Multiscale Strategy Research Articles

Abstract P1180: Multi-scale Imaging And Analysis To Elucidate Cardiac Structure And Contractile Function

Heart failure remains the leading cause of mortality worldwide, largely due to the lack of regenerative capacity of mature cardiomyocytes. The advanced understanding of cardiac architecture and function is critical to uncover the underlying mechanism of cardiac morphogenesis and the remodeling process in response to myocardial infarction. Recent progress has demonstrated that zebrafish and neonatal mice have the capacity to restore cardiac function following the injury, laying the foundation to investigate the cues to structural and functional abnormalities. For this reason, we have developed a light-sheet fluorescence microscope along with the tailored computational methods to investigate the 3-dimensional (3D) architecture of the intact mouse heart, and 4D (3D spatial + 1D temporal) cardiac contractility in zebrafish larvae. This method provides us ~ 2 um isotropic spatial resolution across the atria and ventricles with minimal photo-damage and maximal penetration depth. In combination with established tissue clearing methods, we are able to investigate myocardial trabeculation and compaction in mouse hearts within the 2-minute laser scanning. To unravel the contractile dysfunction following the anticancer therapy, we have coupled our imaging system with retrospective synchronization algorithm for the 4D reconstruction of myocardial deformation from end-systole to end-diastole in zebrafish hearts. To further interpret the deluge of 3D and 4D images, we have also developed a machine learning-based platform for interactive image analysis. Collectively, our multi-scale imaging strategy provides an entry point to assess micro-architecture and cardiac contractile function in the entire heart without physical slicing, holding the great potential to uncover the process of cardiac morphogenesis and regeneration.

BFFNet: a bidirectional feature fusion network for semantic segmentation of remote sensing objects

PurposeHigh-resolution remote sensing images possess a wealth of semantic information. However, these images often contain objects of different sizes and distributions, which make the semantic segmentation task challenging. In this paper, a bidirectional feature fusion network (BFFNet) is designed to address this challenge, which aims at increasing the accurate recognition of surface objects in order to effectively classify special features.Design/methodology/approachThere are two main crucial elements in BFFNet. Firstly, the mean-weighted module (MWM) is used to obtain the key features in the main network. Secondly, the proposed polarization enhanced branch network performs feature extraction simultaneously with the main network to obtain different feature information. The authors then fuse these two features in both directions while applying a cross-entropy loss function to monitor the network training process. Finally, BFFNet is validated on two publicly available datasets, Potsdam and Vaihingen.FindingsIn this paper, a quantitative analysis method is used to illustrate that the proposed network achieves superior performance of 2–6%, respectively, compared to other mainstream segmentation networks from experimental results on two datasets. Complete ablation experiments are also conducted to demonstrate the effectiveness of the elements in the network. In summary, BFFNet has proven to be effective in achieving accurate identification of small objects and in reducing the effect of shadows on the segmentation process.Originality/valueThe originality of the paper is the proposal of a BFFNet based on multi-scale and multi-attention strategies to improve the ability to accurately segment high-resolution and complex remote sensing images, especially for small objects and shadow-obscured objects.

Explainable Deep Ensemble Model for Bearing Fault Diagnosis Under Variable Conditions

Deep learning-based intelligent diagnostic methods have been widely used in aerospace, rail transportation, automotive, rail vehicles, and other fields. However, deep neural networks are composed of multi-layer non-linear transformation to learn the feature maps of input, leading to the black box characteristic. It is usually criticized to be non-transparent and less interpretable in predictions and maintenance decision-making. Meanwhile, due to the variability of machine working conditions in the manufacturing process, there is still a gap between traditional deep learning diagnosis methods and the practical application. To address such problems, an explainable deep ensemble model is proposed for fault diagnosis of bearings which both leverages feature learning ability and improves model interpretability. Firstly, a residual network integrated with bi-directional long short-term memory (BiLSTM) is designed as the backbone learning framework to perform associative feature extraction from the original 1D time-series signals. Then, a multi-branch model with a multi-scale ensemble learning strategy is constructed for feature fusion and final fault diagnosis. After the model is trained, an enhanced gradient-weighted class activation mapping technique, named 1D Grad-CAM ++ is embedded to visualize the region of interest in the convolutional layer, where the prior knowledge gained from the bearing fault mechanism model is effectively combined to provide a better understanding of the learned features and decision-making of model. Finally, three fault diagnosis cases of rolling bearings under different operation conditions are used for algorithm verification. The results demonstrate the proposed approach shows superior explainable capability and diagnosis performance in comparison with other methods.

Multiscale Stamps for Real-time Classification of Alert Streams

In recent years, automatic classifiers of image cutouts (also called “stamps”) have been shown to be key for fast supernova discovery. The Vera C. Rubin Observatory will distribute about ten million alerts with their respective stamps each night, enabling the discovery of approximately one million supernovae each year. A growing source of confusion for these classifiers is the presence of satellite glints, sequences of point-like sources produced by rotating satellites or debris. The currently planned Rubin stamps will have a size smaller than the typical separation between these point sources. Thus, a larger field-of-view stamp could enable the automatic identification of these sources. However, the distribution of larger stamps would be limited by network bandwidth restrictions. We evaluate the impact of using image stamps of different angular sizes and resolutions for the fast classification of events (active galactic nuclei, asteroids, bogus, satellites, supernovae, and variable stars), using data from the Zwicky Transient Facility. We compare four scenarios: three with the same number of pixels (small field of view with high resolution, large field of view with low resolution, and a multiscale proposal) and a scenario with the full stamp that has a larger field of view and higher resolution. Compared to small field-of-view stamps, our multiscale strategy reduces misclassifications of satellites as asteroids or supernovae, performing on par with high-resolution stamps that are 15 times heavier. We encourage Rubin and its Science Collaborations to consider the benefits of implementing multiscale stamps as a possible update to the alert specification.

Shape-Aware Joint Distribution Alignment for Cross-Domain Image Segmentation.

We present an unsupervised domain adaptation method for image segmentation which aligns high-order statistics, computed for the source and target domains, encoding domain-invariant spatial relationships between segmentation classes. Our method first estimates the joint distribution of predictions for pairs of pixels whose relative position corresponds to a given spatial displacement. Domain adaptation is then achieved by aligning the joint distributions of source and target images, computed for a set of displacements. Two enhancements of this method are proposed. The first one uses an efficient multi-scale strategy that enables capturing long-range relationships in the statistics. The second one extends the joint distribution alignment loss to features in intermediate layers of the network by computing their cross-correlation. We test our method on the task of unpaired multi-modal cardiac segmentation using the Multi-Modality Whole Heart Segmentation Challenge dataset and prostate segmentation task where images from two datasets are taken as data in different domains. Our results show the advantages of our method compared to recent approaches for cross-domain image segmentation. Code is available at https://github.com/WangPing521/Domain_adaptation_shape_prior.

Multi-scale structural investigation of uranium-plutonium mixed oxides (U1-yPuy)O2-x with high plutonium content

Uranium-Plutonium mixed oxides U1-yPuyO2-x are considered as the reference fuels for Sodium-cooled Fast neutron Reactors (SFRs). Nevertheless, there is a lack of experimental data on their thermal properties for a Pu content above 0.45. The first step to fill this gap is to manufacture samples with high plutonium contents, which is the purpose of this work. Here, the manufacturing of U1-yPuyO2 samples with y = 0.60, 0.65 and 0.70 by a co-milling powder metallurgy process is detailed. Relying on an optimization of each manufacturing step (blending, co-milling, sieving, pelletizing and sintering), in order to obtain dense, monophasic, stoichiometric and homogenous materials.A multi-scale characterization strategy was used to determine sample microstructural properties including density measurements, Thermal-Ionization Mass Spectrometer (TIMS), ceramography, Electron Probe Micro Analyses, µ-Raman Spectroscopy and X-Ray Diffraction. The results obtained show that dense (> 95% TD), homogenous (chemical distribution of cations and oxygen), monophasic and stoichiometric (O/M = 2.00) samples have been achieved. These properties correspond to those needed to optimize the thermo-physical properties measurements. This also shows that the coupling of multi-scale characterizations is mandatory to evaluate the structural and microstructural properties of this kind of material.

Multi-scale microstructure strengthening strategy in CoCrFeNiNb0.1Mo0.3 high entropy alloy overcoming strength-ductility trade-off

Improving the strength of CoCrFeNi high entropy alloy with a face-centered cubic structure has been a subject of significant research in recent years. However, most strengthening mechanisms are limited by the strength-ductility trade-off. In this study, we proposed an entirely new strategy of multi-scale microstructure strengthening for high entropy alloys that significantly enhances the yield strength and ultimate tensile strength to 751 MPa and 1100 MPa, respectively, while maintaining high elongation of 15.1%. Through annealing at different temperatures, we tailored the microstructure of CoCrFeNiNb0.1Mo0.3 alloy, in which a multi-scale microstructure was achieved through inducing the micro-scale Laves phase, nano-scale precipitate and atomic-scale short-range order structure. Through Orowan bypass mechanism and shearing mechanism, the synergistic interaction effect of multi-scale reinforcements can realize a breakthrough in mechanical properties to further overcome strength-ductility trade off, and thus demonstrate a promising approach to improve engineering performance of high entropy alloys.

Fault Diagnosis With Robustness and Lightweight Synergy Under Noisy Environment

Over the years, deep learning-based fault diagnosis approaches have made remarkable progress. However, the diagnosis accuracy has been on a steady decline because of the noise interference in industrial applications. Additionally, the accuracy improvement is often accompanied by an increase in computational cost, which makes it difficult to achieve engineering-embedded deployment. To find the balance between robustness and complexity, a new intelligent fault diagnosis framework named lightweight multi-scale dilated causal convolutional attention network (Li-MDCCAN) is proposed. Based on dual fusion mechanisms, the proposed network realizes high robustness by adopting multi-scale strategy and attention modules, which explore multi-scale and multi-dimensional features along three parallel paths. Furthermore, lightweight convolutional blocks are introduced to reduce the network’s complexity. The performance of Li-MDCCAN is verified with six different datasets and compared with those of four advanced fault diagnosis networks. The experimental results indicate the effectiveness and superiority of Li-MDCCAN in increasing robustness and decreasing complexity. Moreover, prior sensor placement is discussed to guide applications under limited engineering conditions.

Exploring the mechanism of photosensitizer conjugation on membrane perturbation of antimicrobial peptide: A multiscale molecular simulation study

Antimicrobial peptides (AMPs) exert their biological functions by perturbation with cellular membrane. Conjugation of AMPs with photosensitizer (PS) is a promising strategy for enhancing the efficacy and reducing systemic toxicity of AMPs. However, it is still elusive how the conjugated PS impacts the perturbation of AMPs on cell membrane from molecular level. Here, we addressed this issue by a multiscale computational strategy on pyropheophorbide-a (PPA) conjugated K6L9 (PPA-K6L9), a PS-AMP conjugate developed by us previously. Our atomistic molecular dynamics (MD) simulations revealed that the porphyrin moiety of PPA enhanced the stability of the conjugate in a lipid bilayer membrane model. Moreover, such moiety also maintained the amphipathic structure of K6L9, which is crucial for membrane pore formation. Coarse-grained MD simulations further showed that the conjugates aggregated in membrane environment and formed more stable toroidal pores with respect to K6L9 alone, suggesting the conjugation of PPA may enhance the membrane-disruption activity of K6L9. Consistent with this, our cellular experiments confirmed that PPA-K6L9 was more toxic to 4 T1 tumor cells than K6L9. This study provides insights into the mechanism by which PS-AMP conjugates disrupt cellular membranes and could aid in the design of more potent AMP conjugates.

A Multiscale Strategy for Constructing Finned Microporous Superhydrophobic Surface for Highly‐Efficient Condensation Heat Transfer Enhancement Enabled by Lifespan Regulation of Droplets

AbstractSpontaneous droplet jumping on micro‐/nano‐structured superhydrophobic surfaces has been exploited as an efficient means for enhancing steam condensation heat transfer. However, the good performance of such surfaces quickly decays with raising the degree of subcooling, due to the mismatch between the characteristic length scales and droplet sizes when they grow up. Herein, a novel strategy for multiscale droplet regulation is proposed by combining sub‐millimeter fin structure with a hierarchical microporous superhydrophobic surface. A superior condensation heat transfer performance is attained on such hierarchical superhydrophobic finned tube (F‐SHB), in comparison to the baseline case of superhydrophobic non‐finned (SHB) tube under well‐controlled test conditions. Although the droplet jumping is not as vigorous as that on the SHB tube, the finned geometry of the F‐SHB tube leads to a condensation heat transfer enhancement even under high degrees of subcooling up to 36 K, because of the accelerated departure of large droplets by imposing Laplace force gradient in the presence of V‐shaped sub‐millimeter fins. This multiscale enhancement strategy is shown to enable a cascading regulation over the entire lifespan of condensate droplets. The fabrication of F‐SHB tubes is facile and easy to be scaled up, showing great potential in practical steam condensation applications.

SCLMnet: A dual-branch guided network for lung and lung lobe segmentation

Lung and lung lobe segmentation are two crucial techniques for lung imaging analysis that interact in clinical settings. Lung segmentation assists physicians in comparing different images to select the most appropriate surgical plan, while lobe segmentation provides precise anatomical information to help plan surgical procedures. However, inaccurate lung segmentation edges, mis-segmented lobe boundaries, and tiny targets pose challenges. Therefore, we propose a dual-branch guided convolutional neural network, SCLMnet, for lung and lung lobe segmentation. To completely leverage the semantic information of feature maps, the first branch adds a spatial linkage module (SLM) to focus on low-level features at different spatial levels, highlighting feature representations of lung edges and lung lobe boundaries. A channel linkage module (CLM) is added by matrix inner product to model channel relations, emphasizing the relevance and similarity of feature maps and capturing the interdependency of high-level feature channels to highlight feature representations of the entire lung lobe region. Transmodal synaptic linkage (TSL) and multi-scale fusion strategy guide the feature information of the CLM and SLM and the deep features extracted by the second branch ResUNet to jointly explore useful information in chest computer tomography (CT) images. To evaluate the performance of the state-of-the-art model, we use three publicly available datasets: LUNA16, COVID-19-CT-Seg, and VESSEL12. Compared to the existing methods, SCLMnet achieves average Dice scores of 92.17%, 97.80%, and 99.12%, respectively, demonstrating remarkable performance, which suggests that lung and lung lobe segmentation using CT images with SCLMnet can play an essential role in clinical research.

COMPARISON AND EVALUATION OF TLSS AND MOBILE LIDAR SCANNERS FOR MULTI-SCALE 3D DOCUMENTATION OF CULTURAL HERITAGE

Abstract. Field documentation of archaeological sites and built heritage by point-cloud is very effective in yielding arbitrary 3- dimensional shapes of various objects. There are several options for methods and devices such as TLSs, Mobile scanners, and SfM-MVS-based CRP. These are based on different technology, mechanism, and algorithm. Therefore, it is difficult to compare their precision, resolution, and measurement error in the same manner under practical conditions rather than catalog specification. The authors plan and carry out a field experiment of 3D point-cloud measurement in an archaeological site to obtain sample data for comparison between methods and devices under controlled conditions. Through comparative analysis, we evaluate character and advantage, as well as comparative measurement errors between methods and devices, then argue the combined use of different methods and devices under the “Multi-scale 3D scan strategy”.

Anomaly Segmentation for High-Resolution Remote Sensing Images Based on Pixel Descriptors

Anomaly segmentation in high spatial resolution (HSR) remote sensing imagery is aimed at segmenting anomaly patterns of the earth deviating from normal patterns, which plays an important role in various Earth vision applications. However, it is a challenging task due to the complex distribution and the irregular shapes of objects, and the lack of abnormal samples. To tackle these problems, an anomaly segmentation model based on pixel descriptors (ASD) is proposed for anomaly segmentation in HSR imagery. Specifically, deep one-class classification is introduced for anomaly segmentation in the feature space with discriminative pixel descriptors. The ASD model incorporates the data argument for generating virtual abnormal samples, which can force the pixel descriptors to be compact for normal data and meanwhile to be diverse to avoid the model collapse problems when only positive samples participated in the training. In addition, the ASD introduced a multi-level and multi-scale feature extraction strategy for learning the low-level and semantic information to make the pixel descriptors feature-rich. The proposed ASD model was validated using four HSR datasets and compared with the recent state-of-the-art models, showing its potential value in Earth vision applications.

PolarFormer: Multi-Camera 3D Object Detection with Polar Transformer

3D object detection in autonomous driving aims to reason “what” and “where” the objects of interest present in a 3D world. Following the conventional wisdom of previous 2D object detection, existing methods often adopt the canonical Cartesian coordinate system with perpendicular axis. However, we conjugate that this does not fit the nature of the ego car’s perspective, as each onboard camera perceives the world in shape of wedge intrinsic to the imaging geometry with radical (non perpendicular) axis. Hence, in this paper we advocate the exploitation of the Polar coordinate system and propose a new Polar Transformer (PolarFormer) for more accurate 3D object detection in the bird’s-eye-view (BEV) taking as input only multi-camera 2D images. Specifically, we design a cross-attention based Polar detection head without restriction to the shape of input structure to deal with irregular Polar grids. For tackling the unconstrained object scale variations along Polar’s distance dimension, we further introduce a multi-scale Polar representation learning strategy. As a result, our model can make best use of the Polar representation rasterized via attending to the corresponding image observation in a sequence-to-sequence fashion subject to the geometric constraints. Thorough experiments on the nuScenes dataset demonstrate that our PolarFormer outperforms significantly state-of-the-art 3D object detection alternatives.

UAV-BASED ARCHAEOLOGICAL 3D MONITORING: A RURALSCAPE CASE IN IRAQI KURDISTAN

Abstract. Recently rapid mapping techniques based on UAV photogrammetry increasingly help on-site archaeological documentation works. Multi-temporal data are specifically crucial in diachronic investigation research, and for this purpose the data co-registration and integration can support the accurate 3D digitization of excavation phases and make coherent topological relation among them and phases-related stratigraphic units’ data. In this framework, the level of automation and accuracy control are challenging aspects to streamline the documentation process during excavation activities, however all experimentation phases must be tested and validated in the actual archaeological context, where the boundary conditions are typically demanding. This research is developed during the collaboration project with Cà Foscari University of Venice, started in 2022 campaign, in the excavation site of Tell Zeyd, in Iraqi Kurdistan, The Tell Zeyd Archaeological Project (ZAP) aims to study the rural landscape of the hinterland of Mosul in the long Islamic period, from the Arab conquest in the 7th century to the disintegration of the Ottoman Empire with the First World War, as an ideal observatory on the characteristics of the settlement in its spatial organisation, places of worship, production installations and facilities for storing foodstuffs. The research aims to present the preliminary test and results on an experimental documentation activity based on multi-scale UAV mapping strategy and training with the archaeological expert group, and particularly for automatic co-registration of multi-temporal data, considering different images datasets epochs belonging to subsequent excavation phases.

Hierarchical multiscale analysis for 3D woven composite leaf spring landing gear

In this work, a hierarchical multiscale analysis method was used for 3D woven composite leaf spring landing gear. The structural finite element models at micro, meso and macro scales were established and discussed, respectively. The influence of weave patterns on the tensile properties of 3D woven composites were carried out by experiment and numerical simulation. Finally, the multiscale model was established to predict the mechanical properties of 3D woven composite leaf spring landing gear. The results showed that the maximum prediction error was 9.01% at the microscopic scale, 4.55% at the mesoscopic scale, and 5.61% at the macroscopic scale, thus verifying the effectiveness of the used hierarchical multiscale analysis strategy. The presence of stuffer yarn within the material and the yarn arrangement density were the key factors affecting the structural load-bearing capacity in the design of 3D woven composite leaf spring landing gear.

Identifying the essential roles of light and sonication in dual‐stimuli regulated bulk atom transfer radical polymerization by multiscale simulation

AbstractMore and more attention has been paid to the important roles of external fields in controlled radical polymerization (CRP). However, their essential roles have not been studied thoroughly yet, which hinders the in‐depth understanding of the mechanism and kinetics. Herein, a strategy combining quantum chemical calculations (QCC) and kinetic modeling was adopted to identify the essential roles of light and ultrasonication in the dual‐stimuli regulated bulk atom transfer radical polymerization (ATRP). At the molecular level, the impact of light on Cu‐catalyzed ATRP was investigated. The CuIIBr/Me6TREN has high absorbance at an experimental wavelength of 365 nm (main excitation S0 → D7 accounts for 84.93%). Electron transfer reactions involving Me6TREN are more favorable paths for photochemical reactions, and the mechanism of the copper activation/deactivation pathway is inner sphere electron transfer. At the micro‐scale, a kinetic model based on the method of moment was established with a “series” encounter pair model to consider the influence of ultrasound on diffusional limitation. Simulation results show that the changes of the reaction rate coefficients ka, kda, and kt at high conversion reflect the degree of diffusional limitation by ultrasound. The multiscale modeling strategy applied in this study identifies the essential roles of photo and ultrasonication in dual‐stimuli regulated model systems, which can be extended to other external‐field regulated CRP to improve the mechanistic understanding.

Dynamic hedging strategies across assets and commodities – a wavelet analysis

PurposeThe aim of this research is to explore multiscale hedging strategies among cryptocurrencies, commodities, and GCC stocks. Particularly, this is done by evaluating the connectedness among these asset classes covering a period with COVID-19 implications. Using the wavelet approach, the present study aims to recommend whether there exist different time horizon-based hedging abilities across the asset classes.Design/methodology/approachThe approach used in this study is a multiscale decomposition of time series based on wavelets of daily prices of 13 asset classes. Since the wavelet analysis allows to decompose the time series into its frequency components at different time scales by a filtering process the study covered 1-day, 8-day, and 64-day time horizons to examine the hedging properties across those asset classes.FindingsThe results of this study show that hedging effectiveness differs among stock markets over time. In some cases, cryptocurrencies may keep their hedging properties across time while in others they switch from safe haven to hedge devices. In almost all cases, the three main cryptocurrencies showed diversifying properties as was observed by the multiscale correlation and hedge ratio estimations. In a competing sense, gold showed safe haven properties across time than cryptocurrencies except at an 8-day time scale where hedge ratios were low, positive and statistically different from zero that could be interpreted as a good hedge device in the medium term.Research limitations/implicationsThough this research has considered a set of thirteen asset classes, it was limited to a period in which most cryptocurrencies started trading for the first time which reduces the number of observations compared to Bitcoin prices and stable coins such as Ethereum, Ripple, and Bitcoin Cash. Also, the research was focused on the GCC stock markets which may have different results as compared to other regional markets of Asia or Latin America. A comparative analysis in future could be another area of research in future.Practical implicationsThis study has some significant policy implications. The cryptocurrency market is severely affected by demand and risk shocks to crude oil prices during the COVID-19 period. From the investor's point of view, diversification benefits can be obtained by combining cryptocurrencies along with oil-related products during episodes of financial turmoil and COVID-19 pandemic. The GCC region is constantly endeavoring to adopt more scientific tools and mechanisms of investment, and therefore, this study's results will provide some useful directions to the government, policymakers, financial institutions, and investors.Originality/valueThe current study covers a big bunch of 13 assets spanning across financial and real assets. This is based on literature gap and hence, will be a significant addition to the existing literature. Moreover, the GCC region is emerging as a global investment hub and this study will provide investors dynamic hedging strategies across these asset classes.

Strategies for Achieving Sustainability of Water Supply Systems in Rural Environments with Community Management in Brazil

An urban–rural dichotomy currently exists in policy documentation, accessibility to resources, and human rights. In Brazil, there remains a significant deficit in water and sanitation services coverage, particularly in rural areas. To achieve universal access to water, the legal framework addresses the development of decentralization actions and co-participation in management. As a result, a community-managed water supply (CMWS) has emerged as the social tool in which the community effectively participates in decision-making and oversees the operation and maintenance of water systems. The long-term sustainability of these systems is affected by critical factors in implementation, management, and maintenance. However, no strategic plan exists to solve/mitigate the critical factors commonly associated with CMWS. We proposed an action plan with multiscale intersectoral management strategies to guarantee water security for rural populations. The strategies were separated into areas of responsibility discussed in the recent literature and evaluated by experts. The strategies at the national level were related to institutional aspects and policy formulation. At the local level, the strategies were related to supporting CMWS from project implementation to post-construction. Finally, at the community level, the strategies were related to the quality and financial recognition of the services.

Photovoltaic Power Forecasting Using Multiscale-Model-Based Machine Learning Techniques

The majority of energy sources being used today are traditional types. These sources are limited in nature and quantity. Additionally, they are continuously diminishing as global energy consumption increases as a result of population growth and industrial expansion. Their compensation is made from clean energy and renewable energy. Renewable energy is strongly dependent on climatic conditions; therefore, an aspect of energy management is needed, which is essential in distribution systems, because it enables us to calculate the precise energy used by the load as well as by its many components. It also helps us understand how much energy is required and its origin. The energy management aspect contains two main phases: forecasting and optimization. In this study, we are focused on the forecasting level using intelligent machine learning (ML) techniques. To ensure better energy management, it is very important to predict the production of renewable energy over a wide time period. In our work, several cases are proposed in order to predict the temperature, the irradiance, and the power produced by a PV system. The proposed approach is validated by an experimental procedure and a real database for a PV system. The big data from the sensors are noisy, which pose a major problem for forecasting. To reduce the impact of noise, we applied the multiscale strategy. To evaluate this strategy, we used different performance criteria, such as mean error (ME), mean absolute error (MAE), root mean square error (RMSE), nRMSE and the coefficient of determination (R2). The obtained experimental results show good performance with lower error. Indeed, they achieved an error for nRMSE criteria between 0.01 and 0.37.