以化学改性脱硫灰取代部分炭黑与丁苯橡胶复合制备化学改性脱硫灰/丁苯橡胶复合材料，运用均匀设计结合BP神经网络建立制备工艺参数与力学性能的BP神经网络优化模型，获得制备最优化学改性脱硫灰/丁苯橡胶复合材料的工艺参数，对最优复合材料进行了表征. 结果表明，最优化学改性脱硫灰/丁苯橡胶复合材料的制备参数为促进剂N-叔丁基-2-苯并噻唑次磺酰胺用量1.2 g、硫磺用量1.3 g、硬脂酸用量1.1 g、氧化锌用量2.4 g、硫化时间27 min，该条件下所制复合材料的拉伸强度为20.31 MPa、撕裂强度为45.68 kN/m、邵氏A硬度为66，最优实测值与最优模型预测值吻合较好，相对误差为3.03%~3.22%.
Chemical modified desulfurization ash/styrene butadiene rubber composite materials were prepared using chemical modified desulfurization ash replacing partial carbon black，BP neural network optimization model of preparation process parameters (amounts of accelerator, sulfur, amount of stearic acid, and zinc oxide and curing time) and mechanical properties were established by uniform design in a combination with BP neural network to acquire parameter of optimal chemical modified desulfurization ash/styrene butadiene rubber composite materials. Tensile properties of chemical modified desulfurization ash/styrene butadiene rubber composite materials were tested by referring to rubber, vulcanized or thermoplastic-determination of tensile stress-strain properties, tear strength of chemical modified desulfurization ash/styrene butadiene rubber composite materials were tested by referring to rubber, vulcanized or thermoplastic-determination of tear strength, hardness of chemical modified desulfurization ash/styrene butadiene rubber composite materials was tested by referring to rubber, vulcanized or thermoplastic-determination of indentation hardness, Duromerer method (Shore hardness) (trouser, angle and crecent test pieces). Microstructure of optimal chemical modified desulfurization ash/styrene butadiene rubber composite materials were observed by SEM, mineral composition of optimal chemical modified desulfurization ash/styrene butadiene rubber composite materials were analyzed by XRD. The results showed that the parameters of optimal chemical modified desulfurization ash/styrene butadiene rubber composite materials were amount of accelerator 1.2 g, amount of sulfur 1.3 g, amount of stearic acid 1.1 g, amount of zinc oxide 2.4 g and curing time 27 min. Mechanical properties of optimal chemical modified desulfurization ash/styrene butadiene rubber composite materials were tensile strength 20.31 MPa, tear strength 45.68 kN/m and Shore A hardness 66. The optimal measured values and the model optimal predictive values were in good agreement, the relative error was 3.03%?3.22%. The above research provides technical support and theoretical basis for the development of cheap inorganic filler which can replace partial carbon black.
[1] You C F, Li Y. Desulfurization Characteristics of Rapidly Hydrated Sorbents with Various Adhesive Carrier Particles for a Semidry CFB-FGD System [J]. Environmental Science & Technology, 2013, 47(6): 2754?2759.
[2] Zhou Y G, Peng J, Zhu X, et al. Hydrodynamics of Gas?Solid Flow in the Circulating Fluidized Bed Reactor for Dry Flue Gas Desulfurization [J]. Powder Technology, 2010, 205(1): 208?216.
[3] Zhuang P, Mcbride M B, Xia H, et al. Health Risk From Heavy Metals via Consumption of Food Crops in the Vicinity of Dabaoshan Mine [J]. Science of The Total Environment, 2009, 407(5): 1551?1561
[4] 杨晋涛，范宏，卜志扬，等. 蒙脱土填充补强丁苯橡胶及对橡胶硫化特性的影响 [J]. 复合材料学报, 2005, 22(2):38?45
Yang J T, Fan H, Bu Z Y, et al. Montmorillonite Reinforced SBR and Effect on the Vulcanization of Rubber [J]. Acta Materiae Compositae Sinica, 2005, 22(2): 38?45.
[5] Raza M A, Westwood A, Brown A, et al. Characterization of Graphite Nano Platets and the Physical Properties of Graphite Nano Platelets/Silicone Composite for Thermal Interface Applications [J]. Carbon, 2011, 49(13): 4269?4279
[6] Chen L, Lu L, Wu D J, et al. Slicone Rubber/Graphite Nanosheet Electrically Conducting Nano Composite with a Low Percolation Threshold [J]. Polymer Composites, 2007, 28(4): 493?498
[7] 张浩，许谨，曹现雷. 基于BP神经网络优化制备Cu?Ce/TiO2及其光催化活性研究 [J]. 环境科学学报, 2015, 35(8): 2450?2456.
Zhang H, Xu J, Cao X L. Optimizing Preparation and Photocatalytic Activity of Cu?Ce/TiO2 Bsed on BP Neural Network [J]. Acta Scientise Circumstantise, 2015, 35(8): 2450?2456.
[8] Amjady N, Keynia F. A New Neural Network Approach to Short Term Load Forecasting of Electrical Power Systems [J]. Energies, 2011, 4(3): 488?503.
[9] Nawi N M, Khan A, Rehman M Z. A New Levenberg Marquardt Based Back Propagation Algorithm Trained with Cuckoo Search [J]. Powder Technology, 2013, 11(1): 18?23.
[10] 孙晓燕. RBF神经网络在入侵检测中的应用研究 [D]. 合肥：合肥工业大学, 2009: 34?36.
Sun X Y. Application and Research of RBF Neural Network in Intrusion Detection [D]. Hefei: Hefei University of Technology, 2009: 34?36.
[11] 韩敏，王亚楠. 求解非线性回归问题的Newton算法 [J]. 计算机学报, 2010, 33(5): 841?846.
Han M, Wang Y N. A Newton Algorithm for Nonlinear Regression [J]. Chinese Journal of Computers, 2010, 33(5): 841?846.
[12] 杨铎. 基于Gauss?Newton法的空间管形拟合算法的研究 [J]. 大连大学学报, 2014. 35(3): 19?23.
Yang D. Space Tube-shaped Fitting Slgorithm Based on Spatial Gauss?Newton Method [J]. Journal of Dalian University, 2014, 35(3): 19?23.
[13] 张浩，张欣雨，刘秀玉. 化学改性脱硫灰取代部分炭黑制备环保型丁苯橡胶及性能 [J]. 过程工程学报, 录用
Zhang H, Zhang X Y, Liu X Y. Preparation of Environmentally Friendly Styrene Butadiene Rubber Using Chemical Modified Desulfurization Ash Replacing Partial Carbon Black and Its Properties [J]. Chin. J. Process Eng., Accept
