Data processing: generic control systems or specific application – Specific application – apparatus or process – Product assembly or manufacturing
Reexamination Certificate
2000-05-30
2003-05-13
Picard, Leo (Department: 2125)
Data processing: generic control systems or specific application
Specific application, apparatus or process
Product assembly or manufacturing
C700S031000, C164S151500, C164S452000, C164S454000
Reexamination Certificate
active
06564119
ABSTRACT:
TECHNICAL FIELD
This invention relates to a method of monitoring specific continuous steel casting machine parameters and using this information to predict the possibility for a rupture to occur in a solidified steel shell prior to actual occurrence such that action can be taken to avoid the rupture.
BACKGROUND ART
Continuous steel casting, in the iron and steel industry, is the process of converting liquid steel into solid steel slabs or strands. This transformation of state from liquid to solid is achieved through a process known as continuous casting. In this process, the liquid steel is continuously poured into an open copper mould. Cooling water is supplied internally to the mould walls so that liquid steel in contact with the copper mould solidifies forming a solid shell that contains liquid steel within the interior of the cast strand. The solidified steel shell is continuously withdrawn from the mould into additional cooling chambers of the caster, where the remaining internal liquid steel solidifies under controlled cooling conditions.
During the casting process, ruptures in the solidifying shell can occur due to localized liquid steel not solidifying properly. When such a rupture reaches the end of the mould, molten steel spills through the rupture and causes extensive damage to the caster. This phenomenon is known as a breakout. Breakouts result in a large maintenance cost and production losses and can lead to hazardous conditions that adversely impact production safety. Breakouts can be avoided if the casting speed is reduced whenever the steel does not solidify properly. Reduction in casting speed gives more time for the steel to solidify and also reduces productivity. To avoid the occurrence of a breakout, it is critical to predict improper solidification of the steel shell with enough lead time to take corrective action.
Casters in the steel-making industry typically use breakout detection systems that look for specific patterns in the mould temperature readings. These pattern-matching systems are based on past caster breakout experience. Rules are developed that characterize the patterns in the temperatures prior to the incidence of a breakout. If patterns in the mould temperature readings follow these rules, there is a high likelihood that a breakout will occur. If the conditions of these rules are met, the typical breakout systems output an alarm to the operator to take the necessary action to prevent the breakout or take the action automatically. This normally means slowing down the casting speed. However, only a subset of all process data from the caster operation is used in developing these rules. These rules typically involve finding specific differences and rate of change variations for specific mould temperature readings. Typical rules are of the following style:
the rate of change for thermocouple A is greater than X degrees Celsius for Y consecutive readings;
the reading from thermocouple B is greater than the reading from thermocouple C for Z consecutive readings.
Current industrial breakout detectors generate an alarm only when a predetermined set of rules has been satisfied, indicating that a breakout is imminent. These systems provide a binary signal as output, alarm or not. There is no indication as to when the system is approaching alarm or the severity of the alarm. In some cases, there is not enough lead time to react to prevent the breakout from occurring. This inevitably results in some breakouts occurring without detection. To date, no known system has been able to detect every type of breakout. Having some breakouts is considered part of the cost of operating a continuous caster.
Pattern-matching detection systems of this type are described by Yamamoto et al in U.S. Pat. No. 4,556,099, Blazek et al in U.S. Pat. No. 5,020,585, Nakamura et al in U.S. Pat. No. 5,548,520, and by Adamy in U.S. Pat. No. 5,904,202.
In addition to prior art in the field of breakout detection systems for continuous casters, Applicant is aware of prior art in the area of process monitoring and fault detection. For example, a class of monitoring systems has been described in the
Canadian Journal of Chemometrics
, Vol. 69, by Kresta, MacGregor, and Marlin in 1991 (and by others since), based on the use of a multivariate process model to describe the normal operation of a process. In this approach, new data are supplied to a model in real time, and calculations are made to determine a prediction error and summary, (latent), variables. These calculated data are then tested to determine if the process is operating normally or not. This is basically the approach adopted by Wang et al for detecting faults in wafer fabrication tools as described in U.S. Pat. No. 5,859,964.
A flowchart of a generic monitoring system as described by the published prior art is shown in FIG.
1
.
Such a system is typically deployed on a computer with access to sensor signals from field instruments using a video monitor for output display. The system acquires the process signals as input to a mathematical model and computes output values as depicted in Block
10
. Block
12
provides for the computation of test statistics such as a prediction error to be used in the next step. The decision whether or not the new observation is normal is made in Block
13
. Threshold tests are done on the test statistics to determine the likelihood of the new observation belonging to the set of normal operation. If the new data are deemed normal, the system repeats the process from Block
10
at the next sample interval, but, if the likelihood is sufficiently low, a signal is issued to take corrective action on the process, either manually or automatically. Block
14
provides for determining contributions to the test statistics. Information to direct appropriate actions is displayed. The final block shown in the figure, Block
15
, provides for corrective action to be taken to avert or mitigate the fault detected above. The system continues to loop through the algorithm starting again at Block
10
.
This approach was tested to determine if it was applicable to a continuous casting process by Vaculik in 1995. The results of this off-line work showed the applicability of the technology to the particular process and are fully described in Vaculik's Master Thesis entitled
Applications of Multivariate Projections Methods in the Steel Industry
, M. Eng. Chemical Engineering, McMaster University, Hamilton, Ontario, 1995, the disclosure of which is herein incorporated by reference. What is not included in this work, however, are details required to implement a viable on-line system. The work did provide motivation for the development of an on-line system to detect abnormal operation, including breakouts. Several significant innovations were required to realize the system in its present form. These novelties are departures from prior art and are integral to the successful operation of the system; they are described below.
DISCLOSURE OF INVENTION
The invention is an on-line monitoring and fault detection system for a continuous casting process based on the application of a multivariate model of normal process operation. Additional aspects of the invention deal specifically with on-line system implementation and model development not found in the prior art.
In accordance with this invention, it is proposed to use an extended set of process measurements, beyond the standard mould temperatures, to develop a multivariate statistical model to characterize the casting process. The model is then used in the context of a monitoring system that detects exceptions to normal operation and predicts breakouts in the continuous casting process allowing for corrective action to be taken to avoid a breakout. The system is implemented on a computer using sensor inputs from the casting process to provide input data.
The invention relates to predicting the occurrence of improper solidification of the steel in a caster mould. This prediction process is based on a multivariate statistical model of normal caster operation. The model is
MacCuish R. Blair
Mutha Rajendra K.
Vaculik Vit
Dofasco Inc.
Picard Leo
Rodriguez Paul
Schmidt Ingrid E.
LandOfFree
Multivariate statistical model-based system for monitoring... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Multivariate statistical model-based system for monitoring..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Multivariate statistical model-based system for monitoring... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3019627