本文選題：BP神經(jīng)網(wǎng)絡　切入點：粒子群算法　出處：《東北石油大學》2015年碩士論文　論文類型：學位論文

【摘要】：在原油穩(wěn)定工藝中原油加熱的溫度直接關系到不凝氣、輕烴的產(chǎn)量,原穩(wěn)加熱爐在原油穩(wěn)定過程中起著至關重要的作用,原穩(wěn)加熱爐出口溫度的控制直接影響著產(chǎn)品的產(chǎn)量。因此,通過提高原油穩(wěn)定工藝中原油溫度控制的技術水平,從而提高原油穩(wěn)定裝置的產(chǎn)品率具有重要的意義。本文采用多學科相融合的方法,對神經(jīng)網(wǎng)絡與粒子群優(yōu)化算法在原穩(wěn)加熱爐參數(shù)優(yōu)化方面進行了深入研究。通過理論分析和實際研究,分析某原穩(wěn)裝置加熱爐運行的現(xiàn)狀和存在的問題,得出影響原穩(wěn)加熱爐出口溫度的因素主要有以下幾個方面:原油加熱爐入口溫度、原油進加熱爐流量、原油進加熱爐入口壓力、原油加熱爐燃料氣流量、原油出加熱爐壓力、加熱爐入口調節(jié)閥開度。通過分析和研究現(xiàn)有的神經(jīng)網(wǎng)絡算法的特點、優(yōu)勢以及適用范圍,結合加熱爐的工藝參數(shù)與研究目標之間的關系特點,選用BP神經(jīng)網(wǎng)絡算法建立加熱爐的數(shù)學模型。然后,分析對比各種優(yōu)化算法,最終確定采用粒子群優(yōu)化算法對參數(shù)進行優(yōu)化,采用求取目標函數(shù)最小值的辦法尋找原穩(wěn)加熱爐工藝參數(shù)的最優(yōu)參數(shù)組合。通過本文的研究,得到了原穩(wěn)加熱爐出口溫度控制的最優(yōu)參數(shù)組合,并進行現(xiàn)場應用,對比優(yōu)化前與優(yōu)化后變化,達到了良好的溫度控制結果。該方法有效的提高了原油穩(wěn)定加熱爐的性能,解決了優(yōu)化前原穩(wěn)加熱爐四條支管溫差大、偏流結焦、溫度波動大,調整頻繁等問題。有利于操作人員有針對性的對參數(shù)進行調整,較大的提高了參數(shù)控制水平,為原油穩(wěn)定加熱爐工藝參數(shù)的調整提供重要的理論依據(jù)和實踐指導。
[Abstract]:In crude oil stabilization process, the heating temperature of crude oil is directly related to the output of uncondensed gas and light hydrocarbon, and the original stable heating furnace plays an important role in the process of crude oil stabilization. The control of the outlet temperature of the original stable heating furnace directly affects the output of the product. Therefore, by improving the technical level of the crude oil temperature control in the crude oil stabilization process, Therefore, it is of great significance to improve the product rate of crude oil stabilizer. The neural network and particle swarm optimization (PSO) algorithm are studied in this paper. Through theoretical analysis and practical research, the present situation and existing problems of a primary stabilizer furnace are analyzed. The main factors affecting the outlet temperature of the original stable heating furnace are as follows: the inlet temperature of crude oil furnace, the flow rate of crude oil furnace, the inlet pressure of crude oil furnace, the fuel gas flow rate of crude oil heating furnace, the pressure of crude oil reheating furnace. Through the analysis and research on the characteristics, advantages and applicable range of the existing neural network algorithms, the characteristics of the relationship between the process parameters of the furnace and the research objectives are analyzed and studied. BP neural network algorithm is used to establish the mathematical model of reheating furnace. Then, after analyzing and comparing various optimization algorithms, the particle swarm optimization algorithm is used to optimize the parameters. The optimal parameter combination of the process parameters of the original stable reheating furnace is found by using the method of finding the minimum value of the objective function. Through the research in this paper, the optimal parameter combination of the outlet temperature control of the original stable heating furnace is obtained, and the field application is carried out. Compared with the changes before and after optimization, the results of temperature control are good. This method can effectively improve the performance of crude oil stable heating furnace, and solve the problem that the temperature difference of the four branches of the original stable heating furnace before optimization is large, the slanting coking, and the temperature fluctuating greatly. It is helpful for the operators to adjust the parameters, improve the control level of the parameters, and provide important theoretical basis and practical guidance for the adjustment of the process parameters of crude oil stable heating furnace.
【學位授予單位】：東北石油大學
【學位級別】：碩士
【學位授予年份】：2015
【分類號】：TE963

【參考文獻】

相關期刊論文 前1條

1 劉志中;海林鵬;薛霄;李雯睿;;一種新型混合仿生智能算法及其應用研究[J];計算機應用研究;2013年12期

，

本文編號：1566320