公开(公告)号：EP0240974A3

公开(公告)日：1989-02-22

申请号：EP87105076.1

申请日：1987-04-06

申请人： HITACHI, LTD.

发明人： Baba, Kenji ,  Watanabe, Shoji ,  Hara, Naoki ,  Yoda, Mikio ,  Mori, Shunji ,  Nogita, Shunsuke ,  Yamasawa, Yuzi ,  Akiyama, Toshimitu

IPC分类号： C02F1/52

CPC分类号： C02F1/685 ,  C02F1/5281

摘要： An injection control system of a flocculating agent comprises: a flocculating pool (15) into which the inflow liquid flocculating agent (11) is injected and which forms flocs of suspended matters in the liquid; flocculating agent injection means (12) for injecting the flocculating agent into the flocculating pool (15); floc image pickup means (18) for photographing a state of the flocs in the flocculating pool (15) and for converting luminance data of the flocs into an electric signal; image recognizing means (30) for recognizing the shape of floc by binarizing the image signal derived from the floc image pickup means (18) on the basis of a luminance level of each pixel; flocculation state deciding means (80) for calculating a characteristic amount of a diameter distribution of the flocs on the basis of the floc shapes recognized by the image recognizing means (30); and injection amount control means (100) for controlling an amount of the flocculating agent which is injected from the flocculating agent injection means (12) on the basis of the characteristic amount.

公开(公告)号：EP0521643A1

公开(公告)日：1993-01-07

申请号：EP92305720.2

申请日：1992-06-22

申请人： HITACHI, LTD.

发明人： Enbutsu, Ichiro, Ishinazaka Apt.554 ,  Baba, Kenzi ,  Hara, Naoki ,  Yoda, Mikio ,  Watanabe, Shoji ,  Yahagi, Hayao

IPC分类号： G05B13/02 ,  G06F15/18 ,  G06F9/44

CPC分类号： G06N5/025 ,  G05B13/0285 ,  G06N3/0436 ,  Y10S706/90

摘要： In order to speed up, and simplify, automated learning of rules by a neural network making use of fuzzy logic, data (120) from a system is analyzed by a teaching data creation means (140). This groups the data into clusters and then selects a representative data item from each group for subsequent analysis. The selected data items are passed to a rule extraction means (180). This investigates relationships between the data items, to derive rules, but eliminates rules which have only an insignificant effect on the system. The result are candidate rules which are stored in a first rule base (200). The candidate rules are then compared with rules in a second rule base (240) to check for duplication and/or contradiction. Only those rules which are not duplicated and not contradictory are stored in the second rule base (240). Hence, when fuzzy inference is used to control the system on the basis of rules in the second rule base (240), only valid rules which provide a significant effect on the system are used.

摘要翻译： 为了加速和简化由使用模糊逻辑的神经网络自动学习规则，由教学数据创建装置（140）分析来自系统的数据（120）。 这将数据分组成簇，然后从每个组中选择一个代表性的数据项进行后续分析。 所选择的数据项被传递到规则提取装置（180）。 这调查数据项之间的关系，导出规则，但消除对系统影响不大的规则。 结果是存储在第一规则库（200）中的候选规则。 然后将候选规则与第二规则库（240）中的规则进行比较，以检查重复和/或矛盾。 只有那些不重复而不矛盾的规则被存储在第二规则库（240）中。 因此，当使用模糊推理来基于第二规则库（240）中的规则来控制系统时，仅使用对系统产生重大影响的有效规则。

公开(公告)号：EP0240974A2

公开(公告)日：1987-10-14

申请号：EP87105076.1

申请日：1987-04-06

申请人： HITACHI, LTD.

发明人： Baba, Kenji ,  Watanabe, Shoji ,  Hara, Naoki ,  Yoda, Mikio ,  Mori, Shunji ,  Nogita, Shunsuke ,  Yamasawa, Yuzi ,  Akiyama, Toshimitu

IPC分类号： C02F1/52

CPC分类号： C02F1/685 ,  C02F1/5281

摘要： An injection control system of a flocculating agent comprises: a flocculating pool (15) into which the inflow liquid flocculating agent (11) is injected and which forms flocs of suspended matters in the liquid; flocculating agent injection means (12) for injecting the flocculating agent into the flocculating pool (15); floc image pickup means (18) for photographing a state of the flocs in the flocculating pool (15) and for converting luminance data of the flocs into an electric signal; image recognizing means (30) for recognizing the shape of floc by binarizing the image signal derived from the floc image pickup means (18) on the basis of a luminance level of each pixel; flocculation state deciding means (80) for calculating a characteristic amount of a diameter distribution of the flocs on the basis of the floc shapes recognized by the image recognizing means (30); and injection amount control means (100) for controlling an amount of the flocculating agent which is injected from the flocculating agent injection means (12) on the basis of the characteristic amount.

摘要翻译： 絮凝剂的注射控制系统包括：絮凝池（15），流入液体絮凝剂（11）注入其中并在悬浮物中形成絮凝物; 絮凝剂注入装置（12），用于将絮凝剂注入絮凝池（15）中; 用于拍摄絮凝池（15）中的絮状物的状态并将絮凝物的亮度数据转换成电信号的絮凝图像拾取装置（18） 图像识别装置，用于通过基于每个像素的亮度级二值化从絮凝图像拾取装置（18）导出的图像信号来识别絮体的形状; 絮凝状态决定装置，用于根据由图像识别装置识别的絮状形状来计算絮凝物的直径分布的特征量; 以及用于根据特征量控制从絮凝剂注入装置（12）注入的絮凝剂的量的注入量控制装置（100）。

公开(公告)号：EP0708390B1

公开(公告)日：1999-10-13

申请号：EP95118768.1

申请日：1990-03-13

申请人： Hitachi, Ltd.

发明人： Baba, Kenji ,  Embutsu, Ichiro ,  Watanabe, Shoji ,  Yahagi, Hayao ,  Maruhashi, Fumio c/o Hitachi Intern. Izumigamori ,  Matsuzaki, Harumi ,  Matsumoto, Hiroshi ,  Nogita, Shunsuke ,  Yoda, Mikio ,  Hara, Naoki

IPC分类号： G05B13/02

CPC分类号： G05B13/027 ,  C02F1/008 ,  G05B13/0285 ,  Y10S706/903 ,  Y10S706/906

摘要： This invention discloses a method and a system of making a neuron network model learn the typical past process control results and utilizing this model for the supporting of an operation of a process. The neuron network is given a typical pattern of a plurality of input variables at different points in time as input signals, and is taught corresponding control variables as teaching signals, and untaught input patterns are given to this intelligent neuron network model to have it determine corresponding control variables. A plurality of patterns at predetermined time intervals are preferably used at once as patterns to be taught.

公开(公告)号：EP0708390A3

公开(公告)日：1997-04-02

申请号：EP95118768.1

申请日：1990-03-13

申请人： HITACHI, LTD.

发明人： Baba, Kenji ,  Enbutsu, Ichiro ,  Watanabe, Shoji ,  Yahagi, Hayao ,  Maruhashi, Fumio ,  Matsuzaki, Harumi ,  Matsumoto, Hiroshi ,  Nogita, Shunsuke ,  Yoda, Mikio ,  Hara, Naoki

IPC分类号： G05B13/02

CPC分类号： G05B13/027 ,  C02F1/008 ,  G05B13/0285 ,  Y10S706/903 ,  Y10S706/906

摘要： A method for extracting as knowledge causal relationships between input variables and an output variable of a neural circuit model, said neural circuit model being of a hierarchical structure constructed of an input layer, at least one hidden layer and an output layer and having performed learning a limited number of times by determining weight factors between mutually-connected neuron element models in different layers of the input layer, hidden layer and output layer, wherein with respect to plural routes extending from a neuron element model, corresponding to a particular input variable, of the input layer to a neuron element model, corresponding to a particular output variable, of the output layer by way of the individual neuron element models of the hidden layer, a product of the weight factors for each of the routes is determined, and the products for the plural routes are summed, whereby the sum is employed as a measure for the determination of the causal relationship between the particular input variable and the particular output variable.

公开(公告)号：EP0521643B1

公开(公告)日：1997-03-05

申请号：EP92305720.2

申请日：1992-06-22

申请人： HITACHI, LTD.

发明人： Enbutsu, Ichiro, Ishinazaka Apt.554 ,  Baba, Kenzi ,  Hara, Naoki ,  Yoda, Mikio ,  Watanabe, Shoji ,  Yahagi, Hayao

IPC分类号： G05B13/02 ,  G06F15/18 ,  G06F9/44

CPC分类号： G06N5/025 ,  G05B13/0285 ,  G06N3/0436 ,  Y10S706/90

公开(公告)号：EP0708390A2

公开(公告)日：1996-04-24

申请号：EP95118768.1

申请日：1990-03-13

申请人： HITACHI, LTD.

发明人： Baba, Kenji ,  Enbutsu, Ichiro ,  Watanabe, Shoji ,  Yahagi, Hayao ,  Maruhashi, Fumio ,  Matsuzaki, Harumi ,  Matsumoto, Hiroshi ,  Nogita, Shunsuke ,  Yoda, Mikio ,  Hara, Naoki

IPC分类号： G05B13/02

CPC分类号： G05B13/027 ,  C02F1/008 ,  G05B13/0285 ,  Y10S706/903 ,  Y10S706/906

摘要： A method for extracting as knowledge causal relationships between input variables and an output variable of a neural circuit model, said neural circuit model being of a hierarchical structure constructed of an input layer, at least one hidden layer and an output layer and having performed learning a limited number of times by determining weight factors between mutually-connected neuron element models in different layers of the input layer, hidden layer and output layer, wherein with respect to plural routes extending from a neuron element model, corresponding to a particular input variable, of the input layer to a neuron element model, corresponding to a particular output variable, of the output layer by way of the individual neuron element models of the hidden layer, a product of the weight factors for each of the routes is determined, and the products for the plural routes are summed, whereby the sum is employed as a measure for the determination of the causal relationship between the particular input variable and the particular output variable.

摘要翻译： 一种用于提取输入变量与神经电路模型的输出变量之间的知识因果关系的方法，所述神经电路模型是由输入层，至少一个隐藏层和输出层构成的分层结构，并且已经执行了学习 通过确定输入层，隐层和输出层的不同层中相互连接的神经元元素模型之间的权重因子的有限次数，其中相对于从对应于特定输入变量的神经元元素模型延伸的多个路线， 通过隐藏层的各个神经元元素模型将输入层输入到输出层对应于特定输出变量的神经元元素模型，确定每个路线的权重因子的乘积，并且产品 多数路线被归纳为总和，由此作为衡量第th之间因果关系的措施 特定的输入变量和特定的输出变量。

公开(公告)号：EP0240974B1

公开(公告)日：1992-07-29

申请号：EP87105076.1

申请日：1987-04-06

申请人： HITACHI, LTD.

发明人： Baba, Kenji ,  Watanabe, Shoji ,  Hara, Naoki ,  Yoda, Mikio ,  Mori, Shunji ,  Nogita, Shunsuke ,  Yamasawa, Yuzi ,  Akiyama, Toshimitu

IPC分类号： C02F1/52

CPC分类号： C02F1/685 ,  C02F1/5281