Artificial Neural Network Genetic Algorithm Research Articles

Development of a Genetic Algorithm – Artificial Neural Network model to optimize the Dimensional Accuracy of parts printed by FFF

PurposeFused Filament Fabrication (FFF) is one of the growing technologies in additive manufacturing, that can be used in a number of applications. In this method, process parameters can be customized and their simultaneous variation has conflicting impacts on various properties of printed parts such as dimensional accuracy (DA) and surface finish. These properties could be improved by optimizing the values of these parameters.Design/methodology/approachIn this paper, four process parameters, namely, print speed, build orientation, raster width, and layer height which are referred to as “input variables” were investigated. The conflicting influence of their simultaneous variations on the DA of printed parts was investigated and predicated. To achieve this goal, a hybrid Genetic Algorithm – Artificial Neural Network (GA-ANN) model, was developed in C#.net, and three geometries, namely, U-shape, cube and cylinder were selected. To investigate the DA of printed parts, samples were printed with a central through hole. Design of Experiments (DoE), specifically the Rotational Central Composite Design method was adopted to establish the number of parts to be printed (30 for each selected geometry) and also the value of each input process parameter. The dimensions of printed parts were accurately measured by a shadowgraph and were used as an input data set for the training phase of the developed ANN to predict the behavior of process parameters. Then the predicted values were used as input to the Desirability Function tool which resulted in a mathematical model that optimizes the input process variables for selected geometries. The mean square error of 0.0528 was achieved, which is indicative of the accuracy of the developed model.FindingsThe results showed that print speed is the most dominant input variable compared to others, and by increasing its value, considerable variations resulted in DA. The inaccuracy increased, especially with parts of circular cross section. In addition, if there is no need to print parts in vertical position, the build orientation should be set at 0° to achieve the highest DA. Finally, optimized values of raster width and layer height improved the DA especially when the print speed was set at a high value.Originality/valueBy using ANN, it is possible to investigate the impact of simultaneous variations of FFF machines’ input process parameters on the DA of printed parts. By their optimization, parts of highly accurate dimensions could be printed. These findings will be of significant value to those industries that need to produce parts of high DA on FFF machines.

Chaos Game Optimization-Hybridized Artificial Neural Network for Predicting Blast-Induced Ground Vibration

In this study, we introduced the chaos game optimization-artificial neural network (CGO-ANN) model as a novel approach for predicting peak particle velocity (PPV) induced by mine blasting. The CGO-ANN model is compared with other established methods, including the particle swarm optimization-artificial neural network (PSO-ANN), the genetic algorithm-artificial neural network (GA-ANN), single ANN, and the USBM empirical model. The aim is to demonstrate the superiority of the CGO-ANN model for PPV prediction. Utilizing a dataset comprising 180 blasting events from the Tonglushan Copper Mine in China, we investigated the performance of each model. The results showed that the CGO-ANN model outperforms other models in terms of prediction accuracy and robustness. This study highlights the effectiveness of the CGO-ANN model as a promising tool for PPV prediction in mining operations, contributing to safer and more efficient blasting practices.

Post-Treatment and Hybrid Techniques for Prolonging the Service Life of Fused Deposition Modeling Printed Automotive Parts: A Wear Strength Perspective

<div>This study aims to explore the wear characteristics of fused deposition modeling (FDM) printed automotive parts and techniques to improve wear performance. The surface roughness of the parts printed from this widely used additive manufacturing technology requires more attention to reduce surface roughness further and subsequently the mechanical strength of the printed geometries. The main aspect of this study is to examine the effect of process parameters and annealing on the surface roughness and the wear rate of FDM printed acrylonitrile butadiene styrene (ABS) parts to diminish the issue mentioned above. American Society for Testing and Materials (ASTM) G99 specified test specimens were fabricated for the investigations. The parameters considered in this study were nozzle temperature, infill density, printing velocity, and top/bottom pattern. The hybrid tool, i.e., GA–ANN (genetic algorithm–artificial neural network) has been opted to train, predict, and optimize the surface roughness and sliding wear of the printed parts. Results disclose that the minimum surface roughness obtained with GA–ANN was 1.05482 μm for infill density of 68%, nozzle temperature of 230°C, printing velocity of 80 mm/sec, and for concentric type of top/bottom pattern. In extension of this study, annealing was performed on the specimens printed on the optimized results obtained from the analysis at three different temperatures of 110°C, 150°C, and 190°C and for a fixed period of time of 60 min as a post-treatment process to further study the impact of annealing on the surface roughness and wear rate. The surface roughness of the samples showed a discernible improvement as a result of annealing, which can further make significant inroads in automotive industries.</div>

Application of ionic liquid ultrasound-assisted extraction (IL-UAE) of lycopene from guava (Psidium guajava L.) by response surface methodology and artificial neural network-genetic algorithm

Lycopene-rich guava (Psidium guajava L.) exhibits significant economic potential as a functional food ingredient, making it highly valuable for the pharmaceutical and agro-food industries. However, there is a need to enhance the extraction methods of lycopene to fully exploit its beneficial uses. In this study, we evaluated various ionic liquids to identify the most effective one for extracting lycopene from guava. Among thirteen ionic liquids with varying carbon chains or anions, 1-butyl-3-methylimidazolium chloride demonstrated the highest productivity. Subsequently, a single-factor experiment was employed to test the impact of several parameters on the efficiency of lycopene extraction using this selected ionic liquid. These parameters included extraction time, ultrasonic power, liquid-solid ratio, concentration of the ionic liquid, as well as material particle size. Moreover, models of artificial neural networks using genetic algorithms (ANN-GA) and response surface methodology (RSM) were employed to comprehensively assess the first four key parameters. The optimized conditions for ionic liquid ultrasound-assisted extraction (IL-UAE) were determined as follows: 33 min of extraction time, 225 W of ultrasonic power, 22 mL/g of liquid-solid ratio, 3.0 mol/L of IL concentration, and extraction cycles of three. Under these conditions, lycopene production reached an impressive yield of 9.35 ± 0.36 mg/g while offering advantages such as high efficiency, time savings, preservation benefits, and most importantly environmental friendliness.

Process Optimization for the Extraction of Phenolic Compounds from Pomegranate Peels: Response Surface Methodology-Desirability Function and Artificial Neural Network-Genetic Algorithm

Valorization of agricultural wastes is ongoing topic in industry. Determining the best conditions by artificial neural networks based optimization techniques is the key step to extract valuable compounds efficiently and to obtain high quality extracts. In this study, the response surface methodology (RSM)-desirability function (DF) and artificial neural network (ANN)-genetic algorithm (GA) approaches were compared in modeling and optimization the extraction parameters (temperature, time and ethanol concentration (ratio of ethanol to water, % v/v)) of phenolic compounds in pomegranate peels. The ANN-GA approach providing higher coefficient of determination and lower root mean square deviation showed better predictive capability than the RSM. The optimum time (81.4 min) and ethanol concentration (15.7%) of RSM-DF approach shifted to the lower levels (78.8 min and 15.3%) with the ANN-GA approach while the optimum temperature (54.0°C) shifted to a higher level (59.3°C). The use of these values provided total phenolic content of >1000 mg GAE L-1 and the corresponding antioxidant activity was 11 mmol TE L-1. As a result, increasing temperature up to a critical level decreased the extraction time and ethanol concentration, and it was determined that higher time-temperature combinations must be used for the complete water-based extraction of phenolic compounds from plant wastes in comparison to ethanol-water based extraction.

Building structure-borne noise measurements and estimation due to train operations in tunnel

The perception of structure-borne noise is particularly salient when train passes through the tunnel under the buildings, which has a negative impact on human health. In the process of constructing buildings along metro lines, it is crucial to estimate indoor structure-borne noise levels in order to enhance design and prevent any negative impact on human comfort. This study conducted measurements of structure-borne noise, reverberation time, and train-induced vibrations in Guangzhou, China to investigate the generation, propagation, and dissipation mechanisms of structure-borne noise. An approach based on Short-Time Fourier Transform and Schroeder integral was proposed for obtaining frequency-dependent reverberation time. Additionally, a deep learning-based approach incorporating indoor vibrations, frequency-dependent reverberation time, and room parameters as inputs was proposed based on Genetic Algorithm-Artificial Neural Network. The estimated structure-borne noise levels demonstrated good agreement with measured values, indicating the feasibility of the approach. The finding of this research facilitates a clear comprehension of the generation, distribution, and dissipation mechanisms of indoor structure-borne noise for engineers while also enabling convenient acquisition of indoor structure-borne noise. The estimated noise levels can be effectively utilized during building design processes along metro lines to mitigate adverse impacts on human comfort.

Modeling and Optimization of the Culture Medium for Efficient 4′-N-Demethyl-Vicenistatin Production by Streptomyces parvus Using Response Surface Methodology and Artificial-Neural-Network-Genetic-Algorithm

4′-N-demethyl-vicenistatin is a vicenistatin analogue that has better antitumor activity with promising applications in the pharmaceuticals industry. The harnessing of the complete potential of this compound necessitates a systematic optimization of the culture medium to enable the cost-effective production of 4′-N-demethyl-vicenistatin by Streptomyces parvus SCSIO Mla-L010/ΔvicG. Therefore, in this study, a sequential approach was employed to screen the significant medium compositions, as follows: one-factor-at-a-time (OFAT) and Plackett–Burman designs (PBD) were initially utilized. Cassava starch, glycerol, and seawater salt were identified as the pivotal components influencing 4′-N-demethyl-vicenistatin production. To further investigate the direct and interactive effects of these key components, a three-factor, five-level central composite design (CCD) was implemented. Finally, response surface methodology (RSM) and an artificial-neural-network-genetic-algorithm (ANN-GA) were employed for the modeling and optimization of the medium components to enhance efficient 4′-N-demethyl-vicenistatin production. The ANN-GA model showed superior reliability, achieving the most 4′-N-demethyl-vicenistatin, at 0.1921 g/L, which was 17% and 283% higher than the RSM-optimized and initial medium approaches, respectively. This study represents pioneering work on statistically guided optimization strategies for enhancing 4′-N-demethyl-vicenistatin production through medium optimization.

Controlling recast layer and material removal in reverse µEDM for sustainable fabrication of microelectrode arrays

Microelectrode arrays (MEAs) are essential for researching neurons and heart tissue because they record and stimulate neuronal activity. For manufacturing MEAs on conductive materials, reverse microelectrical discharge machining, or reverse µEDM, presents a viable method. However, achieving precise MEAs requires control over recast layers generated during micromachining. This study focuses on recast layer thickness (RLT) and material removal rate (MRR) as key performance indicators for sustainable reverse µEDM fabrication of MEAs. RLT affects surface stability, while MRR influences energy consumption and cost. Optimising these responses depends on three input factors: feed rate, capacitance, and voltage. ANOVA analysis reveals that capacitance is crucial in both MRR and recast layer thickness (RLT) in reverse µEDM, contributing to 88.71% and 94.04% simultaneously. A hybrid artificial neural network-genetic algorithm for global optimisation was employed. The optimised parameters 46.5nF capacitance, 110 V voltage, and 5 µm/sec feed yielded a minimum RLT of 1.58 µm and a maximum MRR of 0.3 mm³/min. These advancements signify significant improvements in the precision and efficiency of MEAs fabrication through reverse µEDM, contributing to its sustainable implementation.

A Comparative Study on the Estimation of Wind Speed and Wind Power Density Using Statistical Distribution Approaches and Artificial Neural Network-Based Hybrid Techniques in Çanakkale, Türkiye

In recent years, wind energy has become remarkably popular among renewable energy sources due to its low installation costs and easy maintenance. Having high energy potential is of great importance in the selection of regions where wind energy investments will be made. In this study, the wind power potential in Çanakkale Province, located in the northwest of Türkiye, is examined, and the wind speed is estimated using hourly and daily data over a one-year period. The data, including 12 different meteorological parameters, were taken from the Turkish State Meteorological Service. The two-parameter Weibull and Rayleigh distributions, which are the most widely preferred models in wind energy studies, are employed to estimate the wind power potential using hourly wind speed data. The graphical method is implemented to calculate the shape (k) and scale (c) parameters of the Weibull distribution function. Daily average wind speed estimation is performed with artificial neural network–genetic algorithm (ANN-GA) and ANN–particle swarm optimization (ANN-PSO) hybrid approaches. The proposed hybrid ANN-GA and ANN-PSO algorithms provide correlation coefficient values of 0.94839 and 0.94042, respectively, indicating that the predicted and measured wind speed values are notably close. Statistical error indices reveal that the ANN-GA model outperforms the ANN-PSO model.

Advanced chemometric methods for simultaneous quantitation of caffeine, codeine, paracetamol, and p-aminophenol in their quaternary mixture

Two different multivariate techniques have been applied for the quantitative analysis of caffeine, codeine, paracetamol and p-aminophenol (PAP) in quaternary mixture, namely, Partial Least Squares (PLS-1) and Artificial Neural Networks (ANN). For suitable analysis, a calibration set of 25 mixtures with various ratios of the drugs and PAP impurity were established using a 4-factor 5-level experimental design. The most meaningful wavelengths for the chemometric models were chosen using Genetic Algorithm (GA) as a variable selection technique. By using an independent validation set, the validity of the proposed methods was evaluated. A comparative study was established between the three multivariate models (PLS-1, GA–PLS and GA–ANN). The comparison between the various models revealed that the GA–ANN model was superior at resolving the highly overlapped spectra of this quaternary combination. The drugs were successfully quantified in their pharmaceutical dosage form utilizing the GA–ANN models.

Comparison of an Ultrasound-Assisted Aqueous Two-Phase System Extraction of Anthocyanins from Pomegranate Pomaces by Utilizing the Artificial Neural Network-Genetic Algorithm and Response Surface Methodology Models.

As a by-product of pomegranate processing, the recycling and reuse of pomegranate pomaces (PPs) were crucial to environmentally sustainable development. Ultrasound-assisted aqueous two-phase extraction (UA-ATPE) was applied to extract the anthocyanins (ACNs) from PPs in this study, and the central composite design response surface methodology (CCD-RSM) and artificial neural network-genetic algorithm (ANN-GA) models were utilized to optimize the extraction parameters and achieve the best yield. The results indicated that the ANN-GA model built for the ACN yield had a greater degree of fit and accuracy than the RSM model. The ideal model process parameters were optimized to have a liquid-solid ratio of 49.0 mL/g, an ethanol concentration of 28 g/100 g, an ultrasonic time of 27 min, and an ultrasonic power of 330 W, with a maximum value of 86.98% for the anticipated ACN yield. The experimental maximum value was 87.82%, which was within the 95% confidence interval. A total of six ACNs from PPs were identified by utilizing UHPLC-ESI-HRMS/MS, with the maximum content of cyanidin-3-O-glucoside being 57.01 ± 1.36 mg/g DW. Therefore, this study has positive significance for exploring the potential value of more by-products and obtaining good ecological and economic benefits in the future.

Analyzing Temperature Distribution, Mass Transport, and Cell Performance in PEM Fuel Cells with Emphasis on GDL Face Permeability and Thermal Contact Resistance Parameters.

Temperature distribution, mass transport, and current density are crucial parameters to characterize the durability and output performance of proton exchange membrane fuel cell (PEMFC), which are affected by thermal contact resistance (TCR) and gas diffusion layer (GDL) face permeability within both cathode and anode GDL porous jumps. This study examined the effects of TCR and GDL face permeability on a single PEM fuel cell's temperature profiles, mass transport, and cell performance using a three-dimensional, nonisothermal computational model with an isotropic gas diffusion layer (GDL). This model calculates the ideal thermal contact resistance by comparing the expected plate-cathode electrode temperature difference to the numerical and experimental literature. The combined artificial neural network-genetic algorithm (ANN-GA) method is also applied to identify the optimum powers and their operating conditions in six cases. Theoretical findings demonstrate that TCR and suitable GDL face permeability must be considered to optimize the temperature distribution and cell efficiency. TCR and GDL face permeability lead to a 1.5 °C rise in maximum cell temperature at 0.4 V, with a "Λ" shape in temperature profiles. The TCR and GDL face permeability also significantly impacts electrode heat and mass transfer. Case 6 had 1.91, 6.58, and 8.72% higher velocity magnitudes, oxygen mass fractions, and cell performances than case 1, respectively. Besides, the combined ANN-GA method is suitable for predicting fuel cell performance and identifying operation parameters for optimum powers. Therefore, the findings can improve PEM fuel cell performance and give a reference for LT-PEMFC design.

Impact of extrusion processing on bioactive compound enriched plant-based extrudates: A comprehensive study and optimization using RSM and ANN-GA

The present study aimed to investigate the influence of extrusion process parameters, specifically screw speed (300 to 500 rpm) and temperature (130 to 150 °C) on the stability of phytosterols and the physical, functional, and structural characteristics of extrudates made from composite blends containing pea protein isolate (PPI). Response Surface Methodology (RSM) and Artificial Neural Network-Genetic Algorithm (ANN-GA) were employed to optimize the process parameters.Extrudates produced at lower temperatures exhibited higher phytosterol retention compared to those produced at higher temperatures. X-ray diffraction analysis showed that incorporating PPI significantly increased crystallinity from 22% to 33%. Fourier Transform Infrared analysis revealed a reduction in the Amide II stretching of the secondary structure in pea protein isolate following extrusion. Scanning electron microscopy images demonstrated that the addition of PPI and functional oil altered the structure of the extrudates.Statistical comparison of the process parameters using higher R2 values, lower root mean square as well as mean absolute percentage error indicated that the ANN model outperformed the RSM model in predicting all responses. Overall, incorporating PPI and phytosterols into corn-based extrudates improved their nutritional quality. These findings contribute to meeting the demand for fortified foods.

Artificial intelligence model for prediction of cardiovascular disease: An empirical study

Cardiovascular disease (CVD) is a disease related to the heart and blood vessels. Prediction of CVD is essential for early detection and diagnosis, which is however compounded by the complex interplay between medical history, physical examination outcomes, and imaging results. While the existing automated systems are fraught with the usage of irrelevant and redundant attributes, artificial intelligence (AI) helps in the identification of potential CVD populations by prediction models. This work aims at developing an AI model for predicting CVD using different classifications of machine learning techniques. The CVD dataset was obtained from the UCI repository containing about 76 cardiac attributes for training in various machine learning models, which include a hybrid of artificial neural network-genetic algorithm (ANN-GA), artificial neural network, support vector machine (SVM), K-means, K-nearest neighbor (KNN), and decision tree (DT). The performance of the models was measured in terms of accuracy, means square error, sensitivity, specificity, and precision. The results showed that the hybrid model of ANN-GA performs better with an accuracy of 86.4%, compared to the SVM, K-means, KNN, and DT measured at 84.0%, 59.6%, 79.0%, and 77.8%, respectively. It was observed that the system performs better as the number of datasets increases in the database, with a fewer selection of attributes using genetic algorithm for selection. Thus, the ANN-GA model is recommended for CVD prediction and diagnosis.

Response surface methodology and artificial neural network-genetic algorithm for modeling and optimization of bioenergy production from biochar-improved anaerobic digestion

Biochar can be used to improve the anaerobic digestion (AD) of agricultural wastes for higher methane production. However, the interaction of biochar addition with other factors of the anaerobic co-digestion (Co-AD) process has rarely been investigated. In this study, process models based on response surface methodology (RSM) and artificial neural network (ANN) were compared in modeling the methane yield (MY, mL CH4/g VS added) from the Co-AD of poultry litter and wheat straw with biochar addition. Box-Behnken design was applied, with the controlling parameters being carbon to nitrogen ratio (C/N), total solids (TS, %), and biochar addition (Biochar, % TS). Numerical optimization and genetic algorithm (GA) were used as optimization tools for RSM and ANN, respectively. A significant second-order quadratic model was built by RSM (R2 = 0.9981 and RMSE = 0.91), which demonstrated significant interactions between C/N and TS (p < 0.0001), and between C/N and Biochar (p < 0.05). The trained ANN (3−3−1) was less accurate (R2 = 0.9926, RMSE = 1.80) compared to RSM. The optimization results by RSM and ANN coupled with GA (ANN-GA) were both validated with prediction errors <0.5%. The optimization results by ANN-GA should be used since it generated a higher maximum MY of 290.7 ± 0.2 mL CH4/g VS added, under the optimal conditions of C/N ratio 24.46, TS 5.03%, and Biochar 8.73% TS, showing an improvement of 20.6% (compared to the control) through process optimization. The methods can also be applied in other scenarios for process modeling and optimization. The optimized results could support real applications of using additives including biochar, active carbon, nanoparticles, etc., to promote the bioenergy production from AD of agricultural wastes.

Process optimization of microwave-assisted aqueous extraction of tannins and saponins from Malabar Spinach (Basella alba) leaves using ANN-GA and RSM methodology

The effect of microwave-assisted aqueous extraction (MAE) conditions on the recovery and quantification of tannins and saponins from Basella alba leaves has been explored. The experiment was carried out according to the Box-Behnken experimental design of response surface methodology (RSM) with 17 runs. The independent variables of the study were microwave power (100 to 300 W), treatment time (5 to 15 min), and feed/solvent ratio (0.025 to 0.05). The effect of extraction conditions on the response variables, total condensed tannin, total hydrolyzed tannin, and total saponin content, were predicted using RSM and artificial neural network-genetic algorithm (ANN-GA) methodology. The extracted values of total condensed tannin, total hydrolyzed tannin, and total saponin content, using RSM and ANN-GA were compared and evaluated for goodness of fit. The predicted values using RSM and ANN-GA models for total condensed tannin were found to be 0.18 and 0.26 mg TAE g−1, total hydrolyzed tannin were 0.14 and 0.15 mg TAE g−1, and total saponin content were 1.66 and 1.49 mg DE g−1, respectively. When compared to RSM, ANN-GA demonstrated higher value of R2 and lower values of other statistical parameters. Additionally, the predicted value of ANN-GA was more closely aligned with the experimental value.

Tertiary treatment using ultrafiltration in an existing sewage treatment plant for industrial reuse: a modelling approach using artificial neural network with uncertainty estimation

Abstract Navi Mumbai Municipal Corporation of Maharashtra state, India, unified a tertiary treatment plant (TTP) of 20 × 103 m3/day capacity with ultrafiltration technology in an existing Koparkhairane sewage treatment plant (STP) for producing effluent quality usable for industrial purposes. As prior art, an artificial neural network-genetic algorithm (ANN-GA) along with uncertainty estimation using prediction interval is employed to model secondary treated effluent (STE) flow rate (QT) and other quality parameters such as biochemical oxygen demand, chemical oxygen demand, and total suspended solids (TSS) to conclude the reliability of the range in which the input available to TTP. ANN-GA model provides a coefficient of determination above 0.90 for all STE parameters modelled other than TSS. Inferring that a good quantity and quality of 20 × 103 m3/day STP treated water is currently available, where a decreasing trend of QT is also noticed and highlighted. Further, the Wilcoxon signed-rank test on the quality parameter of effluent TTP for industrial reuse standard infers TSS shows infringement during the initial period but started adhering to standards over time. The research delineates at the outset of exploring water reuse policy in India, emphasizing Maharashtra state, modelling STE using ANN-GA and performance evaluation of TTP.

An ensemble neural network for optimising a CNC milling process

Computer numerical control (CNC) milling is a common method for the efficient mass production of products. Process efficiency and product quality have a strong dependency on the cutting process conditions. Furthermore, optimising a process for material removal rate (MRR) and surface roughness (SR), which are measures of process efficiency and product quality, respectively, is a complex optimisation task due to their contrasting relationships with process parameters. In this work, CNC end milling is performed on aluminium 6061 to investigate the influence of key process input variables (feed per tooth, cutting speed, and depth of cut) on the machined part’s SR. Firstly, a full factorial parametric study is conducted and analysed using Analysis of Variance (ANOVA) before an Ensemble Neural Network (ENN) is trained on the experimental data. To capture the complex nonlinear relationships accurately, each base model of the ENN is a combined genetic algorithm-artificial neural network, whose hyperparameters are optimised using a Bayesian optimisation framework. Once trained, the ENN predictive model is exploited to identify optimal input parameter permutations to achieve a predefined SR value while maximising MRR. Analysing the experimental data demonstrates that the SR performance envelope is nonlinear with respect to the input variables. Furthermore, the ANOVA results indicate that feed per tooth is the dominant input parameter with a contribution of 40.2% and that there are strong interactions between the input parameters investigated. The ENN performance was subsequently validated through a further set of experiments producing a mean absolute percentage error in predicted SR of just 2.56%.

Estimation of downwelling surface longwave radiation for cloudy skies by considering the radiation effect from the entire cloud layers

Cloud base height (CBH) is a key parameter to characterize the cloud radiation effect. However, the CBH used in downwelling surface longwave radiation (DSLR) estimation is generally obtained indirectly through cloud top parameters retrieved by passive optical remote sensing instruments, which is of high uncertainty. At the same time, it is unreasonable to replace the effective radiation of the entire cloud layers by only using the cloud base radiation with single layer cloud model. This study proposes a new method to estimate cloudy-sky DSLR, which considers the radiation effect of the entire cloud layers from the cloud base to top. First, the CBH estimation model is established by the genetic algorithm-artificial neural network (GA-ANN) algorithm. The cloud top height and cloud attribute parameters (cloud optical depth, cloud water path, and cloud phase) from the passive remote sensing are used as the input features, and meanwhile CBH data from the active remote sensing are output features in the training and testing process of the model. Then, the cloud base temperature (CBT) is estimated based on the CBH combined with the temperature profile data in the EAR5 reanalysis data. Finally, the effective temperature of the entire cloud layers is calculated by using CBT and cloud top temperature. The verification results of CBH estimation showed that R2 is 0.83, the bias and root mean square error (RMSE) are 0.02 km and 1.56 km, respectively, which indicates a comparable accuracy and higher stability compared with the previous studies. The ground-based measurements in the SURFRAD network are used to validate the newly proposed DSLR estimation method, and the results showed that the bias and RMSE are 5.27 W/m2 and 28.48 W/m2, respectively. Additionally, this study found that although the effective temperature of the entire cloud layers has a weaker linear correlation with DSLR, the radiation contribution generated by cloud still occupies a certain weight, and the maximum ratio of cloud radiation in DSLR estimation can account for 30%. Therefore, the cloud radiation effect must be taken into account in the estimation of cloudy-sky DSLR.

Enhancement of isoprene production in engineered Synechococcus elongatus UTEX 2973 by metabolic pathway inhibition and machine learning-based optimization strategy

An engineered Synechococcus elongatus UTEX 2973-IspS.IDI is used to enhance isoprene production through geranyl diphosphate synthase (CrtE) inhibition and process parameters (light intensity, NaHCO3 and growth temperature) optimization approach. A cumulative isoprene production of 1.21 mg/gDCW was achieved with productivity of 12.6 μg/gDCW/h in culture supplemented with 20 μg/mL alendronate. This inhibition strategy improvises the cumulative isoprene production 5.76-fold in presence of alendronate. The maximum cumulative production of isoprene is observed to be 5.22 and 6.20 mg/gDCW (54.4 and 64.6 μg/gDCW/h) at statistical and artificial neural network genetic algorithm (ANN-GA) optimized conditions, respectively. The overall increase of isoprene production is found to be 29.52-fold using an integrated approach of inhibition and ANN-GA optimization in comparison to unoptimized cultures without alendronate. This study reveals that alendronate use as a potential inhibitor and machine learning based optimization is a better approach in comparison to statistical optimization to enhance the isoprene production.