Yasuo Morooka

Method and apparatus for monitoring the shaft vibration of a rotary machine

From the viewpoint of the preventive maintenance operation of a rotary machine, it is a very important subject to monitor the shaft vibration of the rotary machine and grasp the operating state of the machine. This invention relates to input processings in the case of executing the frequency analysis of a shaft vibration signal by digital processings. This invention describes the rotational frequency synchronization of a sampling frequency, the phase synchronization of the rotary machine to a reference phase, the relationship between an analytical wave number and a sample nunber, etc. in that case.

Method and apparatus for symptom diagnosis by monitoring vibration of shaft of rotary machine

A method of symptom diagnosis by continuously detecting vibration of the shaft of a rotary machine and monitoring a signal indicative of detected shaft vibration for the diagnosis of the operating condition of the rotary machine. In the method, a symptom of unusual operation of the rotary machine is diagnosed. The rotation speed range of the rotary machine is classified into a safety region, an alarm region and a trip region depending on the level of the shaft vibration signal. According to the disclosed method, symptom diagnostic regions are established within the safety region at time intervals of a symptom diagnostic period, and whether or not the level of the detected shaft vibration signal deviates from that of the symptom diagnostic regions is continuously monitored for the diagnosis of a symptom of unusual operation of the rotary machine.

Distributed diagnosis system combining the immune network and learning vector quantization

A distributed diagnosis system combining the immune network (IN) and learning vector quantization (LVQ) is proposed for accurately detecting faulty sensor outputs in control plants. The system has two execution modes, namely, its training mode, where the LVQ extracts a correlation between each two sensors from their outputs when they work properly, and its diagnosis mode, where the LVQ contributes to testing each two sensors using the extracted correlation, and the IN contributes to determining faulty sensors by integrating the local testing results obtained from the LVQ. With the proposed method, faulty sensors, such as age deteriorated ones, which have been difficult to be detected only by checking each sensor output independently, can be specified.

Adjusting neural networks for accurate control model tuning

In this paper, we propose adjusting neural networks (AJNNs), which are an extended model of conventional multilayered neural networks (CNNs), for accurate model tuning with small tuning numbers. The AJNN consists of two multilayered neural networks, namely, a CNN and an error calculation neural network (ECNN) which is added in parallel to the CNN. The ECNN calculates the output error of the CNN and subtracts it from the output of the CNN, to obtain accurate tuning values. A training method for the AJNN is also proposed, where the modified back-propagation developed for reducing the error of the AJNN and the conventional back-propagation for decreasing the output of the ECNN, are applied to the AJNN alternately. The AJNN is evaluated with model tuning of temperature control for reheating furnace plants and is demonstrated to be effective to improve the accuracy of tuning and decrease tuning numbers.<<ETX>>

Interactive Simulator Using a Graphic Terminal for Linear Control System

Abstract Cost and time needed for dynamic simulation of control systems increase very rapidly as controlled process become larger in scale and more complicated. In order to conduct process simulation efficiently, an interactive simulator using a big computer was developed. In the system, engineer can enter data about the controller, described by block diagrams in the simulator, thorough an intelligent terminal interactivelly. The block diagram is displayed on a strage type CRT (Cathode Ray Tube) immediately. With respect to analysis, time response, frequency response and root locations can be calculated and the results are displayed on CRT

Experience in installing a computer control system in a hot strip mill

It is highly desirable to reduce the application period of a computer control system to a hot strip mill. Completion of a hot strip mill computer control system chiefly depends upon the time when slab tracking is operative, and on the setup calculation method and its checking method. For slab tracking, many support programs were devised such as dump, on-line monitor, process input-output simulator, and mill line simulator. To confirm the efficiency of the setup calculation method, a unique checking procedure has been developed and has proven to be quite effective. With these preparations for the installation of a computer control system in Sumitomos 80-inch hot strip mill, it took only five months in the field to put the full system into operation.

Shape control of rolling mills by a neural and fuzzy hybrid architecture

Hitachi Ltd has developed pattern recognition and control techniques which combine neural network and fuzzy logic. Conventionally, skilled operators recognize and manually control waveform patterns based on their sense and experience. The new system recognizes and controls waveform patterns by means of neural networks and fuzzy logics to realize fully automatic shape control of rolling mills. The neural network recognizes spatially distributed waveform patterns from sensor signals, and the fuzzy logics operate multiple final control elements for automatic pattern control. The developed control technique has been applied to automatic shape control system for a Sendzimir Rolling Mill. Shape control for this type of rolling mill is difficult with conventional automatic control systems because of complicated rolling phenomena and the difficulty of creating a control models. Tests with an actual rolling system proved that the new technique achieves more accurate control than the conventional manual operation by skilled operators. The system has been applied at a few plants and is operating favorably.<<ETX>>

Interactive Simulation Method Using Tableau Approach for Nonlinear Control System

Abstract In order to design control systems and simulate processes efficiently, an interactive simulation system using a mainframe computer is newly developed. This system allows the nonlinear block diagrams used widely by control engineers to be entered in the host computer interactively through an intelligent terminal. As simulation method for a nonlinear control system, the tableau approach, a high speed calculation technique for electronic circuit, is employed. One of the advantages of the tableau method is that this method allows larger step than conventional techniques. This is confirmed in a case study.

Development of a Tension Control System for Hot Finishing Mills

Abstract A new tension control system for the hot finishing mill, a “looperless” system, was developed. This system detects inter-stand tensions by applying measured data for the rolling process to a newly developed tension equation. The rotation speed of the roll-drive motors is controlled to achieve a desired tension. Experiment on a hot finishing line showed the tension could be controlled within 0.1 kg/mm 2 . Operation was very stable in the rolling of more than 100 strips.

Adaptively changed winning number LVQ for constructing an accurate control model from enormous and low quality plant data

Construction or tuning of control models is often done using data obtained from an actually working plant. We discuss how to improve these plant data from the viewpoints of decreasing their size to a manageable number without losing their statistical property, excluding ill-suited ones, and dissolving the partial distribution to obtain an accurate control model. We call our procedure plant data purification. First, the learning vector quantization (LVQ) is improved to obtain the desired number of purified data, where, under quantization, the number of winning quantization vectors is changed adaptively and abnormal data determined by similarity with their nearest quantization vector are excluded out of a set of quantized plant data. Then the developed method is further improved to dissolve the partial distribution of data to obtain a uniform distribution. Finally, the proposed method is applied to the construction of a control model used in a continuous galvanizing plant and its effectiveness is demonstrated.