JPH0832281A - Quality analyzing method - Google Patents
Quality analyzing methodInfo
- Publication number
- JPH0832281A JPH0832281A JP16165594A JP16165594A JPH0832281A JP H0832281 A JPH0832281 A JP H0832281A JP 16165594 A JP16165594 A JP 16165594A JP 16165594 A JP16165594 A JP 16165594A JP H0832281 A JPH0832281 A JP H0832281A
- Authority
- JP
- Japan
- Prior art keywords
- defect
- data
- feature extraction
- maintenance information
- failure
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000012423 maintenance Methods 0.000 claims abstract description 39
- 238000000605 extraction Methods 0.000 claims abstract description 36
- 230000007547 defect Effects 0.000 claims description 62
- 238000013528 artificial neural network Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 12
- 230000002950 deficient Effects 0.000 claims description 11
- 238000004458 analytical method Methods 0.000 claims description 9
- 238000003908 quality control method Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 6
- 238000004519 manufacturing process Methods 0.000 description 5
- 210000002569 neuron Anatomy 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 4
- 238000007689 inspection Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 230000001364 causal effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Supply And Installment Of Electrical Components (AREA)
- Multi-Process Working Machines And Systems (AREA)
- General Factory Administration (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
【０００１】[0001]
【産業上の利用分野】本発明は、１又は複数台の設備を
有する製造工程において、設備稼働あるいは製品品質の
データから製品の不良要因あるいは設備のメンテナンス
情報を求める品質分析方法に関するものである。BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a quality analysis method for obtaining a defective factor of a product or equipment maintenance information from equipment operation or product quality data in a manufacturing process having one or a plurality of equipment.
【０００２】[0002]
【従来の技術】従来、製造工程の製品不良の要因を特定
し、製造工程のメンテナンス指示を与える方法として
は、以下のような手法が取られている。2. Description of the Related Art Conventionally, the following method has been used as a method for identifying a cause of a product defect in a manufacturing process and giving a maintenance instruction for the manufacturing process.
【０００３】図５において、設備３１の設備稼働状態あ
るいは製品品質を管理する品質管理データは定期的に通
信手段３２を用いて品質管理データベース３３に収録・
管理されている。例えば、品質管理データベース３３の
内容として、設備を認識する設備番号４１、工程を流れ
る製品の製品番号４２、製品の特定位置に発生した不良
を特定するための検査位置番号４３、不良コード４４、
また製品を計測管理する必要のある場合には計測値４５
と計測時刻４６が表として管理されている。In FIG. 5, quality management data for managing the equipment operating state of the equipment 31 or the product quality is periodically recorded in the quality management database 33 using the communication means 32.
It is managed. For example, as the contents of the quality control database 33, the equipment number 41 for recognizing the equipment, the product number 42 of the product flowing through the process, the inspection position number 43 for specifying the defect occurring at the specific position of the product, the defect code 44,
If the product needs to be measured and managed, the measured value 45
And the measurement time 46 are managed as a table.
【０００４】また品質管理データベース３３の内容は、
各種管理図表、例えば不良発生頻度の大なる順番にパレ
ート図に表現したり、時系列に不良発生頻度を見ること
ができるようにしている。The contents of the quality control database 33 are as follows:
Various control charts, for example, Pareto charts in descending order of occurrence frequency of defects, and defect occurrence frequencies can be viewed in time series.
【０００５】製造管理者は、各種管理図表から不良要因
を経験的に求め、各種設備のメンテナンスを行ってい
る。The manufacturing manager empirically finds the cause of failure from various control charts and carries out maintenance of various equipment.
【０００６】しかしながら、半導体プロセス工程、実装
工程のように多数の設備とプロセスを有する場合、管理
者は多くの分析・判断をする必要があり、メンテナンス
が後追いになってしまい、不良を発生してしまう。そこ
で、最新の品質管理方法では専門の管理者を必要とせ
ず、要因の絞り込み、メンテナンス指示を自動で出力で
きるようになっている。However, when a large number of equipments and processes such as a semiconductor process step and a mounting step are included, the manager needs to make many analyzes and judgments, and the maintenance is postponed, causing defects. I will end up. Therefore, the latest quality control method does not require a specialized administrator, but can narrow down factors and automatically output maintenance instructions.
【０００７】図６に一般的な自動要因絞り込み・メンテ
ナンス指示の方法を示す。品質管理データベースより時
刻毎あるいは製品ロット毎等で集計した管理データを読
み取り、データを各種条件群５１にて判定し、具体的な
不良要因あるいはメンテナンス表５２を出力する。その
条件群５１は、あらかじめｉｆ〜ｔｈｅｎ〜ルールで知
識ベースに登録しておけば、プログラムのフローを変更
する必要はない。また、ルールの追加・変更・削除等も
容易に行うことができる。このような方法を用いたもの
は、エキスパートシステムと呼ばれ、良く知られてい
る。FIG. 6 shows a general automatic factor narrowing / maintenance instruction method. The management data collected for each time or each product lot is read from the quality control database, the data is judged by various condition groups 51, and a specific defect factor or maintenance table 52 is output. If the condition group 51 is registered in the knowledge base in advance with if-then-rules, it is not necessary to change the program flow. In addition, rules can be easily added / changed / deleted. The one using such a method is called an expert system and is well known.
【０００８】[0008]
【発明が解決しようとする課題】しかしながら、上記エ
キスパートシステムには次のような問題がある。However, the above expert system has the following problems.
【０００９】第１に、あらかじめ不良発生とその要因の
関係を明確にルールに記述する必要がある。ところが、
複雑な工程において不良の因果関係を分析して求めるの
は困難な場合が多い。First, it is necessary to clearly describe in advance the relationship between the occurrence of defects and their factors in the rules. However,
In many cases, it is difficult to analyze and determine the causal relationship of defects in complicated processes.
【００１０】第２に、通常品質管理データベース３３の
データ容量は大であり、条件群５１が多くなるとメンテ
ナンス表５２を出力するのに多大な時間を必要とする。Secondly, the data volume of the normal quality control database 33 is large, and if the condition group 51 increases, it takes a lot of time to output the maintenance table 52.
【００１１】第３に、条件群５１は注意深くルールの整
合性を考慮して追加しないと間違ったメンテナンス表５
２を出力してしまうことになり、条件が多くなると条件
を管理する知識ベースの管理が複雑になってしまう。Thirdly, the condition group 51 must be carefully added in consideration of the consistency of the rules, and the wrong maintenance table 5 is required.
2 is output, and the management of the knowledge base for managing the conditions becomes complicated when the number of conditions increases.
【００１２】本発明は、上記従来の問題点に鑑み、１又
は複数台の設備を有する製造工程において、製品の不良
要因あるいは設備のメンテナンス情報を容易に求めるこ
とができる品質分析方法を提供することを目的としてい
る。In view of the above-mentioned conventional problems, the present invention provides a quality analysis method capable of easily obtaining a defect factor of a product or equipment maintenance information in a manufacturing process having one or a plurality of equipment. It is an object.
【００１３】[0013]
【課題を解決するための手段】本発明の品質分析方法
は、設備稼働あるいは製品品質のデータを収録し、その
データより不良特徴抽出データを抽出加工する第１工程
と、既知の典型的不良の不良特徴抽出データとそれに対
応する不良要因あるいは不良に対する設備のメンテナン
ス情報の関係を表現したデータと第１工程の不良特徴抽
出データを照合し、不良要因あるいは設備のメンテナン
ス情報を出力する第２工程とから成ることを特徴とす
る。The quality analysis method of the present invention includes a first step of recording facility operation or product quality data and extracting and processing defect feature extraction data from the data, and a known typical defect. A second process of collating the defect feature extraction data and the corresponding defect factor or the data representing the relationship between the maintenance information of the facility with respect to the defect with the defect feature extraction data of the first process, and outputting the defect factor or the maintenance information of the facility; It is characterized by consisting of.
【００１４】電子部品の実装工程においては、実装設備
の稼働あるいはプリント基板の品質のデータを用いて、
上記方法により実装不良の要因あるいは実装設備のメン
テナンス情報を出力する。In the electronic component mounting process, the data of the operation of the mounting equipment or the quality of the printed circuit board is used to
By the above method, the cause of the mounting failure or the maintenance information of the mounting equipment is output.
【００１５】上記第２工程においては、不良特徴抽出デ
ータをニューラルネットワークの入力層に入力し、それ
に対応する既知の不良要因あるいは設備のメンテナンス
情報を出力層の出力となるように予めネットワークパラ
メータを自動学習させておき、このニューラルネットワ
ークの入力層に不良特徴抽出データを入力して不良要因
あるいは設備のメンテナンス情報を出力層から出力する
のが好適である。In the second step, the defect feature extraction data is input to the input layer of the neural network, and the network parameter is automatically set in advance so that the known defect factor or equipment maintenance information corresponding thereto is output to the output layer. It is preferable to learn and input failure feature extraction data to the input layer of this neural network to output failure factors or equipment maintenance information from the output layer.
【００１６】[0016]
【作用】本発明の品質分析方法によれば、時々刻々変化
する不良の特徴となるデータのみを抽出した不良特徴抽
出データに基づいて製品不良の不良要因あるいはメンテ
ナンス情報を照合して出力するので、あらかじめ不良と
不良要因の関係を明確にルール化する必要がなく、不良
に対応する不良特徴抽出データと不良要因あるいはメン
テンナンス情報の対応関係のデータを持っているだけで
よく、また照合するだけでよいので処理時間もほぼ一定
となる。According to the quality analysis method of the present invention, the defect factor of the product defect or the maintenance information is collated and output based on the defect feature extraction data obtained by extracting only the data that is the feature of the defect that changes from moment to moment. There is no need to clearly rule out the relationship between defects and defect factors in advance, it is sufficient to have defect feature extraction data and defect factor or maintenance information corresponding to defects, and it is only necessary to collate them. Therefore, the processing time becomes almost constant.
【００１７】また、照合方法として、既知の不良に対す
るメンテナンス情報をニューラルネットワークの教師デ
ータとすることで、ネットワークパラメータにて対応関
係を管理できるため、知識ベースの管理が不要となる。Further, as a collating method, the maintenance information for a known defect is used as the teacher data of the neural network, so that the correspondence relationship can be managed by the network parameter, so that the management of the knowledge base becomes unnecessary.
【００１８】[0018]
【実施例】以下、本発明を電子部品の実装工程に適用し
た一実施例について図１〜図３を参照しながら説明す
る。DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which the present invention is applied to a mounting process of electronic parts will be described below with reference to FIGS.
【００１９】図１において、実装ライン１の各実装機あ
るいは検査機の稼働データや品質データが品質管理デー
タベース２に常時収録されている。第１工程で、これら
のデータは、実装条件・不良表３に加工される。図２に
実装条件・不良表３の具体例を示す。実装条件・不良表
３の構成として、縦軸に実装条件の項目１１、横軸に各
実装機や検査機の不良項目１２を設け、一定期間内の不
良発生件数を取る。この実装条件・不良表３は、品質管
理データベース２より一般的なリレーショナルデータベ
ースの検索あるいは結合命令で容易に作成できる。ま
た、この表は通信手段４より転送されたデータで直接作
成することもできる。In FIG. 1, operation data and quality data of each mounting machine or inspection machine in the mounting line 1 are constantly recorded in the quality control database 2. In the first step, these data are processed into the mounting condition / defect table 3. FIG. 2 shows a specific example of the mounting condition / defect table 3. As the configuration of the mounting condition / defective table 3, the vertical axis is provided with the item 11 of the mounting condition and the horizontal axis is provided with the defective item 12 of each mounting machine or the inspection machine, and the number of defects occurring within a certain period is taken. The mounting condition / defect table 3 can be easily created from the quality control database 2 by a general relational database search or a join command. Further, this table can be directly created by the data transferred from the communication means 4.
【００２０】次に、図２に示すように、実装条件・不良
表３の一部又は全部を抽出部１３として取り出し、不良
の特徴を明確に表現した不良特徴抽出表５を作成する。
不良特徴抽出表５の作成方法は、例えば期間内の不良数
を大、中、小、０の４段階の大きさで表現する。不良の
絶対数が小であっても重大なる不良の場合には大にする
必要があるため、メンバシップ関数を用いたファジィ数
で表現してもよい。また、上記例では大きさを４段階に
しているが、後の処理時間に応じて任意に決定できる。Next, as shown in FIG. 2, a part or all of the mounting condition / defect table 3 is taken out by the extraction unit 13 to create a defect feature extraction table 5 that clearly expresses the defect feature.
In the method of creating the failure feature extraction table 5, for example, the number of failures within a period is expressed in four levels of large, medium, small, and 0. Even if the absolute number of defects is small, it needs to be large in the case of serious defects, so it may be expressed by a fuzzy number using a membership function. Further, in the above example, the size is set to four levels, but it can be arbitrarily determined according to the subsequent processing time.
【００２１】又、不良特徴抽出表５にて特定される不良
パターンの典型例とこれに対して発見された不良要因と
実際に行ったメンテナンスの対応関係を対応表データベ
ース６にあらかじめ登録しておく。Further, a typical example of the defective pattern specified in the defective characteristic extraction table 5, the corresponding relationship between the discovered defective factor and the actual maintenance is registered in the correspondence table database 6 in advance. .
【００２２】照合処理部７では、不良特徴抽出表５を入
力データとして対応表データベース６の検索を行い、不
良特徴抽出表５に対応する不良要因とメンテナンス情報
を作成し、不良要因・メンテナンス表８として出力す
る。これらの一連の処理を繰り返し実施することでリア
ルタイムに品質管理ができる。The collation processing unit 7 searches the correspondence table database 6 using the failure feature extraction table 5 as input data, creates failure factors and maintenance information corresponding to the failure feature extraction table 5, and the failure factor / maintenance table 8 Output as. Quality control can be performed in real time by repeatedly performing these series of processes.
【００２３】照合処理の方法として既知のニューラルネ
ットワーク処理を用いる方法もある。不良特徴抽出表５
は、図２に示すように、０、小、中、大に対応する値
０、１、２、３を持つ抽出マトリックス１５として表現
できる。抽出マトリックス１５は、図３のニューラルネ
ットワークの入力層２１に対する入力データとして並べ
替えることが可能である。その際、図４に示すように、
マトリックスの要素ａｉｊは、０、１の値を有するニュ
ーロンａ１ 、ａ2 に分解することができ、これを入力
データとする。このようにしてａijのニューロン列は不
良特徴抽出表５を反映できる。There is also a method of using a known neural network processing as a matching processing method. Defect feature extraction table 5
Can be represented as an extraction matrix 15 with values 0, 1, 2, 3 corresponding to 0, small, medium and large, as shown in FIG. The extraction matrix 15 can be rearranged as input data to the input layer 21 of the neural network of FIG. At that time, as shown in FIG.
The elements a ij of the matrix can be decomposed into neurons a 1 and a 2 having values of 0 and 1 , and this is used as input data. In this way, the bad feature extraction table 5 can be reflected in the neuron sequence of a ij .
【００２４】入力層２１に入力された入力データａ
ijは、あらかじめニューラルネットワーク処理で学習さ
れていたニューロン間の荷重Ｗij、しきい値θk のネッ
トワークパラメータに基づき中間層２２を通り、出力層
２３へと順次計算される。出力層２３には、各ニューロ
ンに対応した不良要因・メンテナンス表８が設定されて
いるため、入力層２１への入力データの変換結果として
出力層２３には不良要因・メンテナンス表８の出力が得
られる。Input data a input to the input layer 21
ij passes through the intermediate layer 22 and is sequentially calculated to the output layer 23 based on the network parameter of the weight W ij between neurons and the threshold value θ k which has been learned in advance by the neural network processing. Since the failure factor / maintenance table 8 corresponding to each neuron is set in the output layer 23, the output of the failure factor / maintenance table 8 is obtained in the output layer 23 as the conversion result of the input data to the input layer 21. To be
【００２５】ネットワークパラメータは、以下の手順で
自動学習することができる。既知の不良特徴抽出表５に
対応する不良要因・メンテナンス表８をＮ通り用意す
る。先の例で不良特徴抽出表５と不良要因・メンテナン
ス表８は、各々入力層２１、出力層２３のデータとして
定義することができるので、ある不良要因・メンテナン
ス表８に対応する出力層２３を教師データとし、対応す
る不良特徴抽出表５の入力パターンを入力層２１に入力
し、教師データと出力層２３の結果が同一になるように
荷重Ｗij、しきい値θk を計算する。これを対応する表
が存在するＮ通りのすべてについて実施する。また、新
たな対応関係を追加した場合も同様の処理をすれば良
い。The network parameters can be automatically learned by the following procedure. N kinds of failure factor / maintenance tables 8 corresponding to the known failure feature extraction table 5 are prepared. In the above example, the failure feature extraction table 5 and the failure factor / maintenance table 8 can be defined as the data of the input layer 21 and the output layer 23, respectively, so that the output layer 23 corresponding to a certain failure factor / maintenance table 8 can be defined. The input pattern of the corresponding defective feature extraction table 5 is input to the input layer 21 as the teacher data, and the weight W ij and the threshold value θ k are calculated so that the results of the teacher data and the output layer 23 are the same. Do this for all N ways for which there is a corresponding table. Also, similar processing may be performed when a new correspondence relationship is added.
【００２６】なお、ニューラルネットワークの処理は、
ソフトウエア手続きのみでも実現できるが、ハードウエ
アにて実現すれば高速の処理が可能となる。The processing of the neural network is
It can be realized only by software procedure, but if realized by hardware, high-speed processing becomes possible.
【００２７】[0027]
【発明の効果】本発明の品質分析方法によれば、以上の
説明から明らかなように、不良の特徴となるデータのみ
を抽出した不良特徴抽出データに基づいて不良要因ある
いはメンテナンス情報を照合して出力するので、あらか
じめ不良と不良要因の関係を明確にルール化する必要が
なく、不良に対応する不良特徴抽出データと不良要因あ
るいはメンテンナンス情報の対応関係データを持ってい
るだけでよく、また照合するだけでよいので処理時間も
ほぼ一定となる。According to the quality analysis method of the present invention, as is clear from the above description, the defect factor or maintenance information is collated based on the defect feature extraction data obtained by extracting only the data that is the feature of the defect. Since it is output, it is not necessary to clearly rule out the relationship between the defect and the defect factor in advance, it is sufficient to have the defect feature extraction data corresponding to the defect and the relationship data of the defect factor or maintenance information, and also to collate. Since it is sufficient, the processing time becomes almost constant.
【００２８】また、照合方法として、既知の不良に対す
るメンテナンス情報をニューラルネットワークの教師デ
ータとすることで、ネットワークパラメータにて対応関
係を管理できるため、知識ベースの管理が不要となる。Further, as the collation method, the maintenance information for the known defect is used as the teaching data of the neural network, so that the correspondence relationship can be managed by the network parameter, so that the management of the knowledge base becomes unnecessary.
【図１】本発明の一実施例の品質分析方法の説明図であ
る。FIG. 1 is an explanatory diagram of a quality analysis method according to an embodiment of the present invention.
【図２】同実施例における特徴抽出データの作成方法の
説明図である。FIG. 2 is an explanatory diagram of a method of creating feature extraction data in the same embodiment.
【図３】同実施例におけるニューラルネットワークの説
明図である。FIG. 3 is an explanatory diagram of a neural network in the example.
【図４】同実施例における抽出マトリックスの要素をニ
ューロンに分解する例の説明図である。FIG. 4 is an explanatory diagram of an example of decomposing the elements of the extraction matrix into neurons in the embodiment.
【図５】従来例の品質分析方法の説明図である。FIG. 5 is an explanatory diagram of a conventional quality analysis method.
【図６】従来例のエキスパートシステムの説明図であ
る。FIG. 6 is an explanatory diagram of a conventional expert system.
１ 実装ライン ２ 品質管理データベース ３ 実装条件・不良表 ５ 不良特徴抽出表 ６ 対応表データベース ７ 照合処理部 ８ 不良要因・メンテナンス表 ２０ ニューラルネットワーク ２１ 入力層 ２３ 出力層 1 Mounting line 2 Quality control database 3 Mounting condition / defect table 5 Defect feature extraction table 6 Correspondence table database 7 Verification processing unit 8 Defect factor / maintenance table 20 Neural network 21 Input layer 23 Output layer
───────────────────────────────────────────────────── フロントページの続き (51)Int.Cl.6 識別記号 庁内整理番号 ＦＩ 技術表示箇所 // Ｂ２３Ｑ 41/08 Ｚ (72)発明者 佐藤 健一 大阪府門真市大字門真1006番地 松下電器 産業株式会社内─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. 6 Identification code Internal reference number FI Technical display location // B23Q 41/08 Z (72) Inventor Kenichi Sato 1006 Kadoma, Kadoma, Osaka Prefecture Matsushita Electric Industrial Within the corporation
Claims (3)
録し、そのデータより不良特徴抽出データを抽出加工す
る第１工程と、既知の典型的不良の不良特徴抽出データ
とそれに対応する不良要因あるいは不良に対する設備の
メンテナンス情報の関係を表現したデータと第１工程の
不良特徴抽出データを照合し、不良要因あるいは設備の
メンテナンス情報を出力する第２工程とから成ることを
特徴とする品質分析方法。1. A first step of recording facility operation or product quality data, and extracting and processing defective feature extraction data from the data, a known typical defective feature extraction data, and a corresponding defective factor or defect. A method of quality analysis, comprising: a second step of collating the data expressing the relationship between the maintenance information of the equipment and the failure feature extraction data of the first step and outputting the failure factor or the maintenance information of the equipment.
の稼働あるいはプリント基板の品質のデータを収録し、
そのデータより不良特徴抽出データを抽出加工する第１
工程と、既知の典型的不良の不良特徴抽出データとそれ
に対応する実装不良の要因あるいは不良に対する実装設
備のメンテナンス情報の関係を表現したデータと第１工
程の不良特徴抽出データを照合し、実装不良の要因ある
いは実装設備のメンテナンス情報を出力する第２工程と
から成ることを特徴とする品質分析方法。2. In the electronic component mounting process, the data of the operation of the mounting equipment or the quality of the printed circuit board is recorded,
The first to extract and process defective feature extraction data from the data
The process, the defect characteristic extraction data of a known typical defect, and the data expressing the relationship between the factors of the mounting defect corresponding to it or the maintenance information of the mounting equipment for the defect and the defect characteristic extraction data of the first process are collated, and the mounting defect is detected. And a second step of outputting the maintenance information of the mounting equipment.
をニューラルネットワークの入力層に入力し、それに対
応する既知の不良要因あるいは設備のメンテナンス情報
を出力層の出力となるように予めネットワークパラメー
タを自動学習させておき、このニューラルネットワーク
の入力層に不良特徴抽出データを入力して不良要因ある
いは設備のメンテナンス情報を出力層から出力すること
を特徴とする請求項１又は２記載の品質分析方法。3. In the second step, the defect characteristic extraction data is input to the input layer of the neural network, and the network parameter is automatically preliminarily set so that the known defect factor or equipment maintenance information corresponding to it is output to the output layer. 3. The quality analysis method according to claim 1, wherein learning is performed and defect feature extraction data is input to the input layer of the neural network to output defect factors or equipment maintenance information from the output layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP16165594A JP3633959B2 (en) | 1994-07-14 | 1994-07-14 | Quality analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP16165594A JP3633959B2 (en) | 1994-07-14 | 1994-07-14 | Quality analysis method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JPH0832281A true JPH0832281A (en) | 1996-02-02 |
| JP3633959B2 JP3633959B2 (en) | 2005-03-30 |
Family
ID=15739315
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP16165594A Expired - Fee Related JP3633959B2 (en) | 1994-07-14 | 1994-07-14 | Quality analysis method |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JP3633959B2 (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7181355B2 (en) | 2000-11-06 | 2007-02-20 | Kabushiki Kaisha Toshiba | Automatic quality control method for production line and apparatus therefor as well as automatic quality control program |
| JP2008051532A (en) * | 2006-08-22 | 2008-03-06 | Hitachi High-Technologies Corp | Autoanalyzer |
| JP2010262949A (en) * | 2009-04-28 | 2010-11-18 | Hitachi High-Tech Instruments Co Ltd | Method of mounting electronic component, and electronic component packaging system |
| JP2010277199A (en) * | 2009-05-26 | 2010-12-09 | Nippon Steel Corp | Facility-diagnosing device, facility-diagnosing method and facility-diagnosing program for production system, and computer-readable recording medium which records the program |
| WO2012094156A2 (en) * | 2011-01-05 | 2012-07-12 | Test Advantage, Inc. | Methods and apparatus for data analysis |
| US8417477B2 (en) | 2001-05-24 | 2013-04-09 | Test Acuity Solutions, Inc. | Methods and apparatus for local outlier detection |
| US8606536B2 (en) | 2002-05-24 | 2013-12-10 | Test Acuity Solutions, Inc. | Methods and apparatus for hybrid outlier detection |
| JP2020023040A (en) * | 2018-07-25 | 2020-02-13 | 株式会社ジェイテクト | Grinding quality estimation model generation device, grinding quality estimation device, operation command data adjustment model generation device of grinding machine and operation command data updating device of grinding machine |
| JP2021152762A (en) * | 2020-03-24 | 2021-09-30 | 株式会社Ｓｃｒｅｅｎホールディングス | Learned-model generating method, learned model, abnormality-factor estimating apparatus, substrate treating installation, abnormality-factor estimating method, learning method, learning apparatus, and learning-data preparing method |
| JP2023024705A (en) * | 2018-07-25 | 2023-02-16 | 株式会社ジェイテクト | Grinding quality estimation model generator, grinding quality estimation device, and operation command data update device of grinder |
| JP2023024706A (en) * | 2018-07-25 | 2023-02-16 | 株式会社ジェイテクト | Grinding quality estimation model generator, grinding quality estimation device, defective quality occurrence factor presumption device, and operation command data update device of grinder |
| US11853899B2 (en) | 2001-01-05 | 2023-12-26 | In-Depth Test Llc | Methods and apparatus for data analysis |
-
1994
- 1994-07-14 JP JP16165594A patent/JP3633959B2/en not_active Expired - Fee Related
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7181355B2 (en) | 2000-11-06 | 2007-02-20 | Kabushiki Kaisha Toshiba | Automatic quality control method for production line and apparatus therefor as well as automatic quality control program |
| US11853899B2 (en) | 2001-01-05 | 2023-12-26 | In-Depth Test Llc | Methods and apparatus for data analysis |
| US8417477B2 (en) | 2001-05-24 | 2013-04-09 | Test Acuity Solutions, Inc. | Methods and apparatus for local outlier detection |
| US8606536B2 (en) | 2002-05-24 | 2013-12-10 | Test Acuity Solutions, Inc. | Methods and apparatus for hybrid outlier detection |
| JP2008051532A (en) * | 2006-08-22 | 2008-03-06 | Hitachi High-Technologies Corp | Autoanalyzer |
| JP2010262949A (en) * | 2009-04-28 | 2010-11-18 | Hitachi High-Tech Instruments Co Ltd | Method of mounting electronic component, and electronic component packaging system |
| JP2010277199A (en) * | 2009-05-26 | 2010-12-09 | Nippon Steel Corp | Facility-diagnosing device, facility-diagnosing method and facility-diagnosing program for production system, and computer-readable recording medium which records the program |
| WO2012094156A3 (en) * | 2011-01-05 | 2012-12-27 | Test Actutty Solutions, Inc. | Methods and apparatus for data analysis |
| WO2012094156A2 (en) * | 2011-01-05 | 2012-07-12 | Test Advantage, Inc. | Methods and apparatus for data analysis |
| JP2020023040A (en) * | 2018-07-25 | 2020-02-13 | 株式会社ジェイテクト | Grinding quality estimation model generation device, grinding quality estimation device, operation command data adjustment model generation device of grinding machine and operation command data updating device of grinding machine |
| JP2023024705A (en) * | 2018-07-25 | 2023-02-16 | 株式会社ジェイテクト | Grinding quality estimation model generator, grinding quality estimation device, and operation command data update device of grinder |
| JP2023024706A (en) * | 2018-07-25 | 2023-02-16 | 株式会社ジェイテクト | Grinding quality estimation model generator, grinding quality estimation device, defective quality occurrence factor presumption device, and operation command data update device of grinder |
| JP2021152762A (en) * | 2020-03-24 | 2021-09-30 | 株式会社Ｓｃｒｅｅｎホールディングス | Learned-model generating method, learned model, abnormality-factor estimating apparatus, substrate treating installation, abnormality-factor estimating method, learning method, learning apparatus, and learning-data preparing method |
Also Published As
| Publication number | Publication date |
|---|---|
| JP3633959B2 (en) | 2005-03-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021184630A1 (en) | Method for locating pollutant discharge object on basis of knowledge graph, and related device | |
| DE69132873T2 (en) | Computer-aided device and process for subsequent changes in production | |
| JPH0832281A (en) | Quality analyzing method | |
| DE102017220140A1 (en) | Polling device, polling method and polling program | |
| CN101833324A (en) | Intelligent fault diagnosis system in tread extrusion process and diagnosis method thereof | |
| CN114880312B (en) | Flexibly-set application system service data auditing method | |
| CN117114420B (en) | Image recognition-based industrial and trade safety accident risk management and control system and method | |
| CN115984158A (en) | Defect analysis method and device, electronic equipment and computer readable storage medium | |
| CN117768618A (en) | Method for analyzing personnel violation based on video image | |
| JPH08314751A (en) | Fault countermeasure supporting method | |
| CN111181785B (en) | Monitoring method and device based on feedback link | |
| CN117331802A (en) | Middle-station data monitoring and analyzing system based on industrial Internet | |
| CN116955071A (en) | Fault classification method, device, equipment and storage medium | |
| DE60208415T2 (en) | METHOD FOR OPTIMIZING TEST DATA | |
| DE102023108157A1 (en) | PROCESSING NATURAL INPUTS FOR MACHINE DIAGNOSTICS | |
| CN108460119A (en) | A kind of system for supporting efficiency using machine learning lift technique | |
| CN115270950A (en) | Refrigerator fault positioning method and system based on tree model | |
| CN112347069A (en) | Manufacturing enterprise prediction type maintenance service method based on large manufacturing data | |
| CN114333179B (en) | Financial terminal fault pre-judging and warning method and system based on big data technology | |
| CN118504555B (en) | Alarm text root cause analysis method, device, medium and equipment | |
| CN112149969B (en) | Extra-high voltage direct current control protection merging unit operation state evaluation method and system | |
| CN112968941B (en) | Data acquisition and man-machine collaborative annotation method based on edge calculation | |
| CN113900865B (en) | Intelligent power grid equipment automatic test method, system and readable storage medium | |
| CN117171591B (en) | Method for analyzing dynamic change of fault correlation of numerical control machine tool | |
| CN117076327B (en) | Automatic interface detection and repair method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A977 | Report on retrieval |
Free format text: JAPANESE INTERMEDIATE CODE: A971007 Effective date: 20040819 |
|
| A131 | Notification of reasons for refusal |
Free format text: JAPANESE INTERMEDIATE CODE: A131 Effective date: 20040914 |
|
| A521 | Request for written amendment filed |
Free format text: JAPANESE INTERMEDIATE CODE: A523 Effective date: 20041115 |
|
| TRDD | Decision of grant or rejection written | ||
| A01 | Written decision to grant a patent or to grant a registration (utility model) |
Free format text: JAPANESE INTERMEDIATE CODE: A01 Effective date: 20041207 |
|
| A61 | First payment of annual fees (during grant procedure) |
Free format text: JAPANESE INTERMEDIATE CODE: A61 Effective date: 20041222 |
|
| R150 | Certificate of patent or registration of utility model |
Free format text: JAPANESE INTERMEDIATE CODE: R150 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20080107 Year of fee payment: 3 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20090107 Year of fee payment: 4 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20100107 Year of fee payment: 5 |
|
| LAPS | Cancellation because of no payment of annual fees |