Test Data Research Articles

Atterberg limits determination and soil classification using Fall cone device on the silty sands and sandy silts

ABSTRACT The use of the Fall cone device to determine the Atterberg limits of soil mixtures with kaolinite and quartz sand is summarised in this research. The existing Fall cone testing procedures were evaluated and presented. The redefined soil classification on the new plasticity chart was implemented and evaluated in this research. For mixtures with higher fines content (sandy silts), liquid and plastic limit values obtained with the Fall cone methods correlate well with the existing data. For mixtures with lower fines content (silty sands), the Fall cone liquid and plastic limit determination methods were performed with inconsistent results. The test data obtained in the present research suggests that the interparticle contacts in densified soil mixtures with low fines content significantly influence the Fall cone methods’ test results. The existing correlation coefficients for the Fall cone flow indexdetermination could be applied to mixtures with higher fines content and liquid limit >30%.

Predicting saturated hydraulic conductivity from particle size distributions using machine learning

Abstract Estimating saturated hydraulic conductivity $$K_f$$ from particle size distributions (PSD) is very common with empirical formulas, while the use of machine learning for that purpose is not yet widely established. We evaluate the predictive power of six machine learning algorithms, including tree-based, regression-based and network-based methods in estimating $$K_f$$ from the PSD solely. We use a dataset of 4600 samples from the shallow Dutch subsurface for training and testing. The extensive dataset provides not only PSD, but also measured conductivities from permeameter tests. Besides training and testing on the entire data set, we apply the six algorithms to data subsets for the soil types sand, silt and clay. We further test different feature/target-variable combinations such as reducing the input to PSD-derived grain diameters $$d_{10}$$ , $$d_{50}$$ and $$d_{60}$$ or estimating porosity from PSD. We test feature importance and compare results to $$K_f$$ estimates from a selection of empirical formulas. We find that all algorithm can estimate $$K_f$$ from PSD at high accuracy (up to $$R^2/NSE$$ of 0.89 for testing data and 0.98 for the entire data set) and outperform empirical formulas. Particularly, tree-based algorithms are well suited and robust. Reducing information in the feature variables to grain diameters works well for predicting $$K_f$$ of sandy samples, but is less robust for silt and clay rich samples. $$d_{10}$$ also shows to be the most influential feature here. An interesting, but not surprising outcome is that PSD is not a suitable predictor for porosity. Overall, our results confirm that machine learning algorithms are a powerful tool for determining $$K_f$$ from PSD. This is promising for applications to e.g. deep-drilling data sets or low-effort and robust $$K_f$$ -estimation of single samples.

Artificial Intelligence (AI) & BOTS in Testing

The knowledge gathered using Artificial Intelligence (AI), coupled with the efficiency of BOTs is a boon for testing activity in the IT world. This can, not only bring down the human errors, but can also significantly speed up the whole testing process and by extension the project timeline. As part of test preparation, the AI needs be exposed to a range of different paths the product can take, so that it can learn operations and the BOTs designed to each of the possible variations in input and outcome. With detailed data gathering and careful design, AI and BOTs can effectively handle every scenario and ensure the end product meets the needs and goals of the organization. Once the BOTS are in place, they can be reused at will to actively test any enhancements or modifications that the product might have to undergo in its lifetime on the IT world. Similarly, with AI constantly observing the application, it can be used to dynamically identify the differences, both in source and the expected result for building targeted testing document and execution. Keywords : AI, BOTs, Automated Testing, IT testing, Machine Learning, Regression testing. Application Evolution monitoring. And learning, Automated Test data gathering, Programmatic test execution

Innovative Admixture Technology Facilitates Rapid Repair of Concrete Pavements

Recent developments in admixture technology can dramatically modify and control both the rheological and hardened properties of concrete. Using the right combination of specific admixtures and appropriate mixture proportions are the first steps towards producing concrete with the desired performance attributes. For example, a system of innovative admixtures can now produce a novel concrete mix that facilitates the rapid repair of concrete pavements that permits opening to traffic within just 4 hours after concrete placement. This paper provides a description of the concrete system and its components, includes test data, and contains a review of successful projects, and demonstrations for local and national transportation agencies.

Experimental study on pressure relief blasting in rock-blast tunnel

Rock burst is a difficult and urgent problem during the construction process of underground engineering in a zone with high surrounding rock pressure. Pressure relief blasting is one of the most effective methods for solving the problem. An in-situ blasting test was carried out in a tunnel under excavation, and the geology stress (geo-stress) of the surrounding rock was tested before and after the pressure relief blasting. The experimental results will be effective for further study of the law of pressure relief blasting and rock-burst prediction. An explicit stress analytic function is also employed to analyze the stress distribution of the tunnel surrounding rock after pressure relief blasting based on the propagation of the blasting stress wave. Furthermore, the stress attenuation coefficient and the variation of the geo-stress distribution of surrounding rock under the pressure relief blasting were discussed in detail with the in-situ blasting test data. Highlights. The main contribution of this paper is focused on: 1. This study measures the in-situ geo-stress of surrounding rock in the ‘Sang zhu ling’ tunnel, both before and after the pressure relief blasting. This data is crucial for ensuring the safety resilience of urban infrastructure. 2. A two-dimensional explicit equation of stress attenuation was derived based on the characteristics of stress propagation, providing a more accurate analytical tool for engineers and researchers, 3. The traditionally obtained stress attenuation coefficient is shown to require modification through further in-situ experiments. Future studies will refine this coefficient, contributing to the development of more reliable methods for safeguarding urban lifelines. These contributions align with the journal’s focus on advancing urban science and engineering by promoting innovative, data-driven solutions for urban safety.

COMPARATIVE ANALYSIS TO PREDICT READING LITERACY BASED ON PISA 2022 USING GRADIENT BOOSTED DECISION TREES AND EXTREME GRADIENT BOOSTING

Reading is a fundamental skill essential for interdisciplinary understanding and serves as a crucial indicator of the educational quality of a nation. PISA provides an international evaluation of students' reading literacy across various countries, including Indonesia. Numerous studies have utilized machine learning algorithms to predict reading literacy; however, achieving high model accuracy remains a significant challenge. This study compares the performance of Gradient Boosting Decision Trees (GBDT) and Extreme Gradient Boosting (XGBoost), two widely recognized machine learning algorithms for predicting reading literacy, utilizing PISA 2022 data from 12.853 Indonesian students and 59 variables from the Student Questionnaire Data File. The GridSearchCV optimization method was employed to select the optimal parameters for each model to ensure the best performance on the test data. GBDT achieved R² of 0.5106, with optimal parameters (n_estimators = 150, learning_rate = 0.2, max_depth = 3, subsample = 0.9). XGBoost reached a higher R² of 0.5247, with parameters (n_estimators = 1000, learning_rate = 0.01, max_depth = 7, colsample_bytree = 0.3, min_child_weight = 20, gamma = 1, alpha = 0), indicating XGBoost's superior performance in predicting reading literacy. Further analysis revealed that the most significant variables in the GBDT model included students' access to technology at home, extracurricular creative activities, socioeconomic status, school involvement in sustainable development, and problem-solving skills. In contrast, significant variables in the XGBoost model included family support, socioeconomic status, school belongingness, family environment's effectiveness in fostering creativity, and student imagination. These findings provide valuable insights for policymakers in education to understand the factors contributing to students' reading literacy and to design more targeted and effective interventions

Forecasting Agricultural Trade Based on TCN-LightGBM Models: A Data-Driven Decision

Facing the increasing complexity and dynamic fluctuations of the global agricultural trade market, accurate forecasting plays a key role in supporting agricultural policy formulation, stabilising the market and optimising resource allocation. In order to increase the precision and stability of agricultural trade predictions, this research suggests a hybrid model built on a temporal convolution network (TCN) and a lightweight gradient boosting tree (LightGBM). The TCN module effectively captures the long-term dependence characteristics of time series data through dilated convolution, which improves the model’s ability to identify seasonal and periodic trends. The LightGBM module, on the other hand, makes use of the characteristics of gradient boosting decision trees and excels at efficiently handling nonlinear relationships and avoiding overfitting. Experimental results show that the TCN-LightGBM model outperforms traditional models in terms of mean square error (MSE), mean absolute error (MAE) and prediction accuracy. Specifically, compared with ARIMA, LSTM, TCN alone or LightGBM alone, TCN-LightGBM achieves a prediction accuracy of 91.3% on the test data, with MSE and MAE of 0.021 and 0.115 respectively, significantly improving prediction accuracy and stability. In addition, parameter sensitivity analysis shows that the TCN-LightGBM model maintains a highly consistent prediction trend under different parameter configurations, which verifies the robustness of the model and its practical application value. This study provides a data-driven decision support tool with high accuracy and strong stability, providing a new solution for agricultural trade forecasting and other complex time series prediction tasks.

Comparison of the Performance of Nonlinear Time-Dependent Constitutive Models Calibrated with Minimal Test Data Applied to an Epoxy Resin

Epoxy resins are extensively employed as adhesives and matrices in fibre-reinforced composites. As polymers, they possess a viscoelastic nature and are prone to creep and stress relaxation even at room temperature. This phenomenon is also responsible for time-dependent failure or creep fracture due to cumulative strain. Several constitutive equations have been used to describe the mechanical time-dependent response of polymers. These models have been proposed over the past six decades, with minimal direct and practical confrontation. Each model is associated with a specific application or research group. This work assesses the predictive performance of four distinct time-dependent constitutive models based on experimental data. The models were deemed sufficiently straightforward to be readily integrated into practical engineering analyses. A range of loading cases, encompassing constant strain rate, creep, and relaxation tests, were conducted on a commercial epoxy resin. Model parameter calibration was conducted with a minimum data set. The extrapolative predictive capacity of the models was evaluated for creep loading by extending the tests to five decades. The selected rheological models comprise two viscoelastic models based on Volterra-type integrals, as originally proposed by Schapery and Rabotnov; one viscoplastic model, as originally proposed by Norton and Bailey; and the Burger model, in which two springs and two dashpots are combined in a serial and parallel configuration. The number of model parameters does not correlate positively to superior performance, even if it is high. Overall, the models exhibited satisfactory predictive performance, displaying similar outcomes with some relevant differences during the unloading phases.

Metrologically Interpretable Soft-Sensing Technique for Non-Invasive Liquid Flow Estimation from Vibration Data

This paper proposes a metrologically interpretable soft sensing method for estimating the liquid flow rates in hydraulic systems from non-invasive vibration frequency power band data. Despite considerable interest in non-invasive flow estimation, state-of-the-art methods provide little to no metrological capabilities. In this work, a dedicated test rig was developed to automatically acquire vibration and flow rate data from a centrifugal pump, in a flow rate range between 0.05 × 10−5m3/s and 9.11 × 10−5m3/s. The vibration data were processed into power bands, which were subsequently used to optimize and train a multilayer perceptron neural network for flow soft sensing. The trained model was compared with models with different vibration processing methods from literature. The power band processing model resulted in a root mean squared error 75.4% smaller than the second-best model in cross-validation, and 51.5% smaller with test data. The uncertainty of the proposed regression model was estimated using a combination of ensemble learning and Monte Carlo simulations, and combined with the reference flow sensor uncertainty to obtain the total combined uncertainty of the soft sensor, found to be between 3.9 × 10−6m3/s and 6.1 × 10−6m3/s throughout the measured flow range. The reference flow sensor accuracy was found to be the largest individual contribution for the final uncertainty, closely followed by the regression model uncertainty.

Impact of thermal crosstalk on dependent failure rates of multilayer ceramic capacitors undergoing lifetime testing

Several research studies have investigated the degradation of BaTiO3-based dielectric capacitor materials, focusing on the impact of composition, defect chemistry, and microstructural design to limit the electromigration of oxygen vacancies under electric fields at finite temperatures. Electromigration can be a dominant mechanism that controls failure rates in the individual multilayer ceramic capacitor (MLCC) components in testing the reliability of failures with highly accelerated lifetime testing (HALT) to determine the mean time to failure of MLCCs surface mounted onto printed circuit boards (PCBs). Conventional assumptions often consider these failures as independent, with no interaction between components on the PCB. However, this study employs a Physics of Failure (PoF) approach to closely examine transient degradation and its impact on MLCC reliability, emphasizing thermal crosstalk and its influence on dependent and independent failure rates. Finite element analysis thermal modeling and infrared thermography were used to assess the impact of circuit layout and component spacing on heat dissipation and thermal crosstalk under various electrical stress conditions. The study distinguishes between dependent and independent failures under a HALT, quantified through a β′ factor reflecting common cause failures due to thermal crosstalk. Through a series of experimental and statistical analyses, the β′ factor is evaluated with respect to temperature, voltage, and component spacing. These insights highlight the importance of understanding the nature of the data in reliability testing of MLCCs and optimizing the layout design of high-density circuits to mitigate dependent failures, improving overall reliability and informing better design and packaging strategies.

Enhancing Localization Accuracy and Reducing Processing Time in Indoor Positioning Systems: A Comparative Analysis of AI Models

This paper presents a comparative study of different AI models for indoor positioning systems, emphasizing improvements in localization accuracy and processing time. This study examines Artificial Neural Networks (ANNs), Long Short-Term Memory (LSTM), Recurrent Neural Networks (RNNs), and the Kalman filter using a real Received Signal Strength Indicator (RSSI) and 9-axis ICM-20948 sensor. An in-depth analysis is provided in this paper for data cleaning and feature selection to reduce errors for all the models. We evaluate these models in terms of localization accuracy and prediction time. The RNN model shows the best performance, achieving a localization error of 0.247 m with a delay of 0.077 s per position location for an area of 12m×9.5m using four anchors. This research highlights the importance of selecting AI models for effective mobile tracking according to test and validation data.

Penilaian Postur Kerja dengan Metode REBA pada UD. Bintang Terang Roti Ketawa

This research aims to determine the level of complaints felt by workers at UD. Bintang Terang Roti Ketawa, because there is the use of manual tools and a mismatch between body dimensions and the facilities used by the workers, seen from the results of the body map questionnaire that was distributed, there are many complaints felt by the workers, therefore the researchers identified body posture problems using the REBA method, from the assessment of the results using the REBA method obtained very high results or it was said that repairs were carried out immediately. Repairs or design of work facilities were carried out using anthropometric data of workers with normality test calculations, data adequacy, and uniformity so that the results obtained were in accordance with the dimensions of the workers' bodies, the results of facility design. The work succeeded in reducing the level of complaints felt by workers so that the design was suitable for use. Keywords: Anthropometrics, REBA, UMKM, Work Facilities

Student physical education teaching achievement test incorporating biomechanics analysis via decision tree algorithm

In the context of China’s economic growth, building a “sports power” is crucial. With the Internet’s development, the nation focuses on physical education curriculum and students’ test results. Currently, Chinese students’ physical education achievements are mainly evaluated by storage, inquiry, and basic statistics. Digital mining is under - utilized, leaving much data unexplored. But digital education’s progress has brought educational data to the fore, enabling schools to find value and patterns for better teaching decisions. In this case, this paper delves into the integration of biomechanics analysis into the assessment of student physical education teaching achievements. Biomechanics offers invaluable insights into the mechanical aspects of human movement during physical activities. By examining biomechanical factors such as joint angles, muscle forces, and movement velocities during various sports, a more profound understanding of students’ physical capabilities and performance can be achieved. Taking the physical test results of students in a specific school as a case study, this research employs data mining technology to explore the rich dataset. The optimized decision tree algorithm is then utilized to analyze the biomechanics - related data. This algorithm enables the identification of key factors that influence students’ physical test outcomes. For instance, it can pinpoint how a student’s running gait mechanics, including stride length, frequency, and leg - muscle activation patterns, affect their performance in track - and - field events. Through this comprehensive approach, the paper not only aims to uncover hidden information within the physical test data but also to elucidate the intricate interplay between biomechanics and students’ overall physical education performance. By dissecting the factors that influence physical test results from a biomechanical perspective, this study provides actionable insights for enhancing physical education teaching methods. Educators can leverage these findings to design personalized teaching programs that cater to the unique biomechanical characteristics of each student, thereby optimizing the effectiveness of physical education and promoting students’ physical development and athletic performance.

Machine Learning Quantum Mechanical/Molecular Mechanical Potentials: Evaluating Transferability in Dihydrofolate Reductase-Catalyzed Reactions.

Integrating machine learning potentials (MLPs) with quantum mechanical/molecular mechanical (QM/MM) free energy simulations has emerged as a powerful approach for studying enzymatic catalysis. However, its practical application has been hindered by the time-consuming process of generating the necessary training, validation, and test data for MLP models through QM/MM simulations. Furthermore, the entire process needs to be repeated for each specific enzyme system and reaction. To overcome this bottleneck, it is required that trained MLPs exhibit transferability across different enzyme environments and reacting species, thereby eliminating the need for retraining with each new enzyme variant. In this study, we explore this potential by evaluating the transferability of a pretrained ΔMLP model across different enzyme mutations within the MM environment using the QM/MM-based ML architecture developed by Pan, X. J. Chem. Theory Comput. 2021, 17(9), 5745-5758. The study includes scenarios such as single point substitutions, a homologous enzyme from different species, and even a transition to an aqueous environment, where the last two systems have MM environment that is substantially different from that used in MLP training. The results show that the ΔMLP model effectively captures and predicts the effects of enzyme mutations on electrostatic interactions, producing reliable free energy profiles of enzyme-catalyzed reactions without the need for retraining. The study also identified notable limitations in transferability, particularly when transitioning from enzyme to water-rich MM environments. Overall, this study demonstrates the robustness of the Pan et al.'s QM/MM-based ML architecture for application to diverse enzyme systems, as well as the need for further research and the development of more sophisticated MLP models and training methods.

Research on the confinement mechanism and calculation method of axial load-bearing capacity of steel-reinforced concrete-filled square steel tubular columns

There are significant discrepancies between the experimental results and numerical analysis outcomes for steel-reinforced concrete-filled steel tubular (SRCFST) columns. This is primarily due to the lack of understanding of the confinement mechanisms of SRCFST and the corresponding models for confined concrete. Therefore, based on the confinement mechanisms of square steel tubes and steel reinforcements on concrete, this paper calculates the effective lateral confining stress to determine the strength enhancement factors for concrete in various confined regions. Drawing from the classical Mander model, the key parameters are modified to establish stress–strain models for concrete in different confined regions. By comparing and analyzing axial compression test data for steel tube-reinforced concrete short columns with different parameters, the study identifies the most suitable axial compression bearing capacity calculation formula for these columns. Additionally, the introduction of stress–strain models into finite element analysis, compared with experimental results, demonstrates that the proposed confinement mechanism and the established stress–strain model for steel-reinforced concrete are reasonable and effective. The theoretical calculation results for the ultimate bearing capacity based on this are well aligned with the simulation and experimental values.

Understanding tinnitus symptom dynamics and clinical improvement through intensive longitudinal data

Intensive longitudinal sampling enhances subjective data collection by capturing real-time, dynamic inputs in natural settings, complementing traditional methods. This study evaluates the feasibility of using daily self-reported app data to assess clinical improvement among tinnitus patients undergoing treatment. App data from a multi-center randomized clinical trial were analysed using time-series feature extraction and nested cross-validated ordinal regression with elastic net regulation to predict clinical improvement based on the Clinical Global Impression—Improvement scale (CGI-I). With 50% app compliance (N = 129, 8480 entries), the model demonstrated good fit to the test data (McFadden R2 = 0.82) suggesting its generalizability. Clinical improvement was associated with linear declines in tinnitus-related thoughts, jaw tension, tinnitus loudness, increases in happiness, and variability changes in tinnitus loudness and distress. These findings suggest that daily self-reported data on tinnitus symptoms is sensitive to treatment response and provides insights into specific symptom changes that occur during treatment.

Remaining Useful Life Prediction of Lithium-ion Battery based on AConvST-LSTM-Net-TL

Abstract The accurate prediction of the Remaining Useful Life (RUL) of lithium-ion batteries can significantly enhance the safety and reliability of these batteries, thereby reducing operational risks. However, numerous existing methodologies operate under the assumption that both training and testing data adhere to the same distribution pattern, hindering the application of successful laboratory-based models to different target batteries. Hence, this study introduces a transfer learning model for predicting lithium-ion battery life, utilizing an attention mechanism-driven convolutional neural network (ACNN) in conjunction with a spatiotemporal Long Short-Term Memory network (ST-LSTM). Initially, the deep charging process data is leveraged for battery capacity estimation, where a Generalized Additive Model (GAM) is employed to delineate the nonlinear trend of battery capacity decay and characterize the capacity degradation. Following feature extraction, the Maximum Mean Discrepancy (MMD) value is computed to assess the distribution disparity between the two domains. Subsequently, the AConvST-LSTM-Net computes capacity estimations, loss functions for both source and target domains, merging these with the MMD value to formulate a comprehensive loss function. In conclusion, by employing transfer learning, the refined model is adapted to the target domain dataset, enabling the accurate prediction of the RUL of lithium-ion batteries. This approach is validated using the NASA lithium-ion battery dataset and the National Big Data Alliance for New Energy Vehicles (NDANEV), demonstrating the superior accuracy and robustness of the AConvST-LSTM-Net-TL model in comparisn to other prediction methods.

Assessing the Very High Cycle Fatigue Behavior and Frequency Effect of Structural Steel Welds

ABSTRACTThe very high cycle fatigue behavior of S275J2+N flux‐core arc welded joints was investigated using the ultrasonic fatigue testing method at a loading frequency of 20 kHz. A bespoke specimen design featuring the weld toe was employed to more suitably represent in‐service welded joints than typical ultrasonic fatigue testing specimens. This revealed that fracture occurs above 10 million cycles, beyond the classically accepted fatigue limit. Additionally, comparative fatigue tests were performed at 10 Hz to investigate the frequency effect on the fatigue behavior. A 35% increase in fatigue strength was measured at 20 kHz when compared to 10 Hz, indicating that a correction must be considered in the use of ultrasonic fatigue testing data for design purposes. Fracture surfaces for both test frequencies showed similar morphologies and typical characteristics for fatigue failures of ductile metals.

The Data Heterogeneity Issue Regarding COVID-19 Lung Imaging in Federated Learning: An Experimental Study

Federated learning (FL) has emerged as a transformative framework for collaborative learning, offering robust model training across institutions while ensuring data privacy. In the context of making a COVID-19 diagnosis using lung imaging, FL enables institutions to collaboratively train a global model without sharing sensitive patient data. A central manager aggregates local model updates to compute global updates, ensuring secure and effective integration. The global model’s generalization capability is evaluated using centralized testing data before dissemination to participating nodes, where local assessments facilitate personalized adaptations tailored to diverse datasets. Addressing data heterogeneity, a critical challenge in medical imaging, is essential for improving both global performance and local personalization in FL systems. This study emphasizes the importance of recognizing real-world data variability before proposing solutions to tackle non-independent and non-identically distributed (non-IID) data. We investigate the impact of data heterogeneity on FL performance in COVID-19 lung imaging across seven distinct heterogeneity settings. By comprehensively evaluating models using generalization and personalization metrics, we highlight challenges and opportunities for optimizing FL frameworks. The findings provide valuable insights that can guide future research toward achieving a balance between global generalization and local adaptation, ultimately enhancing diagnostic accuracy and patient outcomes in COVID-19 lung imaging.

THE COMPARISON OF LONG SHORT-TERM MEMORY AND BIDIRECTIONAL LONG SHORT-TERM MEMORY FOR FORECASTING COAL PRICE

Coal remains vital for global energy despite recent demand fluctuations due to the COVID-19 pandemic and geopolitical tensions. The International Energy Agency (IEA) projected a decline in global coal demand starting in early 2024, driven by increasing renewable energy adoption. As one of the top coal exporters, Indonesia must adjust to these changes. This study aims to forecast future coal prices using historical data from Indonesia's Ministry of Energy and Mineral Resources (KESDM), applying and comparing Long Short-Term Memory (LSTM) and Bidirectional LSTM (BiLSTM) models. While BiLSTM has shown advantages in other contexts and studies, its effectiveness for coal price forecasting remains underexplored. To ensure robust predictions, we employ walk-forward validation, which divides the data into six segments and evaluates 90 hyperparameter combinations across all segments. The BiLSTM model consistently outperforms the LSTM model, achieving lower average RMSE and MAPE values. Specifically, BiLSTM records an average MAPE of 7.847 and RMSE of 10.485, compared to LSTM's 10.442 and 11.993, respectively. The Diebold-Mariano (DM) test using squared error and absolute error loss functions further corroborates these findings, with most segments showing significant improvements in favor of BiLSTM, indicated by negative DM-test statistics and p-values below 0.01 or 0.10. This superior performance continues into the testing data, where BiLSTM maintains lower error metrics and a significant result of the DM test, underscoring its reliability for forecasting. In the final stage, the forecasts from both models indicate a nearly linear downward trend in coal prices over the next 18 months, aligning with the International Energy Agency's 2023 projection of a structural decline in coal demand driven by the sustained growth of clean energy technologies.