Research On Process Monitoring Of Secondary Cooling Water In Continuous Casting And The Reliability Of Its Data

Along with the development of continuous casting technique, improving output and quality of products becomes an important problem in continuous casting research. Cooling water in secondary cooling process is the main factor in the yield of process and the quality of products, and adopting the second class optimization dynamically technique in cooling water is the valid means of raising the product's quantity and the production efficiently. So making assurance of the data credibility in'the cooling process and the normal working appearance of the equipments is very necessary. It is needed to discover and solve the problems of the equipments in the cooling process timely, such as the nozzles'saturated, leaking and blocking. At the same time, in the continuous casting process, the data provided to the second class system must be reliability; the fault data must be discovered in time; the fault data must be avoided from the optimization and management such as cooling water in secondary cooling process. The paper chooses the No.5 machine of continuous casting in San Ming Steelworks as the study object, and establish monitoring model in the continuous casting process based on the actual data.Firstly, carry on the data in pre-process measurement. The data quantity has direct influence with the establishment of the process monitoring model and monitoring effect, so it is needed to carry on the data in pre-process measurement when establishing the statistic model to monitor the process. The paper analyzes four important factors influenced the data quality:no-dimension, having the outlier, de-noising, the quantity of monitoring data, and put forward the method of the data processing aimed at any influenced factor:data standardization, cleaning the outlier, filtering with wavelet, choosing the quantity of monitoring data.Secondly, with principal component analysis method for theory tool, based on the process production data of the No.5 machine of continuous casting in the San Ming Steelworks in Fijian, aiming at the device breakdown of the continuous casting process for cooling water (such as nozzles'saturated, blocking and leaking), making use of the principal component analysis method, establish continuous casting process monitoring model. Aiming at the characteristic of slow process changes in the process, establish the Recursive PCA monitoring model based on PCA. Meanwhile, aiming at the disadvantage of the T2 control limit, put forward the improving method. By comparing with the PCA monitoring model, RPCA monitoring model can reduce the missing and failing to report phenomenon effectively.Thirdly, on the foundation of PCA, make further study of the continuous casting process monitoring method based on independent component analysis. The independent component obtained by ICA satisfies the independent characteristic under the meaning under of statistics. It is stronger than the irrelevant characteristic of principal components obtained by PCA. By comparing with principal component analysis process monitoring, the ICA method has better monitoring ability. It can monitor the occurrence of the breakdown intelligently. Meanwhile, it has better supervision effect to the breakdown that couldn't be discovered by the PCA method.Finally, establish the applied platform of the continuous casting process monitoring system. In order to apply the multivariate statistical monitoring method in the produce spot, design and develop the continuous casting process online monitoring system in the environment of VB 6.0. By monitoring, analyzing and diagnosis in the continuous casting system, especially the cooling device of the secondary cooling process online, carry out the breakdown prevention of the system, and make sure the smoothly progress of the production.