Analysis of specific discharge rate functions in industrial scale cement grinding multi-compartment ball mills to assist ball mill modeling

Specific discharge rates in industrial scale multi-compartment cement grinding ball mills were determined using industrial scale data to assist modeling of multi-compartment ball mills. Multi-compartment ball mill model structures were proposed to estimate normalized specific discharge rate functions. Multi-compartment mills were divided into perfectly mixed mill segments to describe material transport mechanism in the mill. Specific discharge rate functions were determined by using the estimated mill hold-up for perfectly mixed mill segments at the steady state condition of the cement grinding circuits. Estimated normalized discharge rate functions were found to be different than the basic pattern observed in semi-autogenous mills due to the differences in design and operational conditions of the mills. Coarse particle accumulation was found to have a significant effect on discharge rate function pattern. Estimated normalized discharge rate functions could be used in modeling of multi-compartment ball mills.     

Simultaneous enhancement of the efficiency and stability of organic solar cells using PEDOT:PSS grafted with a PEGME buffer layer

We report a simple processing method to simultaneously improve the efficiency and stability of organic solar cells (OSCs). Poly(4-styrene sulfonate)-doped poly(3,4-ethylenedioxy-thiophene (PEDOT:PSS), widely used as hole transport layer (HTL) in OSCs, tends to accelerate the degradation of devices because of its hygroscopic and acidic properties. In this regard, we have modified PEDOT:PSS to reduce its hygroscopic and acidic properties through a condensation reaction between PEDOT:PSS and poly(ethylene glycol) methyl ether (PEGME) in order to improve the efficiency and stability of OSCs. As a result, the power conversion efficiency (PCE) increased by 21%, from 2.57% up to 3.11%. A better energy level alignment by the reduced work function of the modified PEDOT:PSS with a highest occupied molecular orbital (HOMO) level of poly(3-hexylthiophene-2,5-diyl) (P3HT) is considered the origin of the improved the efficiency. The half-life of OSCs with PEDOT:PSS modified with PEGME buffer layer also increased up to 3.5 times compared to that of devices with pristine PEDOT:PSS buffer layer.     

Deterministic and probabilistic ship pitch prediction using a multi-predictor integration model based on hybrid data preprocessing,reinforcement learning and improved QRNN

The deterministic and probabilistic prediction of ship motion is important for safe navigation and stable real-time operational control of ships at sea. However, the volatility and randomness of ship motion, the non-adaptive nature of single predictors and the poor coverage of quantile regression pose serious challenges to uncertainty prediction, making research in this field limited. In this paper, a multi-predictor integration model based on hybrid data preprocessing, reinforcement learning and improved quantile regression neural network (QRNN) is proposed to explore the deterministic and probabilistic prediction of ship pitch motion. To validate the performance of the proposed multi-predictor integrated prediction model, an experimental study is conducted with three sets of actual ship longitudinal motions during sea trials in the South China Sea. The experimental results indicate that the root mean square errors (RMSEs) of the proposed model of deterministic prediction are 0.0254°, 0.0359°, and 0.0188°, respectively. Taking series #2 as an example, the prediction interval coverage probabilities (PICPs) of the proposed model of probability predictions at 90%, 95%, and 99% confidence levels (CLs) are 0.9400, 0.9800, and 1.0000, respectively. This study signifies that the proposed model can provide trusted deterministic predictions and can effectively quantify the uncertainty of ship pitch motion, which has the potential to provide practical support for ship early warning systems.     

Experimental study on vibration-induced clamping force reduction in fuel cell structures

When applied to transportation systems, fuel cell structures are exposed to external mechanical disturbance including shocks and harmonic excitations from operating components. To minimize performance degradation from machine operations, the fuel cell structure needs to be examined via vibration reliability tests. In this study, the reduction in the clamping force of the stack by random vibrations was investigated by experiments. The stack mass and gasket were clamped with bolts for vibration tests. The vibration induced shear movements between clamped stacks. To estimate the vibration input magnitudes, the Dirlik method was used. The reduction in the stack clamping force was estimated using the Basquin's power law. The clamping force decreased by the shear vibration input to the stack structure. The degree of clamping force reduction was larger for the heavier stack. When the stacks were separated by the gasket the reduction became smaller. Through the Dirlik method, the vibration reliability of the stack was evaluated. This information provides severity of the external vibration on the stack functionality.     

Failure load prediction and optimisation for adhesively bonded joints enabled by deep learning and fruit fly optimisation

Adhesively bonded joints have been extensively employed in the aeronautical and automotive industries to join thin-layer materials for developing lightweight components. To strengthen the structural integrity of joints, it is critical to estimate and improve joint failure loads effectually. To accomplish the aforementioned purpose, this paper presents a novel deep neural network (DNN) model-enabled approach, and a single lap joint (SLJ) design is used to support research development and validation. The approach is innovative in the following aspects: (i) the DNN model is reinforced with a transfer learning (TL) mechanism to realise an adaptive prediction on a new SLJ design, and the requirement to re-create new training samples and re-train the DNN model from scratch for the design can be alleviated; (ii) a fruit fly optimisation (FFO) algorithm featured with the parallel computing capability is incorporated into the approach to efficiently optimise joint parameters based on joint failure load predictions. Case studies were developed to validate the effectiveness of the approach. Experimental results demonstrate that, with this approach, the number of datasets and the computational time required to re-train the DNN model for a new SLJ design were significantly reduced by 92.00% and 99.57% respectively, and the joint failure load was substantially increased by 9.96%.     

Efficiency,mechanism and microbial community of Cd(II) removal by mixed bacteria enriched from heavy metals mine soil

Mixed bacteria were enriched from heavy metals mine soil for cadmium (Cd(II))-containing wastewater treatment. Batch adsorption experiment results showed that the optimal pH, temperature, initial Cd(II) concentration, and biomass dosage were 6.0, 30 °C, 20 mg/L, and 1 g/L, respectively. Living biomass exhibited better Cd(II) removal efficiency (91.97%) than autoclaved biomass (79.54%) under optimal conditions. The isotherms and kinetics of living biomass conformed to the Langmuir isotherm model and pseudo-first-order kinetic model, respectively. FTIR results implied that amine groups, hydroxyl groups and phosphoric acid play an important role in the Cd(II) adsorption process, while XRD results showed that crystalline Cd(OH)₂ and CdO were obtained. After Cd(II)-containing wastewater treatment exposure, the dominant bacteria genera included Comamonas (39.94%), unclassified_f__Enterobacteriaceae (34.96%), Ochrobactrum (14.07%), Alcaligenes (4.84%), Bordetella (2.07%), Serratia (1.04%), and Bacillus (1.01%). Function prediction showed that the abundance of metabolic genes changed significantly. This study proposes the potential application of mixed bacteria for Cd(II)-containing wastewater treatment.     

Correlation between motion of source bender element and feedback signal induced by self-monitoring

Since the introduction of bender element tests to soil testing, the reliability of the estimated travel time has been the most serious problem. The author has previously shown a potential solution whereby removing the response of the bender element subsystem from the whole response could dramatically improve the accuracy of the travel time estimation. In order to lay the foundation for estimating the response of the bender element subsystem, this paper examines the correlation between the displacements of the element and the induced feedback signals by employing a self-monitoring element. The response of the self-monitoring element is modeled as a transfer function involving two internal transfer functions that relate the input signals to the displacements and the displacements to the feedback signals, respectively. Using a laser displacement sensor, the displacements are directly measured through the entire surface and reveal the three-dimensional bending motion of the element oscillating in both longitudinal and width directions. The feedback signals are similar to, but inconsistent with, the tip displacements, and an attempt is made to correct the feedback signals. Finally, a conclusion is given on the potential for estimating the response of the bender element subsystem using the self-monitoring element.     

基于BERT-Att-biLSTM模型的医学信息分类研究

目前,对互联网上虚假健康信息的研究多集中于谣言识别,而对医学信息自动分类的研究较少。采用基于双向编码的语言表征模型和注意力增强的双向长短时记忆模型(BERT-Att-BiLSTM模型),对健康信息文本进行分类,实现自动识别虚假健康信息。实验结果表明,BERT-Att-BiLSTM模型可以高效地对医学信息进行分类,其中BERT模型相较于BiLSTM模型,性能提升明显;与融合Word2Vec的BiLSTM模型相比,BERT-Att-BiLSTM模型效果更佳。     

面向可解释性的物体拓扑结构骨架表征方法

目的 模式识别中，通常使用大量标注数据和有效的机器学习算法训练分类器应对不确定性问题。然而，这一过程缺乏知识表征和可解释性。认知心理学和实验心理学的研究表明，人类往往不使用代价如此巨大的机制，而是使用表征、归纳、推理、解释和约束传播等与符号主义人工智能方法类似的手段来应对物体识别中的不确定性并提供可解释性。因此本文旨在从传统的符号计算出发，利用骨架拓扑结构表征提供一种可解释性的思路。方法 以骨架树为基本手段来形成物体拓扑结构特征和几何特征的形式化表征，并基于泛化框架对少量同类表征进行知识抽取来形成关于物体类别的知识概括显式化表征。结果 在形成物体类别的概括表征实验中，通过路径重建直观展示了同类属物体上得到的最一般表征的几何物理意义。在可解释性验证实验中，通过跨数据的拓扑应用展示了新测试样本相对于概括表征的特定差异，表明该表征具有良好的可解释性。最后在形状补全的不确定性推理实验中，不仅可以得到识别结论，而且清晰展示了识别背后做出的判断依据，进一步验证了该表征的可解释性。结论 实验表明一般化的形式表征能够应对尺寸、颜色和形状等不确定性问题，本文方法避免了基于纹理特征所带来的不确定性，适用于任意基于基元的表征方式，具有更好的鲁棒性、普适性和可解释性，计算代价更小。