数据挖掘在生产工艺参数调节中的应用
文献类型：学位论文
| 作者 | 李开拓 |
| 学位类别 | 硕士 |
| 答辩日期 | 2017-05-24 |
| 授予单位 | 中国科学院沈阳自动化研究所 |
| 授予地点 | 沈阳 |
| 导师 | 彭慧 |
| 关键词 | 工业 数据挖掘 参数调节 特征选择 模式匹配 |
| 其他题名 | Application of Data Mining in Production Process Parameter Adjustment |
| 学位专业 | 控制工程 |
| 中文摘要 | 目前，我国的大部分制造企业智能化水平不足，仍是根据现场经验进行参数调节来满足工艺要求。随着新一代信息技术与制造业的深度融合，大量的生产过程数据得以采集存储起来，通过数据挖掘等技术深入分析蕴藏在海量工业生产过程数据中的潜在规律，建立基于数据驱动的工业过程参数调节方法得到了广泛的应用。基于此，本文提出了一种基于操作模式匹配的生产过程参数调节方法，同时将该方法应用于干燥机生产过程数据，最后通过基于案例推理的工艺指标预测模型验证了该方法的有效性。本文研究的主要内容有： （1）针对直接从现场采集的工业数据中存在重复值、缺失值和噪声值等问题。本文对干燥机6个月的生产过程数据进行了数据集成、数据清洗、数据变换和数据归约等必要的数据预处理，并对生产过程数据进行了初步分析。 （2）针对数据中存在的大量不相关和冗余特征会对数据挖掘模型产生负面影响，从而降低模型性能的问题。本文提出了基于多种相关性度量的过滤式特征选择算法，另外，本文提出的算法也考虑了特征之间的交互性。将工业生产过程数据应用于新算法，同时结合现场专家经验，最终确定对工艺指标影响比较大的8个变量。 （3）针对工业生产过程难以建立机理模型、优化调节参数困难等问题，本文提出了一种基于操作模式匹配的生产过程参数调节方法。该方法利用历史优秀生产数据建立优良样本库，并在此基础上通过建立典型操作模式库来指导生产过程参数调节。如果当前生产状况不满足实际要求，需要对参数进行调节时，基于操作模式匹配的参数调节方法首先根据当前生产状况与典型操作模式库中的操作模式进行操作模式匹配，寻找相似性最大的典型操作模式，最后提取出典型操作模式中的操作参数作为当前操作参数调节的目标值。同时将该方法应用于干燥机生产过程数据。 （4）针对基于操作模式匹配的生产过程参数调节方法难以验证的问题，本文通过基于案例推理的工艺指标预测方法验证了参数调节方法的有效性。 |
| 英文摘要 | At present, the intelligence level of the manufacturing enterprises in our country is insufficient, and adjust the parameters according to the field experience to meet the technological requirements. With the deep integration of the new generation of information technology and manufacturing, the large amounts of production process data can be collected and stored, through data mining technology in-depth analysis of latent rules in the massive industrial process, and establish industrial process parameters adjustment method based on data driven has been widely used. Based on this, this paper presents a method of adjusting production process parameters based on pattern matching, this method was also used in the dryer production process data, and through the case reasoning process prediction model to verify the effectiveness of the method finally. The main contents of this paper are: (1) For some problems in the industrial data collected directly from the field, such as repeated values, missing values and noise values. In this paper, we make use of data integration, data cleaning, data transformation and data reduction to preprocess the 6 months data of the drying machine, and preliminary analysis the data. (2) For a large number of irrelevant and redundant features in the data will have a negative impact on the data mining model, so as to reduce the performance of the model. In this paper, we propose a filter feature selection algorithm based on multiple correlation measures. The industrial production process data is applied to the new algorithm, and ultimately determine the important process parameters of the 8 variables with the experience of experts. (3) For the difficulty of establishing the mechanism model and optimizing the parameters of the industrial process, this paper puts forward a method to adjust the production process parameters based on the operation mode matching. Based on the excellent production data, this paper establishes an excellent sample library, and establish a typical operation pattern database based on the excellent sample library to guide the adjustment of production process parameters. If need to adjust the parameters for the current situation does not meet the actual production requirements, operation parameters adjustment method based on pattern matching first mode of operation according to the current situation of the production and operation of the typical pattern in the library to find most similar mode of typical operation, finally extract the typical operating parameters as the current mode of operation operating parameter adjustment target value. And the method is also applied to the drying process data. (4) In order to solve the problem that the parameter adjustment method of production process based on operation pattern matching is difficult to verify, the effectiveness of the method is verified by the method of case-based reasoning. |
| 语种 | 中文 |
| 产权排序 | 1 |
| 源URL | [http://ir.sia.cn/handle/173321/20528]  |
| 专题 | 沈阳自动化研究所_数字工厂研究室 |
| 推荐引用方式 GB/T 7714 | 李开拓. 数据挖掘在生产工艺参数调节中的应用[D]. 沈阳. 中国科学院沈阳自动化研究所. 2017. |
入库方式： OAI收割
来源：沈阳自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。