本文選題：鋼鐵工業(yè) + 熱軋產(chǎn)品��； 參考：《東北大學(xué)》2013年碩士論文

【摘要】：鋼鐵工業(yè)是我國(guó)能源消耗的重點(diǎn)行業(yè),能源消耗大,資源利用率低,在能源需求日趨緊張的今天,所面臨的形勢(shì)異常嚴(yán)峻。本文通過(guò)對(duì)鋼鐵企業(yè)典型產(chǎn)線(xiàn)以及產(chǎn)品的能源介質(zhì)消耗預(yù)測(cè)問(wèn)題的研究,提出了三種基于數(shù)據(jù)驅(qū)動(dòng)的預(yù)測(cè)算法,來(lái)預(yù)測(cè)未來(lái)一段時(shí)間的能源消耗值,并分別開(kāi)發(fā)了面向工序和面向產(chǎn)品的鋼鐵企業(yè)能源消耗預(yù)測(cè)系統(tǒng),目的在于提高鋼鐵企業(yè)能源利用率,減少生產(chǎn)成本。本文的主要內(nèi)容如下：(1)以國(guó)內(nèi)某鋼鐵企業(yè)生產(chǎn)過(guò)程中面臨的能源介質(zhì)消耗預(yù)測(cè)問(wèn)題作為研究背景,通過(guò)現(xiàn)場(chǎng)調(diào)研,提煉出鋼鐵企業(yè)面向典型工序的能源介質(zhì)消耗預(yù)測(cè)問(wèn)題,并進(jìn)一步提煉出面向熱軋產(chǎn)品的能源介質(zhì)消耗預(yù)測(cè)問(wèn)題。(2)針對(duì)能源介質(zhì)預(yù)測(cè)問(wèn)題的動(dòng)態(tài)特點(diǎn),分別提出了三種預(yù)測(cè)算法：傳統(tǒng)的線(xiàn)性回歸算法和最小二乘支持向量機(jī)算法,以及基于參數(shù)模型的近似動(dòng)態(tài)規(guī)劃算法。對(duì)于基于參數(shù)模型的近似動(dòng)態(tài)規(guī)劃算法,分別采用隨機(jī)梯度法與遞推最小二乘法獲取一階和二階模型參數(shù),并通過(guò)比較兩種參數(shù)模型的值函數(shù),最終確定預(yù)測(cè)模型。對(duì)三種算法進(jìn)行了數(shù)值實(shí)驗(yàn),試驗(yàn)結(jié)果表明基于參數(shù)模型的近似動(dòng)態(tài)規(guī)劃算法要明顯優(yōu)于傳統(tǒng)的線(xiàn)性回歸算法以及最小二乘支持向量機(jī)算法。(3)針對(duì)鋼鐵企業(yè)面向工序的能源消耗預(yù)測(cè)問(wèn)題,設(shè)計(jì)并開(kāi)發(fā)了鋼鐵生產(chǎn)工序能源預(yù)測(cè)系統(tǒng),該系統(tǒng)是一個(gè)綜合性管理系統(tǒng),可以準(zhǔn)確有效預(yù)測(cè)未來(lái)一段時(shí)間的各個(gè)工序的介質(zhì)消耗值,實(shí)現(xiàn)了介質(zhì)平衡分析和指標(biāo)管理功能。(4)針對(duì)熱軋產(chǎn)品的能源介質(zhì)消耗精細(xì)計(jì)量問(wèn)題,設(shè)計(jì)并開(kāi)發(fā)了熱軋產(chǎn)品能耗精細(xì)計(jì)量與預(yù)測(cè)系統(tǒng),該系統(tǒng)主要針對(duì)熱軋產(chǎn)線(xiàn)的具體產(chǎn)品做能源介質(zhì)消耗預(yù)測(cè),為精細(xì)化的能源管理提供了有效決策支持。
[Abstract]:Iron and steel industry is the key industry of energy consumption in our country. The energy consumption is large and the utilization ratio of resources is low. In this paper, three data-driven prediction algorithms are proposed to predict the energy consumption for a period of time in the future by studying the energy consumption prediction problem of typical production lines and products in iron and steel enterprises. A process oriented and product-oriented energy consumption prediction system for iron and steel enterprises is developed in order to improve the energy utilization ratio and reduce the production cost of iron and steel enterprises. The main contents of this paper are as follows: (1) based on the research background of energy medium consumption prediction in the production process of a domestic iron and steel enterprise, through field investigation, the energy medium consumption prediction problem for typical processes in iron and steel enterprises is extracted. The problem of energy medium consumption prediction for hot rolled products is further refined. According to the dynamic characteristics of the energy medium prediction problem, three prediction algorithms are proposed: the traditional linear regression algorithm and the least square support vector machine algorithm. And approximate dynamic programming algorithm based on parameter model. For the approximate dynamic programming algorithm based on parameter model, random gradient method and recursive least square method are used to obtain the first and second order model parameters, and the prediction model is finally determined by comparing the value functions of the two parameter models. Numerical experiments are carried out on three algorithms. The experimental results show that the approximate dynamic programming algorithm based on parameter model is obviously superior to the traditional linear regression algorithm and least squares support vector machine algorithm. The energy prediction system of iron and steel production process is designed and developed. The system is a comprehensive management system, which can accurately and effectively predict the medium consumption value of each working procedure in the future. The function of medium balance analysis and index management is realized. Aiming at the problem of energy medium consumption fine measurement of hot rolled products, a fine energy consumption measurement and prediction system for hot rolled products is designed and developed. The system mainly makes energy medium consumption prediction for the specific products of hot rolling production line, and provides effective decision support for fine energy management.
【學(xué)位授予單位】：東北大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2013
【分類(lèi)號(hào)】：F426.31;TP311.52
，

本文編號(hào)：1786033