Average Relative Error Values Research Articles

A Prediction Formula for The Estimation of Sediment Load in The Upper Reach of Al-Gharraf River

The presence of deposition in the river decreases the river flow capability's efficiency due to the absence of maintenance along the river. In This research, a new formula to evaluate the sediment capacity in the upstream part of Al-Gharraf River will be developed. The current study reach lies in Wasit province with a distance equal to 58 km. The selected reach of the river was divided into thirteen stations. At each station, the suspended load and the bedload were collected from the river during a sampling period extended from February 2019 till July 2019. The samples were examined in the laboratory with a different set of sample tests. The formula was developed using data of ten stations, and the other three stations were used for validation. The determination coefficient, root mean square error and average relative error values were equal to 0.987,0.97 kg/s and 7%, respectively. Also, the values of the sediment load that resulted from the formula close to the results of the HEC-RAS model from a previous study, and the determination coefficient, root mean square error, and average relative error values were equal to 0.988, 0.88 kg/s, and 7 % respectively for the simulated model.

Study on strength influencing factors and strength model of cemented backfill with mixed aggregate

The strength of cemented backfill is one of the important indexes affecting the filling effect. There are many factors that affect the strength of cemented backfill. In order to explore the influence law of the cemented backfill strength, firstly, the strength test of cemented backfill with mixed aggregate was carried out on the basis of physical and chemical analysis of test materials and the particle size distribution analysis of mixed aggregate. Then, according to the test results, the effects of particle size distribution of mixed aggregate, solid mass concentration Cw,%, cement content Bw,%, and their interaction on the strength of backfill were analyzed. Finally, the strength model of cemented backfill with mixed aggregate was established by using industrial filling parameters such as aggregate gradation, solid mass concentration Cw,%, and cement content Bw,%, and the reliability of the model was analyzed, and the model verification test was carried out. The results show that the strength of cemented backfill is affected by single factors such as aggregate gradation, solid mass concentration Cw,%, and cement content Bw,% and is also affected by the interaction of these factors. Except for a few points, the relative error of the model is less than 10%. The verification test results show that the average relative error values of 3-day, 7-day, and 28-day strength of cemented backfill with mixed aggregate are 6.32%, 2.46%, and 0.92%, respectively, which shows that the model has a good applicability for predicting the strength of the backfill with mixed aggregate in Jinchuan mine.

Estimates of greenhouse gas emission in Turkey with grey wolf optimizer algorithm-optimized artificial neural networks

The main purpose of this study was to predict Turkey’s future greenhouse gas (GHG) emissions using an artificial neural network (ANN) model trained by a grey wolf optimizer (GWO) algorithm. Gross domestic product, energy consumption, population, urbanization rate, and renewable energy production data were used as predictor variables. To probe the accuracy of the proposed model, the new ANN-GWO model’s performance was compared with the performance of ANN-BP (back propagation), ANN-ABC (artificial bee colony), and ANN-TLBO (teaching–learning-based optimization) models using multiple error criteria. According to calculated error values, the ANN-GWO models predicted GHG emissions more accurately than classical ANN-BP, ANN-ABC, and ANN-TLBO models. According to the average relative error values calculated for the test set, ANN-GWO performs 32.23% better than ANN-BP, 35.29% better than ANN-ABC, and 19.33% better than ANN-TLBO. Using the ANN-GWO model, GHG emissions were forecasted out to 2030 under three different scenarios. The predictions obtained, consistent with a prior forecasting study in the literature, indicated that GHG emissions are expected to outpace official predictions (model prediction range for 2030, 956.97–1170.54 Mt CO2 equivalent). The present study demonstrated that GHG emissions can be predicted accurately with an ANN-GWO model, and that the GWO optimization method is advantageous for predicting future GHG emissions.

Measurement enhancement of TDLAS based on variable weighted cross-correlation tomography for the simultaneous reconstruction of 2D temperature and concentration

Tunable diode laser absorption tomography is a technology that enables the visualization of 2D temperature and concentration in combustion fields. This study proposes a method called variable weighted cross-correlation tomography to improve the traditional two-wavelength scheme affected by signal noise and bias error. The proposed approach combines the multiplicative algebraic reconstruction technique and pattern matching at nine wavelengths by adding peak and bottom wavelengths. In addition, this method iteratively calculates using a corrective multiplication line-of-sight and a modified return process to simultaneously estimate the images of 2D temperature and concentration. Numerical tests are performed using various thermodynamic models with additive noise. The validation experiments involving premixed methane-air flames demonstrate the good agreement between the average relative error values of the reconstructed temperature and the temperature measured using thermocouple (3.32 %), which can be ascribed to the introduction of modified broadening coefficients.

Energy-related carbon dioxide emission forecasting of four European countries by employing data-driven methods

Carbon dioxide emission of countries is deeply dependent on the energy system. Share of different energy resources in primary energy consumption of the countries has principal role in the emission of energy-related carbon dioxide. As well as energy consumption, the level of economic activities performs substantial role in the emission of greenhouse gases. By using data-driven methods such as artificial neural networks (ANNs), the emission of greenhouse gases can be precisely modeled. In this work, two types of ANNs, group method of data handling (GMDH) and multi-layer perceptron (MLP), are employed for estimating carbon dioxide, as one of the most important greenhouse gases, emission of four European countries including UK, Germany, Italy and France; in this regard, consumptions of various energy resources in addition to GDP, as an indicator for economic activities, of the mentioned countries are used as the inputs for modeling and forecasting. Comparison of the actual and predicted data reveals great performance of the employed approaches in modeling. R-squared values of the regression by using both GMDH and MLP are 0.9999. In addition, according to the values of average relative error of the models, it is found that using MLP is preferred due to its lower value compared with GMDH. Obtained values of absolute relative deviations of GMDH and MLP models are 0.39% and 0.33%, respectively.

Artificial neural network model for reservoir petrophysical properties: porosity, permeability and water saturation prediction

Prediction of reservoir petrophysical properties from well-logs data has evolved from the use of experts’ knowledge and statistics to the use of artificial intelligence (AI) models. Several AI models for this purpose are available in the literature; regrettably, most of them are multiple-inputs single-output (MISO) models. Meaning, these models used numerous well-logs to predict a single reservoir petrophysical property. In this study, multiple-inputs multiple-outputs (MIMO) artificial neural network (ANN) model to predict reservoir petrophysical properties: porosity ( $$\varphi$$ ), permeability ( $$k$$ ), and water saturation ( $$S_{{\text{w}}}$$ ), was developed based on wireline logs (gamma ray, resistivity, density and depth interval logs) from 15 fields in the Niger Delta region. The developed ANN model is a feed-forward back-propagation (FFBP) network with 12 neurons in its hidden layer with the Levenberg–Marquardt algorithm as the best learning algorithm than Bayesian regularization and Scaled conjugate gradient. The performance of the developed model resulted in the overall correlation coefficient (R) and Mean Square Error (MSE) values of 0.9945 and 0.7310, respectively. Again, the generalization potential of the developed ANN model with new datasets was determined using five performance indicators: R, coefficient of determination (R2), MSE, root mean square error (RMSE) and average relative error (ARE). The results obtained showed that $$\varphi$$ had R of 0.9243, R2 of 0.8544, MSE of 1.7243, RMSE of 1.3131 and ARE of 0.0652, while $$k$$ had R of 0.9810, R2 of 0.9624, MSE of 0.0003, RMSE of 0.0173 and ARE of 0.0036. Also, $$S_{{\text{w}}}$$ resulted in R, R2, MSE, RMSE and ARE values of 0.9631, 0.9276, 0.0049, 0.0700 and 0.0158, respectively. Furthermore, comparing the performance of the developed ANN model with some existing ones indicated that it performed better than some existing models. Additionally, the developed ANN model is replicable as its threshold weights and biases required to replicate the model are made available, unlike other models in the literature. Furthermore, average contribution factor analysis of the input variables depicted that resistivity log, density log, gamma-ray log and depth interval had the factor of 44.23%, 22.68%, 20.16% and 12.98%, respectively, on the developed ANN model. Hence, the developed ANN model is a more robust tool to predict reservoir’s porosity, permeability and water saturation in the Niger Delta region.

Verification of a Nondestructive Method for Assessing the Humidity of Saline Brick Walls in Historical Buildings

The paper presents the results of the verification of the neural method for assessing the humidity of saline brick walls. The method was previously developed by the authors and can be useful for the nondestructive assessment of the humidity of walls in historic buildings when destructive intervention during testing is not possible due to conservation restrictions. However, before being implemented in construction practice, this method requires validation by verification on other historic buildings, which to date has not been done. The paper presents the results of such verification, which has never been carried out before, and thus extends the scope of knowledge related to the issue. For experimental verification of the artificial neural network (ANN), the results of moisture tests of two selected historic buildings, other than those used for ANN learning and testing processes, were used. An artificial unidirectional multilayer neural network with backward error propagation and the algorithm for learning conjugate gradient (CG) was found to be useful for this purpose. The obtained satisfactory value of the linear correlation coefficient R of 0.807 and low average absolute error |Δf| of 1.16% confirms this statement. The values of average relative error |RE| of 19.02%, which were obtained in this research, were not very high for an in-situ study. Moreover, the relative error values |RE| were mostly in the range of 15% to 25%.

Support vector regression: A novel soft computing technique for predicting the removal of cadmium from wastewater

The presence of toxic heavy metals in the wastewater coming from industries is of great concern across the world. In the present work, a novel soft computing technique support vector regression (SVR)technique has been used to predict the removal of cadmium ions from wastewater with agricultural waste ‘rice polish’ as a low-cost adsorbent, with contact time, initial adsorbate concentration, p H of the medium, and temperature as the independent parameters. The developed SVR-based model has been compared with the widely used multiple regression (MR) model based on the statistical parameters such as coefficient of determination (R 2 ), average relative error (AARE) etc. The prediction performance of SVR-based model has been found to be more accurate and generalized in comparison to MR model with low AARE values of 0.67% and high R 2 values of 0.9997 while MR model gives an AARE value of 29.27% and 0.2161 as coefficient of determination (R 2 ). Furthermore, it has also been observed that the SVR model effectively predicts the behavior of the complex interaction process of cadmium ions removal from waste water under various experimental conditions.

Three-Dimensional Morphological Measurement Method for a Fruit Tree Canopy Based on Kinect Sensor Self-Calibration

Perception of the fruit tree canopy is a vital technology for the intelligent control of a modern standardized orchard. Due to the complex three-dimensional (3D) structure of the fruit tree canopy, morphological parameters extracted from two-dimensional (2D) or single-perspective 3D images are not comprehensive enough. Three-dimensional information from different perspectives must be combined in order to perceive the canopy information efficiently and accurately in complex orchard field environment. The algorithms used for the registration and fusion of data from different perspectives and the subsequent extraction of fruit tree canopy related parameters are the keys to the problem. This study proposed a 3D morphological measurement method for a fruit tree canopy based on Kinect sensor self-calibration, including 3D point cloud generation, point cloud registration and canopy information extraction of apple tree canopy. Using 32 apple trees (Yanfu 3 variety) morphological parameters of the height (H), maximum canopy width (W) and canopy thickness (D) were calculated. The accuracy and applicability of this method for extraction of morphological parameters were statistically analyzed. The results showed that, on both sides of the fruit trees, the average relative error (ARE) values of the morphological parameters including the fruit tree height (H), maximum tree width (W) and canopy thickness (D) between the calculated values and measured values were 3.8%, 12.7% and 5.0%, respectively, under the V1 mode; the ARE values under the V2 mode were 3.3%, 9.5% and 4.9%, respectively; and the ARE values under the V1 and V2 merged mode were 2.5%, 3.6% and 3.2%, respectively. The measurement accuracy of the tree width (W) under the double visual angle mode had a significant advantage over that under the single visual angle mode. The 3D point cloud reconstruction method based on Kinect self-calibration proposed in this study has high precision and stable performance, and the auxiliary calibration objects are readily portable and easy to install. It can be applied to different experimental scenes to extract 3D information of fruit tree canopies and has important implications to achieve the intelligent control of standardized orchards.

Adsorption of a non-steroidal anti-inflammatory drug onto MgAl/LDH-activated carbon composite – Experimental investigation and statistical physics modeling

A Layered Double Hydroxides (LDH) with Mg and Al composite (MgAl/LDH-biochar) was synthesised by the co-precipitation method, using bovine bone activated carbon as support structure for the LDH particles. The adsorbent was applied for diclofenac sodium (DS) removal from synthetic water. Batch experiments provided a DS removal of almost 80 % under optimized conditions. The performances of the MgAl/LDH-biochar composite (i.e. adsorption capacity as much as 5.96 mg/g) result significantly better when compared to the raw biochar (2.35 mg/g) at first 15 min. Despites the equilibrium data provided good adjustments for all tested models, Redlich-Peterson model showed the highest values of R2 and lower values of average relative error (ARE) and Akaike’s information criterion (AIC). The kinetic data of DS adsorption was best represented by the pseudo-second order model. The regenerability of MgAl/LDH-biochar was also evaluated through six adsorption/desorption cycles using either methanol or sodium chloride (0.5 mol/L) as desorbing agents, showing that the material can be reused until the third cycle maintaining a good adsorption capacity. The thermodynamic studies suggested that the adsorption process was spontaneous and exothermic, since the adsorption capacity decreased when the temperature increased. Finally, for a deeper insight in the DS adsorption mechanism, monolayer model formulated via statistical physics was adopted to describe the adsorption geometry of this relvant water pollutant and also to further characterize the interactions between DS and the tested adsorbent.

Real Estate Evaluation Model Based on Genetic Algorithm Optimized Neural Network

With the rapid development of society, the real estate economy, as an important part of Chinese economy, is showing a growing trend. But it is also the most likely to generate bubble economy, causing financial risks; it will trigger a series of social contradictions and cause social unrest in severe cases. Therefore, it is urgent to improve and optimize the real estate evaluation model. In this study, the real estate was evaluated based on the neural network model optimized by genetic algorithm. Through sorting out and summarizing the real estate data in a period of time, the corresponding model was established and the test data were obtained. The average relative error value of the genetic algorithm optimized neural network model was 3.552, which was smaller than that of the Back-Propagation (BP) neural network prediction model. The experimental conclusion that the new network model was better than the traditional model was obtained. This work opens up a new route of real estate evaluation.

Lake level dynamics exploration using deep learning, artificial neural network, and multiple linear regression techniques

Estimating the lake level dynamics accurately on a daily or finer timescale is important for a better understanding of ecosystems, especially the lakes in Badain Jaran Desert, China. In this study, lake level dynamics of Sumu Barun Jaran are simulated and predicted on a 2-h timescale using the deep learning (DL) model, which is structured for the first time in this area by considering critical environmental factors. Two machine learning methods, namely multiple linear regression (MLR) and the three-layered back-propagation artificial neural network (ANN), are also adopted for the prediction purpose. The performances of these models are evaluated by comparing the values of average relative error, the mean squared error, and the coefficient of determination. The result shows that the DL model performs better than MLR and ANN on these three criteria, and this DL model is beneficial for exploring the mechanism of lake level dynamics in Badain Jaran Desert.

Forecasting of Surface Ozone Concentration by Using Artificial Neural Networks in Rural and Urban Areas in Central Poland

This paper presents the development of artificial neural network models for the prediction of the daily maximum hourly mean of surface ozone concentration for the next day at rural and urban locations in central Poland. The models were generated with six input variables: forecasted basic meteorological parameters and the maximum O3 concentration recorded on the previous day and number of the month. The training data set covered the period from April 2015 to September 2015. An analogous data set of input variables, for the 2014 year, not used during the process of training the networks, was used as test data to examine the quality of these models. From the results of simulations for the year 2014, the average relative error values were equal to 15.3% and 15.7% for Belsk and Warsaw stations, respectively. The mean error (ME) value indicates a tendency to overestimate the predicted values by 4.8 µg/m3 for Belsk station and to underestimate the predicted values by 0.9 µg/m3 for Warsaw station. The analysis of days when the relative error value was >50% revealed that all predictions with extremely high relative error value were associated with relatively low daily maximum surface ozone concentration values that occurred suddenly due to a sharp drop in day-to-day ozone concentration values.

Preparation and characterization of MWCNT\u2013COOH\u2013cellulose\u2013MgO NP nanocomposite as adsorbent for removal of methylene blue from aqueous solutions: isotherm, thermodynamic and kinetic studies

In this work, MWCNT–COOH–cellulose nanocomposite was prepared with MWCNT–COOH, SOCl2 as leaving group and cellulose as an adsorbent (A1); MgO nanoparticles have been successfully coated on the surface of nanocomposite by directly adding to A1 (MWCNT–COOH–cellulose–MgO; A2), magnesium nitrate was added to A1 and MgO nanoparticles were achieved by coating indirectly on the surfaces (MWCNT–COOH–cellulose–MgO; A3). These nanocomposites (A1, A2, and A3) were used for the removal of methylene blue (MB) dye from aqueous solution. For characterization of adsorbent surfaces, the FT-IR, TEM, SEM, and XRD analysis were used. The effects of initial concentration of MB dye, contact time, and temperature on the adsorption were studied. According to the results, 55 min was selected as the optimum contact time for the removal process. The equilibrium data of adsorption were well fitted and the Langmuir (type III) model had the best agreement because it possesses high value of linear regression, high value R2, and least value of average relative error, ARE (%). Parameters of thermodynamics including enthalpy (∆H°), entropy (∆S°), and Gibbs energy (∆G°) were calculated. Kinetic data had best agreement with pseudo-second-order model.Graphical abstractSchematic synthesis of MWCNT–COOH–cellulose–MgO NP nanocomposite as adsorbent

Quantity comprehension of optical hysteresis loop

A theoretical model for quantity comprehension optical hysteresis loop is proposed, the model is derived from solution to decay rate differential equation together with boundary conditions. Because there are only four parameters in the model, it is easy to apply it. The model is verified by comparison with experimental data of vanadium oxides materials, the predicted results have an excellent consistency with experimental results. The maximum value of average relative error is 7.19% between measurements and theory, and the minimum data for the correlation coefficient in 18 fitted transmission-temperature curves is 0.994. Besides, it should be noteworthy that the calculation of hysteresis width is also very easy via parameters in this model. To the best knowledge of author, this is the first systemic theory to explain optical hysteresis loop.

Towards an ensemble based system for predicting the number of software faults

Software fault prediction using different techniques has been done by various researchers previously. It is observed that the performance of these techniques varied from dataset to dataset, which make them inconsistent for fault prediction in the unknown software project. On the other hand, use of ensemble method for software fault prediction can be very effective, as it takes the advantage of different techniques for the given dataset to come up with better prediction results compared to individual technique. Many works are available on binary class software fault prediction (faulty or non-faulty prediction) using ensemble methods, but the use of ensemble methods for the prediction of number of faults has not been explored so far. The objective of this work is to present a system using the ensemble of various learning techniques for predicting the number of faults in given software modules. We present a heterogeneous ensemble method for the prediction of number of faults and use a linear combination rule and a non-linear combination rule based approaches for the ensemble. The study is designed and conducted for different software fault datasets accumulated from the publicly available data repositories. The results indicate that the presented system predicted number of faults with higher accuracy. The results are consistent across all the datasets. We also use prediction at level l (Pred(l)), and measure of completeness to evaluate the results. Pred(l) shows the number of modules in a dataset for which average relative error value is less than or equal to a threshold value l. The results of prediction at level l analysis and measure of completeness analysis have also confirmed the effectiveness of the presented system for the prediction of number of faults. Compared to the single fault prediction technique, ensemble methods produced improved performance for the prediction of number of software faults. Main impact of this work is to allow better utilization of testing resources helping in early and quick identification of most of the faults in the software system.

Co-modified MCM-41 as an effective adsorbent for levofloxacin removal from aqueous solution: optimization of process parameters, isotherm, and thermodynamic studies.

Antibiotics are emerging contaminants due to their potential risks to human health and ecosystems. Poor biodegradability makes it necessary to develop effective physical-chemical methods to eliminate these contaminants from water. The cobalt-modified MCM-41 was prepared by a one-pot hydrothermal method and characterized by SAXRD, N2 adsorption-desorption, SEM, UV-Vis DR, and FTIR spectroscopy. The results revealed that the prepared 3% Co-MCM-41 possessed mesoporous structure with BET surface areas at around 898.5m2g-1. The adsorption performance of 3% Co-MCM-41 toward levofloxacin (LVF) was investigated by batch experiments. The adsorption of LVF on 3% Co-MCM-41 was very fast and reached equilibrium within 2h. The adsorption kinetics followed the pseudo-second-order kinetic model with the second-order rate constants in the range of 0.00198-0.00391gmg-1min-1. The adsorption isotherms could be well represented by the Langmuir, Freundlich, and Dubinin-Radushkevich (D-R) isotherm equations. Nevertheless, D-R isotherm provided the best fit based on the coefficient of determination and average relative error values. The mean free energy of adsorption (E) calculated from D-R model was about 11kJmol-1, indicating that the adsorption was mainly governed by a chemisorption process. Moreover, the adsorption capacity was investigated as a function of pH, adsorbent dosage, LVF concentration, and temperature with help of respond surface methodology (RSM). A quadratic model was established, and an optimal condition was obtained as follows: pH8.5, adsorbent dosage of 1gL-1, initial LVF concentration of 119.8mgL-1, and temperature of 31.6°C. Under the optimal condition, the adsorption capacity of 3% Co-MCM-41 to LVF could reach about 108.1mgg-1. The solution pH, adsorbent dosage, LVF concentration, and a combination of adsorbent dose and LVF concentration were significant factors affecting the adsorption process. The adsorption thermodynamic functions were also determined. The negative ΔH 0 (-33.50kJmol-1) and ΔS 0 (-43.57Jmol-1K-1) suggested that the adsorption was an exothermic process accompanied by decreasing disorder. This study may indicate that 3% Co-MCM-41 is a promising adsorbent for removing emerging pollutants of LVF from water.

ADSORPTION OF BASIC YELLOW 28 (BY 28) AND ACID YELLOW 23 (AY 23) DYES ONTO CHITIN

The present study investigated the sorption of Basic Yellow 28 (BY 28) and Acid Yellow 23 (AY 23) by chitin flakes. The study determined the influence of pH value on adsorption effectiveness and the adsorption capacity of chitin flakes. The results were described with Freundlich, Langmuir, Sips and double Langmuir isotherms. Similar values of adsorption capacities were achieved for both tested dyes using Langmuir, Sips and Langmuir2 models, i.e. 16.804, 17.740 and 18687 mg/g d.m. for BY28 as well as 24.195, 27.930 and 24.196 mg/g d.m. for AY23, respectively. The isotherms were compared with the use of average relative error (ARE) of approximation. In the case of both dyes, the best fit to experimental data was achieved with the use of tri-parametric Sips equilibrium isotherm, which was indicated by ARE values of 3.10% (BY 28) and 5.26% (AY 23).

木兰围场典型落叶松-杨桦混交林生物量及固碳能力研究

A mixed Dahurian larch and polar birch forest,artificial Larix principis-rupprechtii forest,natural Betula platyphylla secondary forest,and natural Populus davidiana secondary forest,which are all part of the Bei-gou Forest Farm in Mulan Forestry Bureau were studied here. The layered cut method was used to measure aboveground biomass in all four forests. The layered mining method was used to measure underground biomass. The trunks were divided into segments 0. 5m in diameter and 5 cm thick. The north-south diameter of each disk was recorded. Tree roots were divided into 0. 5 cm layers,and the crown projection area was dug up. Fresh weights of the trunks( with bark),branches,leaves,and roots of model trees were recorded in the open air. Samples of each organ were dried to( 105℃) constant weight and weighed. To determine biomass,leaf and branch samples were collected from different parts of the crown cover. They were weighed and dried,and their moisture content was calculated. For trunks,disk sections were obtained and weighed before and after drying. Moisture content was calculated. The leaf,branch,and root biomass of each model tree was calculated. Incrementswere measured in analytic trees. Then models of biomass and increments were established. Carbon reserve and carbon sequestration were estimated. Significant power functions and logistic equations were chosen as biomass prediction models.DBH( diameter at breast height) served as an independent variable and the napierian logarithm of biomass and total biomass of organs as the dependent variable. Optimization was performed according to regression curve and residual sum of squares( SSE),discriminant coefficient( R2),total relative error( RS),average relative error( E1),absolute value of average relative error( E2),AIC( Akaike Information Criterion),and BIC( Bayesian Information Criterion). DBH increments were measured with analytic trees,wooden cores,and DBH increments of the past five years as a baseline. A power function regression model was established. Then the DBH increments of trees over the past five years were estimated.Then the fixed carbon content of each forest was calculated for that period using the carbon content of different tree species and organs. Results showed the Dahurian larch and polar birch mixed forest stored more carbon than other forests. The Dahurian larch,Betula platyphylla,and Populus davidiana in the mixed forest were all superior to their respective artificial and natural forests. The mean DBH( diameter at breast height) was higher by about 6. 7%,12. 8%,and 4. 1%. The mean was higher by about 12. 1%,1. 4%,and 11. 1%. The fixed carbon content growth rates of Dahurian larch,Betula platyphylla,and Populus davidiana in the mixed forest were 29. 74%,28. 36%,and 34. 52%,respectively. The rate of carbon fixation was 27. 09% in the artificial Larix principis-rupprechtii forest; 26. 34% in the natural Betula platyphylla secondary forest; and 26. 24% in the natural Populus davidiana secondary forest. The rates of carbon fixation of Dahurian larch,Betula platyphylla,and Populus davidiana in the Dahurian larch and polar-birch mixed forest were 2. 65%,2.02%,and 8. 28% higher,respectively,than in the corresponding monoculture forests.

Forecasting Impact Injuries of Unrestrained Occupants in Railway Vehicle Passenger Compartments

Objective: In order to predict the injury parameters of the occupants corresponding to different experimental parameters and to determine impact injury indices conveniently and efficiently, a model forecasting occupant impact injury was established in this work.Methods: The work was based on finite experimental observation values obtained by numerical simulation. First, the various factors influencing the impact injuries caused by the interaction between unrestrained occupants and the compartment's internal structures were collated and the most vulnerable regions of the occupant's body were analyzed. Then, the forecast model was set up based on a genetic algorithm–back propagation (GA-BP) hybrid algorithm, which unified the individual characteristics of the back propagation–artificial neural network (BP-ANN) model and the genetic algorithm (GA). The model was well suited to studies of occupant impact injuries and allowed multiple-parameter forecasts of the occupant impact injuries to be realized assuming values for various influencing factors. Finally, the forecast results for three types of secondary collision were analyzed using forecasting accuracy evaluation methods.Results: All of the results showed the ideal accuracy of the forecast model. When an occupant faced a table, the relative errors between the predicted and experimental values of the respective injury parameters were kept within ±6.0 percent and the average relative error (ARE) values did not exceed 3.0 percent. When an occupant faced a seat, the relative errors between the predicted and experimental values of the respective injury parameters were kept within ±5.2 percent and the ARE values did not exceed 3.1 percent. When the occupant faced another occupant, the relative errors between the predicted and experimental values of the respective injury parameters were kept within ±6.3 percent and the ARE values did not exceed 3.8 percent.Conclusions: The injury forecast model established in this article reduced repeat experiment times and improved the design efficiency of the internal compartment's structure parameters, and it provided a new way for assessing the safety performance of the interior structural parameters in existing, and newly designed, railway vehicle compartments.