基于LM-BP神经网络的煤炭地下气化选址决策探讨

发布时间：2018-04-29 01:39

本文选题：煤炭地下气化 + 选址决策　；参考：《中国矿业大学》2017年硕士论文

【摘要】：煤炭地下气化技术是一种既能实现煤炭资源绿色、安全、高效开采,又能通过改变煤气后续利用方式实现煤气资源洁净与综合利用的煤炭开采新技术。科学选址是煤炭地下气化技术中重要一环,直接影响到后续气化炉的建立,产气的稳定性和优劣性,以及最终的经济效益和环境效益。因此,需要建立一个行之有效的煤炭地下气化选址决策系统。本文引入BP人工神经网络,综合采用文献调研、理论分析、数值模拟以及仿真应用相结合的研究方法,提出可行的选址模型,取得了如下创新性成果:(1)本文着重对煤炭地下气化可行性的资源条件影响因素进行了全面分析,如地质构造、水文地质条件、煤层赋存条件、煤质等,选取了14项因素作为煤炭地下气化项目可行性评价指标,并根据地下气化项目的特点,确定了各评价指标的合理取值范围,建立煤炭地下气化选址评估体系。(2)通过对神经网络的分析,本文首次将BP人工神经网络引入煤炭地下气化选址决策中,设计出基于标准BP神经网络和LM-BP神经网络的煤炭地下气化选址评估模型的网络结构。利用Matlab,对优化的LM-BP神经网络与标准BP神经网络进行对比实验和分析,寻求出最优的网络相关参数。实验证明,优化的LM-BP神经网络具有更好的性能。对三个样本案例进行仿真应用,仿真结果表明,基于LM-BP神经网络的煤炭地下气化评估模型评估准确可靠,能够指导煤炭地下气化的选址评估。对煤炭地下气化发展有重要意义。(3)本文在对仿真实例详细分析的同时,对BP人工神经网络在煤炭地下气化中的应用进行两方面的扩展,即可根据煤气组分预测热值和根据气化日期预测热值,具有进一步研究的价值。
[Abstract]:Underground coal gasification technology is a new coal mining technology which can not only realize green, safe and efficient mining of coal resources, but also realize clean and comprehensive utilization of gas resources by changing the way of gas utilization. Scientific location selection is an important part of underground coal gasification technology, which directly affects the establishment of subsequent gasifier, the stability and quality of gas production, as well as the final economic and environmental benefits. Therefore, it is necessary to establish an effective decision system for underground coal gasification location. In this paper, BP artificial neural network is introduced, and a feasible location model is put forward by combining the research methods of literature investigation, theoretical analysis, numerical simulation and simulation application. In this paper, the factors affecting the feasibility of underground coal gasification are analyzed, such as geological structure, hydrogeological conditions, coal seam occurrence conditions, coal quality, etc. This paper selects 14 factors as the feasibility evaluation index of underground coal gasification project, and according to the characteristics of underground gasification project, determines the reasonable value range of each evaluation index. Based on the analysis of neural network, BP artificial neural network is introduced into the decision of underground coal gasification site selection for the first time. The network structure of coal underground gasification location evaluation model based on standard BP neural network and LM-BP neural network is designed. The optimized LM-BP neural network and the standard BP neural network are compared and analyzed by Matlab, and the optimal network parameters are found out. Experiments show that the optimized LM-BP neural network has better performance. The simulation results show that the evaluation model of underground coal gasification based on LM-BP neural network is accurate and reliable and can guide the evaluation of underground coal gasification location. In this paper, the application of BP artificial neural network in underground coal gasification is extended in two aspects, while the simulation example is analyzed in detail. The calorific value can be predicted according to the composition of gas and the date of gasification, which has the value of further study.
【学位授予单位】：中国矿业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD84

【参考文献】