study of Ria de Aveiro

Salt marshes are among the most valuable ecosystems on earth; however, previous research has reported global losses in their extent accompanied by an impairment of their ecological processes. Nonetheless, investigations on salt marsh vegetation ecology are still lacking in many sites, including in Ria de Aveiro. As such, this study aimed to (1) monitor, characterize, and understand the evolution of salt marsh vegetation communities at three different study sites with different environmental conditions and stages of degradation, where we also sought to (2) identify the main delimiting abiotic factors associated with the distribution of the main species. To do so, a multidisciplinary approach that involved the survey of vegetation from permanent transects and the collection of ecological, physicochemical, and hydrodynamic data at sampling points within the monospecific stands of the main species was reported. The results showed that, of the abiotic factors deemed as the main delimiting forces of salt marsh vegetation, the soil elevation and hydroperiod were the most restrictive factors, as they ultimately influenced species composition at the different study sites and explained most of the variation observed between the studied monospecific stands.

Modeling left-truncated degradation data using random drift-diffusion Wiener processes

ABSTRACT For products whose performance characteristic (PC) gradually degrades with time, one usually observes its degradation levels repeatedly to predict its remaining useful life (RUL). Due to the limited storage space of the server and the low resolution of a measurement instrument, we seldom record the low-magnitude degradation values at the early degradation stage in applications. Such observation setting introduces left-truncated degradation data, in which the data collection starts later than the unit’s installation. This brings sampling biases and complicates the degradation data analysis. Moreover, due to the uncontrollable factors in applications, the degradation drift and the degradation diffusion may differ among various units. Motivated by an application of high-speed train bearings, we propose a Wiener process model for the left-truncated degradation data and jointly consider the drift-diffusion random effects. Closed-form formulas are available in the expectation-maximization (EM) algorithm for estimating the model parameters. We derive the RUL distribution in closed form. We also extend the proposed model to the multivariate degradation process. The parameters are estimated with the help of the Monte Carlo EM (MCEM) algorithm. An additional laser application illustrates the performance of the proposed model in RUL prediction, which may help to design a predictive maintenance strategy

Implicit Kalman filtering method for remaining useful life prediction of rolling bearing with adaptive detection of degradation stage transition point

Remaining useful life (RUL) prediction is a vital task in rolling bearing prognostics and health management (PHM) process. Kalman filtering (KF) is one of the hot spots in the research area of RUL prediction. However, three dispiriting shortcomings in KF methods are still unavoidable, including: (1) difficulty in tracking the unknown time-varying noise information, (2) the subjectivity for setting time to start prediction (TSP), and (3) short-term accuracy of the predicting results based on linear predictors. To improve the capability of KF methods, this work adopts the variational Bayesian technique to adaptively describe noise information and considers linear and nonlinear factors of multi-channel signals to recognize the degradation stage transition point of bearing as TSP. Moreover, this work proposes an implicit Kalman filtering method to predict the RUL. The effectiveness of the proposed method is validated on XJTU-SY and IMS-Rexnord bearing data. Results show that the proposed method can recognize the TSP and improve the long-term accuracy of the prediction result during the accelerated degradation stage.

A hybrid method for cutting tool RUL prediction based on CNN and multistage Wiener process using small sample data

Based on the multistage and nonlinear characteristics of cutting tool wear, a hybrid method for cutting tool remaining useful life (RUL) prediction based on convolutional neural network (CNN) and multistage Wiener process using small sample data is proposed in this paper. First, wavelet transform is used to analyse the vibration data collected for cutting tools in the time–frequency domain to obtain an initial image data set. A CCVAE (conditional variational autoencoder with CNN) network is built to expand the initial data set and solve the problem of unbalanced data in different stages of cutting tool wear. The expanded data is used as the input of the CNN to monitor the tool wear state and amount of wear. Then, a nonlinear multistage Wiener process is established to describe the cutting tool wear degradation process and achieve accurate RUL prediction for the cutting tool. Specifically, a nonlinear Wiener process model corresponding to different degradation stages based on the change points between the cutting tool wear states output by the CNN is established. Maximum likelihood estimation and Bayesian methods are used to estimate and update the parameters, respectively, and the RUL value and corresponding probability density function (PDF) are obtained under different wear conditions. Finally, through experimental research and comparative analysis, it is found that the multistage nonlinear Wiener model accurately simulates cutting tool wear degradation, which verifies the feasibility and performance of the method proposed in this paper.

Long-Term Degradation Assessment of a Polyurethane-Based Surgical Adhesive—Assessment and Critical Consideration of Preclinical In Vitro and In Vivo Testing

Tissue adhesives constitute a great possibility to improve conventional wound closure. In contrast to sutures, they enable nearly immediate hemostasis and can prevent fluid or air leaks. In the present study, a poly(ester)urethane-based adhesive was investigated which already proved to be suitable for different indications, such as reinforcing vascular anastomosis and sealing liver tissue. Using in vitro and in vivo setups, the degradation of the adhesives was monitored over a period of up to 2 years, to evaluate long-term biocompatibility and determine degradation kinetics. For the first time, the complete degradation of the adhesive was documented. In subcutaneous locations, tissue residues were found after 12 months and in intramuscular locations, tissue degradation was complete after about 6 months. A detailed histological evaluation of the local tissue reaction revealed good biocompatibility throughout the different degradation stages. After full degradation, complete remodeling to physiological tissue was observed at the implant locations. In addition, this study critically discusses common issues related to the assessment of biomaterial degradation kinetics in the context of medical device certification. This work highlighted the importance and encouraged the implementation of biologically relevant in vitro degradation models to replace animal studies or at least reduce the number of animals in preclinical testing prior to clinical studies. Moreover, the suitability of frequently used implantation studies based on ISO 10993-6 at standard locations was critically discussed, especially in light of the associated lack of reliable predictions for degradation kinetics at the clinically relevant site of implantation.

Use of Drone RGB Imagery to Quantify Indicator Variables of Tropical-Forest-Ecosystem Degradation and Restoration

Recognizing initial degradation levels is essential to planning effective measures to restore tropical forest ecosystems. However, measuring indicators of forest degradation is labour-intensive, time-consuming, and expensive. This study explored the use of canopy-height models and orthophotos, derived from drone-captured RGB images, above sites at various stages of degradation in northern Thailand to quantify variables related to initial degradation levels and subsequent restoration progression. Stocking density (R2 = 0.71) and relative cover of forest canopy (R2 = 0.83), ground vegetation (R2 = 0.71) and exposed soil + rock (R2 = 0.56) correlated highly with the corresponding ground-survey data. However, mean tree height (R2 = 0.31) and above-ground carbon density (R2 = 0.45) were not well correlated. Differences in correlation strength appeared to be site-specific and related to tree size distribution, canopy openness, and soil exposure. We concluded that drone-based quantification of forest-degradation indicator variables is not yet accurate enough to replace conventional ground surveys when planning forest restoration projects. However, the development of better geo-referencing in parallel with AI systems may improve the accuracy and cost-effectiveness of drone-based techniques in the near future.

Pyrolysis and thermal degradation studies of non-isocyanate polyurethanes modified by polyhedral oligomeric silsesquioxanes

This work reports on the thermal properties of non-isocyanate polyurethanes (NIPUs) obtained via the polyaddition reaction of polyether diamine with five membered tri(cyclic carbonate) by using a prepolymerization method. The non-isocyanate polyurethanes were further chemically modified by trifunctional polyhedral oligomeric silsesquioxane (POSS) - triglycidylisobutyl POSS (3epPOSS), resulting in hybrid composite materials containing 5 to 15 wt% POSS. Thermal properties of the obtained materials were studied by various methods, such as Thermogravimetric analysis (TGA), Pyrolysis with gas chromatography and mass spectrometry (Py-GC/MS), and Microscale Combustion Calorimetry (MCC), and were later discussed in dependence of their composition and structure assessed by Infrared Spectroscopy (FTIR), X-ray Diffraction (XRD), and Scanning Electron Microscopy (SEM) techniques. FTIR spectra confirm the successful synthesis of NIPUs and their hybrids with POSS. The results of XRD and SEM revealed homogeneous distribution of silsesquioxane in the bulk of the NIPU matrix. Depending on the POSS concentration in the PU hybrids, various self-assembling structures were formed. TGA analysis indicated a two- and three-step degradation process of NIPU materials in inert and oxidative atmospheres, respectively. Significantly higher degradation temperatures and a remarkably reduced degradation rate for composite materials compared to unmodified NIPU matrix were observed during the second degradation step under an oxidative atmosphere. Based on the overall stabilization effect (OSE), hybrids containing 10 and 15 wt% of 3epPOSS were more thermally stable than the unmodified NIPU matrix. Py-GC/MS demonstrated that the products of NIPU thermal decomposition originate from hydroxyurethane moieties, main substrate core structures, and others, such as catalysts. Incorporation of 3epPOSS into the NIPU matrix decreases its flammability, as evidenced by the results of the MCC, with a nearly 20 % reduction effect noted for the composite with of 15 wt% of POSS, compared to the unmodified NIPU matrix. Importantly, chemical modification of NIPU by POSS caused a reduction of the heat of combustion in the final stage of degradation by ca. 30 %.

Bayesian inference of mesoscale mechanical properties of mortar using experimental data from a double shear test

In this work, we propose Bayesian parameter estimation of a nonlinear mechanics based model describing the behaviour of mortar subjected to double shear test with externally bonded carbon fibre reinforced polymer (CFRP) plates. With the Bayesian approach, it is possible to identify mechanical material parameters of different phases of the mortar mesostructure, i.e. hardened cement paste, aggregates and interface transition zone (ITZ). Due to nonlinearity of the concerned problem, we use a novel sequential approach for the parameter inference, which does not require coupling between the finite element solver and software for the stochastic analysis. The model geometry and material mesostructure are learned based on micro-computed tomography (μCT) scans of the real specimen, whereas the unknown boundary conditions are assumed to be uncertain and are also identified from experimental data. Mortar is modelled through a discrete lattice model consisting of spatial Timoshenko beams with embedded discontinuities. The latter allows the description of the distinct stages in material degradation, from the appearance of microscopic material damage to its evolution into macroscopic cracks leading to localised failure.

Resilience-Oriented Co-Deployment of Remote- Controlled Switches and Soft Open Points in Distribution Networks

Soft open points (SOPs) can perform load transferring and voltage support, helping to expand the range of service restoration in distribution networks (DNs) under fault scenarios. In addition, in the field of DN restoration, SOPs are usually controlled in coordination with switch actions (e.g., the network reconfiguration). However, in the planning level, the coordinated deployment of SOPs and remote-controlled switches (RCSs) has never been studied in the literature. To this end, for the first time, this paper proposes a co-deployment framework for SOPs and RCSs incorporating active management of flexible resources for resilience improvement in DNs, in which the coupled multiple recovery stages are also modeled and included. The objective is to minimize the investment cost of SOPs and RCSs, along with the cost caused by de-energized loads, with the consideration of exhaustive DN operation constraints over proactive islanding, degradation, isolation and restoration stages. Furthermore, resilience-cost tradeoff is analyzed to help DN managers determine equipment investment plans. The formulated model is further transformed into a mixed-integer second-order cone programming model that can be thus efficiently solved. Simulation results on modified IEEE 34-node and IEEE 123-node test systems verify the effectiveness of the proposed model.

Insights into the moisture‐induced degradation mechanisms on field‐deployed CIGS modules

AbstractMoisture ingress is known to be one of the most frequent and detrimental causes of photovoltaic (PV) modules degradation. In an original approach, the present work applies the full analytic capability of lab‐scale equipment to characterize a commercial Cu(In,Ga)(Se,S)2 (CIGS) module that degraded in the field due to moisture ingress. By extracting (coring) and unpackaging samples from different areas of the module, we are able to track the propagation of the degradation mechanisms with respect to both material and electrical properties. The three following locations in the device were found to be affected by the water ingress: first, the encapsulant, ethylene‐vinyl acetate (EVA), resulting in delamination and acetic acid formation; second, the molybdenum (Mo) back contact, which oxidizes and corrodes entirely at the P3 scribes; lastly, the ZnO front contact, through the formation of Zn‐based hydroxides and its further reaction with the Mo migrating from the degraded scribes. The main impact on the electrical properties is an increase in series resistance and a decrease in both open‐circuit voltage and short‐circuit current. In the last degradation stage, that is, the most degraded area of the module, a complete performance loss is observed. The results of this work further our understanding of the physics and chemistry involved in the moisture‐induced degradation in CIGS modules. In particular, the study highlights the interactions between the various degradation products and associated degradation mechanisms. These results also illustrate the broader potential of the coring and unpackaging method for reliability studies of full‐sized, field‐degraded commercial PV modules.

Deep ensemble learning enables highly accurate classification of stored red blood cell morphology

Changes in red blood cell (RBC) morphology distribution have emerged as a quantitative biomarker for the degradation of RBC functional properties during hypothermic storage. Previously published automated methods for classifying the morphology of stored RBCs often had insufficient accuracy and relied on proprietary code and datasets, making them difficult to use in many research and clinical applications. Here we describe the development and validation of a highly accurate open-source RBC morphology classification pipeline based on ensemble deep learning (DL). The DL-enabled pipeline utilized adaptive thresholding or semantic segmentation for RBC identification, a deep ensemble of four convolutional neural networks (CNNs) to classify RBC morphology, and Kalman filtering with Hungarian assignment for tracking changes in the morphology of individual RBCs over time. The ensembled CNNs were trained and evaluated on thousands of individual RBCs from two open-access datasets previously collected to quantify the morphological heterogeneity and washing-induced shape recovery of stored RBCs. Confusion matrices and reliability diagrams demonstrated under-confidence of the constituent models and an accuracy of about 98% for the deep ensemble. Such a high accuracy allowed the CNN ensemble to uncover new insights over our previously published studies. Re-analysis of the datasets yielded much more accurate distributions of the effective diameters of stored RBCs at each stage of morphological degradation (discocyte: 7.821 ± 0.429 µm, echinocyte 1: 7.800 ± 0.581 µm, echinocyte 2: 7.304 ± 0.567 µm, echinocyte 3: 6.433 ± 0.490 µm, sphero-echinocyte: 5.963 ± 0.348 µm, spherocyte: 5.904 ± 0.292 µm, stomatocyte: 7.080 ± 0.522 µm). The effective diameter distributions were significantly different across all morphologies, with considerable effect sizes for non-neighboring classes. A combination of morphology classification with cell tracking enabled the discovery of a relatively rare and previously overlooked shape recovery of some sphero-echinocytes to early-stage echinocytes after washing with 1% human serum albumin solution. Finally, the datasets and code have been made freely available online to enable replication, further improvement, and adaptation of our work for other applications.