基于灰色BP神经网络的水泥强度预测模型研究

水泥强度的预测具有多变量、非线性和大时滞特性,因此传统线性回归方法的结果不准确。除此之外,传统的神经网络预测可能对少量样本不够精确。本文建立灰色BP模型,以此来预测水泥的强度。建立一个多因素灰色模型GM(1,N)用于水泥化学成分的样本数据进行预处理,得到新的数据来作为建立预测模型的样本数据,通过BP神经网络建立预测模型。最终通过建立的灰色BP神经网络预测模型来预测28天水泥强度。仿真结果表明:灰色BP预测模型的效果比BP预测的要准确。     

黄霉素发酵过程的灰色关联度分析及其灰色预测模型

将灰色理论应用于黄霉素摇瓶发酵过程。对黄霉素摇瓶发酵过程中的各控制因素进行了灰色关联度分析,并建立了GM(1,7)灰色模型。结果表明不同参数对摇瓶发酵过程中黄霉素效价的影响程度是不同的,其影响顺序为:总糖>氨基氮>pH值>还原糖>种子液生物量>残油;黄霉素的发酵生产与种子液生物量和pH值呈正相关性,与总糖、还原糖、氨基氮和残油呈负相关性;GM(1,7)灰色模型的平均相对误差仅为4.804%,预测模拟的结果较好。这说明灰色理论中的灰色关联度分析和灰色预测模型可以很好的应用于黄霉素的发酵过程。     

基于时序InSAR监测的VMD-SSA-LSTM矿区地表形变预测模型研究

矿区地表形变引起的滑坡、塌陷等灾害对矿山安全生产造成了极大影响,因此对地表形变进行预测对于矿山安全开采和灾害预警预报具有重要的现实意义。采用时序InSAR技术获取矿区面域式地表形变数据,结合变分模态分解算法(VMD)对其进行分解,通过构建麻雀优化算法(SSA)与长短期记忆神经网络(LSTM)相结合的预测模型,对分解后的形变分量进行综合预测;以云南省玉溪市大红山矿区为例,利用该模型对其地表形变进行了预测,结果表明：VMD算法在一定程度上可以解决矿山地表形变数据时序特征复杂且难以直接分析判别的问题;SSA-LSTM预测模型通过智能搜索算法自动寻找预测网络模型超参数,有效减少了人为因素对模型预测效果的干扰,模型预测结果的决定系数R²均大于0.96;同时,该模型采用“多维输入”的方式训练模型,提高了训练效率;该模型在预测精度、训练效率等方面均取得了较好的效果。采用VMD-SSA-LSTM预测算法对时间序列InSAR监测下的形变数据进行分析能有效预测矿区地表形变,可为矿山安全生产管理提供参考。     

基于灰色系统的混凝土酸性化预测模型研究及应用

基于灰色系统理论、Fick定律,通过研究粉煤灰掺量、pH值对酸性化速率系数K的影响,对酸性化数据进行灰色建模,建立了综合考虑粉煤灰掺量、pH值、时间依赖性的混凝土酸性化预测模型。采用实验室加速试验方法,探讨了提高酸性水环境下混凝土结构使用寿命的有效方法。结果表明:酸性化预测模型能有效反映酸性化过程的实际状况,且所需数据量少,无需知道原始数据的先验特征;酸性水环境下混凝土表面采用水性渗透型无机防水剂、清水混凝土保护涂料进行防腐,混凝土结构寿命的增加率的效果远高于增加保护层厚度的,使混凝土结构的使用寿命有了质的提升。     

中国卫生陶瓷进出口趋势的灰色预测模型的研究

研究了中国卫生陶瓷进出口趋势的灰色预测模型。将中国卫生陶瓷1997年至2005年进出口量与进出口额数据进行分析处理和方根变换，用包络线法和新陈代谢法建立GM（1，1）灰色预测模型，该模型不仅贴近实际，还可以对未来五年中国卫生陶瓷进出口趋势进行预测。     

地表变形对框架结构的受力影响

主要对高层框架结构在常规条件下、地表变形下进行对比分析,得出建筑物主要构件在只有地表变形下内力的变化规律。同时对建筑物在地表变形作用下的基底位移、水平方向的位移进行了对比分析,得出建筑物基底位移在地表曲率变形下呈不规则的抛物线型,水平方向的位移在地表变形下值很小。     

工程开挖引起的地表移动及变形测量

本文研究了多种地矿和地下工程开挖所引致的地表移动和变形的特点。本文首先介绍地表移动预计和地表及其近邻建筑物的变形测量,通过研究与分析,掌握变形规律,从而避免工程开挖灾害的发生,判定安全程度。     

灰色-马尔可夫的改进模型及在陶瓷工业产值预测

灰色模型是经济预测方法中最有效的方法之一,但灰色模型对波动性大的序列预测时精度较差.针对这一不足,提出灰色—马尔可夫的改进模型.新模型将原始序列进行对数运算提高数据光滑度,采用距离法降低灰色模型的相对误差,马尔可夫链对预测值进一步修正.将改进模型运用到景德镇陶瓷工业产值预测中,并与传统灰色模型、灰色-马尔可夫模型预测结果进行对比.结果表明改进模型的相对误差远小于灰色模型,预测精度优于灰色-马尔可夫模型,更能反应经济发展趋势.     

基于灰色理论的露天矿生产成本分析

应用灰关联分析方法,通过实例分析了影响矿山生产成本的诸多因素,并找出影响矿山生产成本的主要因素,从而为矿山抓住主要矛盾,降低主要成本因素对总成本的影响,提高矿山经济效益提供参考依据.     

基于LSTM-RNN模型的铁水硅含量预测

针对高炉炼铁是一个动态过程,具有大延迟,工况复杂的特性。采用LSTM-RNN模型进行硅含量预测,充分发挥了其处理时间序列时挖掘前后关联信息的优势。首先根据时间序列趋势及相关系数选择自变量,并采用复杂工况的实际生产数据进行验证。然后用程序自动求解最优参数进行硅含量预测。最后将LSTM-RNN模型与PLS模型及RNN模型的结果进行对比,验证该方法的优势。研究发现LSTM-RNN模型预测误差稳定,预测精度较高,比传统的统计学及神经网络方法取得了更好的预测精度。     

露天矿事故分析与预测及安全信息管理

结合某露天矿的实际情况,利用事故致因理论和调查模式,讨论了该矿主要事故致因,用故障树(FT)对列车运输事故作了详细的分析,揭露了发生事故的主要原因,提出了防止事故的措施。运用灰色系统理论建立了权系数一阶单变量灰色预测模型,用于预测矿山事故的发生。该模型用实际数据作了验证。该数据库包括9个模块,具有完整的数据存入、修改、删除、统计、查询和报表打印等功能。该成果对其他矿山有参考性。     

Analysis and prediction of the width of spreading carbon fiber tow based on gray system theory

This article studied the effects of the speed of spreading fiber and tension of spreading fiber on the width of spreading carbon fiber tow. Based on the single tension of spreading fiber, the univariate gray prediction model GM (1,1) was established, which showed high feasibility and accuracy. Under different tensions of spreading fiber, the intrinsic correlation of the width sequence were analyzed and the multivariate gray prediction model MGM(1,n) was established to predict width of spreading fiber. By transforming the identity of multiple factors in the uncertainty system, it can overcome the limitation that multiple factors develops without keeping same pace. The analysis shows that there is an intrinsic correlation between the widths of spreading fiber tow under different tensions of spreading fiber. Based on the gray correlation degree and the adjacent sequence, the gray prediction model established has high numerical precision. There is good guiding significance for precision optimization in actual production.     

矿山地表塌陷坑综合治理

针对刘店铁矿开采现状及地表塌陷坑破坏情况,对矿山塌陷状况进行了分析,指出矿山塌陷坑亟须治理的必要性和紧迫性,并根据矿山现有技术条件,提出了塌陷坑治理的基本原则,制定了塌陷坑治理的合理方案及相关技术配套措施。     

Folding rate prediction based on neural network model

One of the most important challenges in biology is to understand the relationship between the folded structure of a protein and its primary amino acid sequence. A related and challenging task is to understand the relationship between sequences and folding rates of proteins. Previous studies found that one of contact order (CO), long-range order (LRO), and total contact distance (TCD) has a significant correlation with folding rate of protein. Although the predicted results from TCD can provide better results, the deviation is also large for some proteins. In this paper, we adopt back-propagation neural network to study the relationship between folding rate and protein structure. In our model, the input nodes are CO, LRO, and TCD, and the output node is folding rate. The number of nodes in the hidden layer is seven. Our results show that the relative errors for the predicted results are even lower than other methods in the literature. We also observe a best excellent correlation between the folding rate and contact parameters (including CO, LRO, and TCD), and find that the folding rate depends on CO, LRO and TCD simultaneously. This means that CO, LRO and TCD are similarly important in folding rate of protein. Some comparisons are made with other methods.     

Development of a porosity prediction model based on shrinkage velocity and glass transition temperature

Abstract

Pore formation and evolution is a common physical phenomenon observed in food materials during different dehydration processes. This change affects heat and mass transfer process and many quality attributes of dried product. Many mathematical models ranging from emperical to classical models proposed in the literature for predicting porosity during drying of food materials. Classical model is in its infancy as the required materials properties during drying are not avaiable for the material charecterisation. Empirical and semi-empirical models are reasonably well developed in establishing relationships between pore evolution and moisture content and determining experimental based coefficients. However, there are no simplistic models that considered process conditions and material properties together to predict the porosity. The purpose of this work is to develop a simplistic theoretical model for pore formation taking both process parameters and changing material properties during drying into consideration. A new “shrinkage velocity” approach has been introduced and the model has been developed based on this shrinkage velocity taking into account the main factors that influence the porosity including the glass transition temperature. Experimental results show good agreement with simulated results and thus validated the model. This study is expected to enhance the current understanding of pore formation of deformable materials during drying.     

Treatment of gray wastewater and heavy metal removal from aqueous medium using hydrogels based on novel crosslinkers

Water contamination by heavy metals is a serious environmental problem. Several adsorbents with high adsorption capacity have been reported by researchers. In this paper we report the application of a new crosslinker allyl sorbitol (AS), and two other crosslinkers viz. allyl mannitol (AM) and allyl pentaerythritol (AP), toward the synthesis of polymer hydrogels. These hydrogels have been utilized for the removal of heavy metal ions from water. Chemically crosslinked polyacrylic acid hydrogels have been prepared using these crosslinkers viz. AS or AM or AP. The synthesized crosslinked hydrogels have been utilized for the removal of heavy metal ions from water. The maximum copper (Cu) metal ion removal of 81.8, 81.6, 86.8% and adsorption capacity of 409, 408, 434 mg/g were found with AS, AM and AP crosslinker based hydrogels respectively. While the maximum nickel (Ni) metal ion removal of 96.8, 96.6, 99.0% and adsorption capacity of 484, 483, 495 mg/g were found with AS, AM and AP crosslinkers based hydrogels respectively. The maximum chemical oxygen demand (COD) reduction of 72, 74, 78% and total dissolved solids reduction of 86%, 83%, 95% were found with AS, AM, AP based polymer gel respectively. Approximately, 439.2, 451.4, and 475.8 (mg COD removed/g polymer gel) of pollutants reduction capacity was achieved during their treatment.     

柠檬酸钠表面改性重钙粉体的研究

研究了柠檬酸钠用量、反应温度、反应时间、浆料浓度对重钙粉体表面改性的影响。结果表明,柠檬酸钠用量为重钙粉体的6.0%(质量分数),改性温度65℃,改性时间45 min,浆料浓度为12.5%时,重钙粉体沉降体积降为0.65 mL/g,活化度可达到67.9%,吸油值降为230 mg/g,粘度值降低为120 mPa.s,pH值8.50。