本文關(guān)鍵詞：基于專利知識的機(jī)械產(chǎn)品創(chuàng)新設(shè)計方法研究　出處：《浙江大學(xué)》2013年博士論文　論文類型：學(xué)位論文

  更多相關(guān)文章： 專利 專利知識 創(chuàng)新設(shè)計 知識提取 知識表示 自動分類 創(chuàng)新性 作用結(jié)構(gòu) 靈感啟發(fā)

【摘要】：在綜合分析了專利文本的處理方法、基于專利知識的創(chuàng)新設(shè)計方法及其應(yīng)用系統(tǒng)的研究現(xiàn)狀后,本文針對當(dāng)前英文專利知識應(yīng)用于創(chuàng)新設(shè)計的研究中存在的一些關(guān)鍵問題,提出了新的解決方法,主要研究進(jìn)展體現(xiàn)在以下五個方面： 一、專利自動分類方法研究 研究了專利文本部分的表述特點(diǎn),分析了包括特征選擇算法和分類算法的現(xiàn)有分類方法用于專利文本自動分類的缺點(diǎn)�；谛畔⒃鲆嫠惴ê蚄鄰近算法,提出了改進(jìn)的信息增益算法用于特征選擇,改進(jìn)的K鄰近算法用于專利分類,在此基礎(chǔ)上進(jìn)一步提出了一種高效的專利自動分類方法。最后以美國專利的分類試驗為例,說明該專利自動分類方法能準(zhǔn)確有效地實(shí)現(xiàn)專利文本分類,并且有助于在基于專利的創(chuàng)新設(shè)計中的檢索和分析應(yīng)用。 二、專利創(chuàng)新性評估方法研究 根據(jù)專利中所包含知識的特點(diǎn),采用動詞和名詞分別代表原理設(shè)計知識和結(jié)構(gòu)設(shè)計知識。通過引入信息熵的概念,分別提出了專利知識新穎度和相容度的量化評估方法。結(jié)合形態(tài)學(xué)分析方法,建立了組合專利知識的創(chuàng)新性評估模型,提出了基于形態(tài)學(xué)分析的組合專利知識創(chuàng)新性評估過程。以具有自鎖功能的直線驅(qū)動機(jī)構(gòu)設(shè)計中所采用的專利知識評估作為例子,通過試驗對比分析了專利數(shù)量和專利知識詞匯數(shù)量對創(chuàng)新性評估結(jié)果的影響,驗證了提出的新穎度和相容度量化評估方法以及組合專利知識創(chuàng)新設(shè)計方法的有效性。 三、機(jī)械產(chǎn)品專利作用結(jié)構(gòu)知識獲取方法研究 針對機(jī)械產(chǎn)品專利,分析了現(xiàn)有專利作用結(jié)構(gòu)知識提取和表示方法所存在的問題,提出了針對英文專利的作用結(jié)構(gòu)知識獲取方法,包括技術(shù)對象和技術(shù)關(guān)系兩方面提取任務(wù)。提出了作用結(jié)構(gòu)知識圖形化表示方法,建立了基于可擴(kuò)展標(biāo)記語言結(jié)構(gòu)XSD和統(tǒng)一建模語言UML的專利作用結(jié)構(gòu)知識表示模型。研究了機(jī)械產(chǎn)品專利語言的特點(diǎn),采用最大熵模型和領(lǐng)域內(nèi)專利術(shù)語詞典識別并提取技術(shù)對象,通過建立組成類動詞庫識別核心動詞提取技術(shù)關(guān)系。該專利作用結(jié)構(gòu)知識表示模型具有建模便捷和提取過程計算機(jī)自動處理等特點(diǎn)。結(jié)合美國專利實(shí)例,闡述了從專利中提取作用結(jié)構(gòu)知識,寫入XSD結(jié)構(gòu)化文檔,最后轉(zhuǎn)化為作用結(jié)構(gòu)表示圖的獲取過程。 四、基于專利靈感啟發(fā)的機(jī)械產(chǎn)品創(chuàng)新設(shè)計研究 總結(jié)了專利中有助于靈感啟發(fā)的知識類型,分析了各類知識的特點(diǎn),提出了基于正則表達(dá)式和非確定有限自動機(jī)的專利靈感啟發(fā)知識提取方法,并結(jié)合專利作用結(jié)構(gòu)知識構(gòu)建了用于靈感啟發(fā)的專利知識表示模型。根據(jù)授權(quán)發(fā)明專利的發(fā)明形式,結(jié)合專利知識類型,系統(tǒng)性地提出了靈感啟發(fā)方式,發(fā)現(xiàn)了一些至今未出現(xiàn)的新啟發(fā)方式。在此基礎(chǔ)上,提出了基于專利靈感啟發(fā)的創(chuàng)新設(shè)計過程,通過構(gòu)建以靈感啟發(fā)方式和專利知識為核心的靈感啟發(fā)環(huán)境,激發(fā)設(shè)計者的靈感輔助創(chuàng)新設(shè)計。以具有自鎖功能的液壓油缸創(chuàng)新設(shè)計為例,驗證了這一創(chuàng)新設(shè)計過程的可行性和有效性。 五、基于專利知識的計算機(jī)輔助創(chuàng)新系統(tǒng)研究 以前述理論研究為基礎(chǔ),開發(fā)了由知識庫、專利庫、規(guī)則庫和功能模塊組成的基于專利知識的計算機(jī)輔助創(chuàng)新系統(tǒng),實(shí)現(xiàn)了基于專利知識的啟發(fā)創(chuàng)新設(shè)計流程。結(jié)合液壓油缸產(chǎn)品設(shè)計,說明了該系統(tǒng)功能模塊的操作界面及操作方法,初步驗證了該系統(tǒng)的有效性。通過該系統(tǒng)產(chǎn)生的創(chuàng)新設(shè)計產(chǎn)品獲得國家發(fā)明專利。
[Abstract]:In a comprehensive analysis of the patent text processing method, based on the present research situation of system innovation design method of patent knowledge and its application, aiming at some key problems of the current English patent knowledge used in the innovative design of the proposed new solutions, the main research progress is reflected in the following five aspects:
First, research on the automatic classification method of patent
Study on the expression characteristics of patent text, analysis of the existing classification methods including feature selection algorithm and classification algorithm for automatic text classification. The disadvantages of patent information gain algorithm and K neighbor algorithm based on information gain algorithm for feature selection is proposed. The improved algorithm for K adjacent patent classification, on this basis, the automatic classification method for an efficient patent. Finally to test the United States Patent Classification as an example, the automatic classification method can accurately and effectively implement the patent patent text classification, and contributes to the innovation of design patent in the retrieval and analysis based on the application.
Two, research on the method of patent innovation evaluation
According to the characteristics of the knowledge contained in patents, using verbs and nouns respectively represent the principle of design knowledge and design knowledge structure. By introducing the concept of information entropy, respectively put forward quantitative evaluation method of novelty and compatibility of patent knowledge. Combined with morphological analysis method, established evaluation model of innovative combination of patent knowledge, are proposed. Knowledge innovation patent portfolio assessment process based on morphological analysis. In line with self-locking function driven by mechanism design patent evaluation as an example, affect the number of patent number and patent knowledge vocabulary for innovative assessment results were analyzed by comparative test, verify the validity of the proposed novelty and compatibility evaluation method and combination of patent knowledge innovation design method.
Three, research on knowledge acquisition method of mechanical product patent
According to the analysis of the existing mechanical product patent, patent structure knowledge extraction and representation methods of the existing problems, proposes a method for structure of knowledge for English patents, including two aspects of the relationship between technology and object extraction technology task. Put forward the role of knowledge of the structure of the graphical representation method of knowledge representation model structure, the function of patent extensible markup language the structure of XSD and the unified modeling language based on UML is established. The characteristics of patent language of mechanical products, as to using the maximum entropy model and in the field of patent glossary identification and extraction technology, composition of animal lexicon identifying core extraction technology. The relationship between verbs by establishing the function of patent structure knowledge representation model with modeling convenience and feature extraction the process of computer automatic processing. Combined with the United States patent case, expounds the effect of extract from the patent knowledge structure, write into XS D structured documents, finally converted to the process of obtaining the graphical representation of the functional structure.
Four, research on innovation design of mechanical products based on the inspiration of patent
Summary of the patent contributes to the inspiration of knowledge type, analyzes the characteristics of various types of knowledge, puts forward the regular expression and extraction method of non deterministic finite automaton patent inspiration based on knowledge and inspiration for the patent knowledge representation model is constructed with the function of patent knowledge structure. According to the invention patent authorization form the combination of patent knowledge types, systematically put forward the inspiration, has not found some new inspiration. On this basis, put forward the innovative design process of patent based on inspiration, through building the inspiration and patent knowledge as the core of the inspiration, creative design inspiration assisted designers to stimulate. The innovative design of hydraulic cylinder with self-locking function as an example, to verify the feasibility of this innovative design process and effectiveness.
Five, research on computer aided innovation system based on patent knowledge
On the basis of the research on the theory, developed by the knowledge base, patent database, computer aided innovation system based on patent knowledge rules and function modules, realizes the inspiration of innovative design process based on the knowledge of the patent. The combination of hydraulic cylinder product design, illustrates the operating interface of the system function module and operation method, preliminary verification the effectiveness of the system. The product innovation design produced by the system won the national invention patent.

【學(xué)位授予單位】：浙江大學(xué)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2013
【分類號】：TH122

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 胡思康;曹元大;;Web網(wǎng)頁知識獲取技術(shù)[J];北京理工大學(xué)學(xué)報;2006年12期

2 王磊;檀潤華;韋子輝;楊雨;李強(qiáng);;基于文化和技術(shù)的產(chǎn)品外觀進(jìn)化模式[J];包裝工程;2008年07期

3 馬力輝;檀潤華;;基于TRIZ進(jìn)化理論和TOC必備樹的沖突發(fā)現(xiàn)與解決方法[J];工程設(shè)計學(xué)報;2007年03期

4 李萌;;基于TRIZ和DEA理論的產(chǎn)品概念設(shè)計方法[J];系統(tǒng)工程;2007年02期

5 蔣宗禮;徐學(xué)可;李帥;;文本分類中基于詞條聚合的特征抽取[J];哈爾濱工程大學(xué)學(xué)報;2008年11期

6 張惠;邱清盈;馮培恩;王朝霞;;產(chǎn)品專利設(shè)計知識獲取方法研究[J];哈爾濱工程大學(xué)學(xué)報;2009年07期

7 朱顥東;鐘勇;;基于貝葉斯粗糙集的文本特征選擇方法[J];河南師范大學(xué)學(xué)報(自然科學(xué)版);2009年04期

8 徐沛娟;李雄飛;惠s，

本文編號：1358142