Experimental study on mixed particles flow rate of dual circulating fluidized bed and prediction of kernel extreme learning machine model

doi:10.16085/j.issn.1000-6613.2018-1869

Abstract

Abstract:

The particles circulation flow rate between the two beds is the key to ensure the normal operation of the dual circulating fluidized bed. The study of relationship between circulating flow rate (G _s,mix) and various control parameters (gasification chamber gas velocity, riser gas velocity, initial bed material mass, quartz sand particle diameter and rice husk quality in the mixture) were carried out on a self-designed dual circulating fluidized bed (DCFB) experimental system，and the change of rice husk mass fraction(X _r) in circulating materials was analyzed. In addition, based on the experimental measurement data, the KELM model was established to predict the G _s,mix and the X _r of the circulating materials, and compared with the ELM model. It was found that the KELM model had smaller predicted MAPE and RMSE values, and the required time of complete prediction was shorter than ELM. This indicated that KELM model can achieve good prediction of G _s,mix and X _r under various control parameters and provided a new method for the research of DCFB system and similar gasification system.

Key words: dual circulating fluidized bed, mixed particles, circulating flow rate, control parameters, kernel extreme learning machine model

摘要：

两床间颗粒循环流率是保证双循环流化床正常运行的关键。在自行设计的双循环流化床（DCFB）实验系统上进行循环流率（G _s,mix）与各控制参数（气化室风速、提升管风速、初始床层物料量、混合物料中石英砂粒径以及稻壳质量分数）变化关系的研究，并对循环物料中稻壳质量分数（X _r）变化进行分析。此外，根据实验测量数据，建立核极限学习机（KELM）模型并进行G _s,mix和循环物料中X _r的预测，同时与极限学习机（ELM）模型进行比较，发现KELM模型具有更小的预测平均绝对值误差和均方根误差且预测所需时间较短，表明该模型可实现各控制参数下DCFB系统中G _s,mix和X _r的良好预测，为DCFB系统及类似气化系统的数据模型研究提供一种新方法。

关键词: 双循环流化床, 混合物料, 循环流率, 控制参数, 核极限学习机

CLC Number:

TK 229

Xin YANG, Zherui MA, Hongwei CHEN, Zhenghui ZHAO. Experimental study on mixed particles flow rate of dual circulating fluidized bed and prediction of kernel extreme learning machine model[J]. Chemical Industry and Engineering Progress, 2019, 38(05): 2103-2111.

杨新, 麻哲瑞, 陈鸿伟, 赵争辉. 双循环流化床混合颗粒流率实验研究与核极限学习机模型预测[J]. 化工进展, 2019, 38(05): 2103-2111.

Figures/Tables 16

References 22

1	陈鸿伟, 张志才, 李忠猛, 等 . 锥形布风板双循环流化床颗粒循环流率的研究及预测[J]. 动力工程, 2003, 23(4): 2525-2529.
	CHEN H W , ZHANG Z C , LI Z M , et al . Study and prediction of particle circulation flow rate in a double circulating fluidized bed with conical air distribution plate [J]. Power Engineering, 2003, 23(4): 2525-2529.
2	CHU C Y , HWANG S J . Attrition and sulfation of calcium sorbent and solids circulation rate in an internally circulating fluidized bed [J]. Power Technology, 2002, 127: 185-195.
3	韩超一, 吴文龙, 陶蕾, 等 . 双循环流化床提升管中气固流动特性及接触效率研究[J]. 石油炼制与化工, 2016, 47(3): 16-24.
	HAN C Y , WU W L , TAO L , et al . Study on gas-solid flow characteristics and contact efficiency in double-circulating fluidized bed riser [J]. Petroleum Processing and Petrochemicals, 2016, 47(3): 16-24.
4	耿察民, 钟文琪, 邵英娟, 等 . 双循环流化床颗粒分布特性的三维数值模拟[J]. 中国科学院大学学报, 2016, 33(2): 258-264.
	GENG C M , ZHONG W Q , SHAO Y J , et al . Three-dimensional numerical simulation of particle distribution characteristics of double-circulating fluidized bed [J]. Journal of University of Chinese Academy of Sciences, 2016, 33(2): 258-264.
5	陈鸿伟, 刘焕志, 李晓伟, 等 . 双循环流化床颗粒循环流率与BP神经网络预测[J]. 中国电机工程学报, 2010, 33(32): 25-29.
	CHEN H W , LIU H Z , LI X W , et al . Circulating flow rate of double circulating fluidized bed particles and BP neural network prediction [J]. Proceedings of the CSEE, 2010, 33(32): 25-29.
6	韩磊, 于旷世, 朱治平, 等 . 双循环流化床冷态实验研究[J]. 锅炉技术, 2011, 42(1): 26-30.
	HAN L , YU Y S , ZHU Z P , et al . Study on cold state experiment of double circulating fluidized bed [J]. Boiler Technology, 2011, 42(1): 26-30.
7	BI H , ZHU J X . Static instability analysis of circulating fluidized beds and concept of high-density risers [J]. AIChE Journal, 1998, 39(8): 1272-1280.
8	MIAO Q , JESSE Z , BARGHI S , et al . A hydrodynamic model prediction of a circulating fluidized bed with low-density particles [J]. The Canadian Journal of Chemical Engineering, 2002, 90: 1027-1032.
9	SIDDHARTHA S , BRAHIM S A , BADRUL M J , et al . Hydrodynamic properties of a cold model of dual fluidized bed gasifier: a modeling and experimental investigation [J]. Chemical Engineering Research and Design, 2016, 109: 791-805.
10	王伟, 范晓旭, 那永杰, 等 . 循环流化床多联供试验台的冷态实验研究[J]. 锅炉技术, 2006, 37(4): 40-43.
	WANG W , FAN X X , NA Y J , et al . Cold state experimental study of circulating fluidized bed multi-joint test bench [J]. Boiler Technology, 2006, 37(4): 40-43.
11	韩磊 . 双循环流化床冷态实验研究[D]. 北京: 中国科学院工程热物理研究所, 2010: 32-41.
	HAN L . Experimental study on cold state of double circulating fluidized bed [D]. Beijing: Institute of Engineering Thermophysics, Chinese Academy of Sciences, 2010: 32-41.
12	李佑楚, 陈丙瑜, 王凤鸣, 等 . 快速流态化流动[J]. 化工学报, 1979, 30(2): 143-152.
	LI Y C , CHEN B Y , WANG F M , et al . Fast fluidization flow [J]. Journal of Chemical Industry and Engineering (China), 1979, 30(2): 143-152.
13	RYU H J, LIM N Y, BAE D H, et al . Minimum fluidization velocity and transition velocity of fast fluidization of oxygen carrier particle for chemical-looping combustor [J]. Hwahak Konghak, 2003, 41(5): 624-631.
14	金涌, 祝京旭, 汪展文, 等 . 流态化工程原理[M]. 北京: 清华大学出版社, 2001: 25-63.
	JIN Y , ZHU J X , WANG Z W , et al . Principle of fluidization engineering [M]. Beijing: Tsinghua University Press, 2001: 25-63.
15	WEI F , CHENG Y , JIN Y , et al . Axial and lateral dispersion of fine particles in a binary-solid riser [J]. The Canadian Journal of Chemical Engineering, 1998, 76: 19-26.
16	HSINHONG S , CHEN Y C , SHYH J H . Solids circulation and attrition rate and gas bypassing in an internally circulating fluidized bed [J]. Industrial & Engineering Research, 2003, 42: 5915-5923.
17	PAUDEL B , FENG Z . Prediction of minimum fluidization velocity for binary mixtures of biomass and inert particles [J]. Powder Technology, 2013, 237: 134-140.
18	SHARMA M , KUMAR A , PATIL K , et al . Fluidization characteristics of a mixture of gasifier solid residues, switchgrass and inert material [J]. Powder Technology, 2013, 235: 661-668.
19	HUANG G B , ZHOU H M , DING X J , et al . Extreme learning machine for regression and multiclass classification [J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2012, 42(2): 513-529.
20	张雷, 张小刚, 陈华 . 基于Gath-Geva算法和核极限学习机的多阶段间歇过程软测量[J]. 化工学报, 2018, 69(6): 2576-2585.
	ZHANG L , ZHANG X G , CHEN H . Soft sensors for multi-stage batch processes based on Gath-Geva algorithm and kernel extreme learning machine [J]. CIESC Journal, 2018, 69(6): 2576-2585.
21	李琨, 韩莹, 佘东生, 等 . 基于IFOA-KELM-MEA模型的游梁式抽油机采油系统井下工况的短期预测[J]. 化工学报, 2017, 68(1): 188-198.
	LI K , HAN Y , SHE D S , et al . IFOA-KELM-MEA model based transient prediction on down-hole working conditions of beam pumping units [J]. CIESC Journal, 2017, 68(1): 188-198.
22	ZHANG M , LIU X G , ZHANG Z Y . A soft sensor for industrial melt index prediction based on evolutionary extreme learning machine [J]. Chinese Journal of Chemical Engineering, 2016, 24(8): 1013-1019.

种类	平均粒径 /μm	密度 /kg?m^?3	球形度	临界流化速度/m?s^?1	起始输送速度/m?s^?1
石英砂	210	2450	0.76	0.183	2.84
石英砂	330	2450	0.76	0.216	3.36
石英砂	490	2450	0.76	0.257	4.06
稻壳	900	630	0.53	0.385	3.52

种类	平均粒径 /μm	密度 /kg?m^?3	球形度	临界流化速度/m?s^?1	起始输送速度/m?s^?1
石英砂	210	2450	0.76	0.183	2.84
石英砂	330	2450	0.76	0.216	3.36
石英砂	490	2450	0.76	0.257	4.06
稻壳	900	630	0.53	0.385	3.52

模型	RMSE_Tr	MAPE_Tr/%	RMSE_Te	MAPE_Te/%	时间/s
ELM(G _s,mix)	0.4194kg·m^?2·s^?1	1.67	0.7035kg·m^?2·s^?1	2.49	19.56
KELM(G _s，mix)	0.1117kg·m^?2·s^?1	0.38	0.6972kg·m^?2·s^?1	2.35	14.35
ELM(X _r)	0.0779%	1.66	0.1023%	2.24	21.72
KELM(X _r)	0.0442%	0.90	0.0611%	1.48	15.85

模型	RMSE_Tr	MAPE_Tr/%	RMSE_Te	MAPE_Te/%	时间/s
ELM(G _s,mix)	0.4194kg·m^?2·s^?1	1.67	0.7035kg·m^?2·s^?1	2.49	19.56
KELM(G _s，mix)	0.1117kg·m^?2·s^?1	0.38	0.6972kg·m^?2·s^?1	2.35	14.35
ELM(X _r)	0.0779%	1.66	0.1023%	2.24	21.72
KELM(X _r)	0.0442%	0.90	0.0611%	1.48	15.85

C	——	正则化参数
d _p	——	混合颗粒中石英砂平均粒径，mm
f(x)	——	网络输出
G _s,mix	——	混合颗粒循环流率，kg/(m²·s)
H _b	——	气化室中床层物料高度，m
h (x)/ H	——	隐藏层特征映射矩阵
I	——	对角矩阵
m _int	——	初始物料量，kg
m _s,mix	——	一定时间内的收集到的混合物料质量，kg
S _r	——	提升管横截面积，m²
T	——	样本输出向量
Te	——	测试样本
Tr	——	训练样本
U _b	——	气化室风速,m/s
U _mf	——	混合颗粒中石英砂的临界流化风速，m/s
U _r	——	提升管风速，m/s
X _r	——	稻壳-石英砂混合颗粒中稻壳质量分数，%
X _r,0	——	初始稻壳-石英砂混合颗粒中稻壳质量分数，%
x/ x_i/x_j	——	输入样本/输入样本向量，其中i和j取[1,N]内的正整数
y_i	——	实际测量值
?_i	——	模型预测值
β	——	隐藏层与输出层连接权值
γ	——	核函数参数