电力变压器状态评估及预测方法的研究

发布时间：2018-04-05 08:32

本文选题：综合集成赋权　切入点：多参数灰色预测模型　出处：《青岛科技大学》2017年硕士论文

【摘要】：电力变压器是电力系统中极其重要的设备之一,保障它的安全、稳定运行对人们的生产、生活具有很重要的意义。为了使变压器能够得到合理的检修,论文对电力变压器的状态评估及预测方法做了深入研究。不仅建立了比较系统的状态评估指标体系,而且针对电力变压器状态的评估和预测都做了大量的仿真和实例论证,最终确定了改进的评估和预测方法真实有效,具有科学性。具体的工作内容如下:(1)论文对电力变压器的状态评估指标体系做了研究。科学有效的评估指标体系是状态评估的基础,论文在搜集整理大量技术标准、规程导则、专家经验以及变压器实际运行状态数据的基础上,深入研究了电力变压器状态评估的指标体系。最终从油色谱试验、电气试验、油化试验三个方面建立了完整的电力变压器状态评估指标体系,保证了变压器状态评估和预测的准确性。(2)论文对电力变压器的状态评估做了研究。论文将变压器的状态等级进一步细化,将其划分为5个等级。在此基础之上论文不仅将集对分析理论和模糊理论应用到了电力变压器状态评估中,而且从变压器状态评估的每一步仔细剖析,提出了一种主客观综合集成赋权的算法。该算法综合考虑主观专家经验和客观事实,使变压器状态评估的指标权重更加可靠、合理。运用主观赋权法中的5级标度法充分体现了专家经验对变压器状态评估的作用;运用客观赋权法中的熵值法充分体现了各指标数据的变化情况在权重确定中的作用,符合客观事实,最后采用最小二乘法的原理综合集成了主客观权重,得出的指标权重从理论上讲最科学。最终评估出来的变压器状态等级结果与其他方法相比最为准确。这一改进的评估算法为变压器的状态检修提供了有力保障。(3)论文对电力变压器的状态预测做了研究。本文主要对状态预测过程中的数据预处理和预测模型方面进行了研究。在数据预处理方面的创新性在于使用了平均弱化算子处理方式,电力变压器的特征气体数据经过平均化处理和弱化处理之后,它的平滑性更加符合预测模型的要求,预测的精度相对而言就有所提高。在预测模型方面的改进在于论文运用了多参数灰色预测模型MGM(1,N)模型对电力变压器的特征气体进行分组预测。分组方式是以电力变压器常用故障诊断方法三比值诊断法为基础进行分组,这个方法可以避免多参数灰色预测模型预测所需要的关联度分析,从而简化了模型的计算步骤,计算时间也得到了节省。通过对实验结果的分析可以看出采用了基于三比值法分组的平均弱化多参数灰色预测模型在预测方面准确度更高,更能把握系统的规律性。最后利用理想点解理论和预测出来的数据对电力变压器的状态进行了评估,结果表明真实、可靠。
[Abstract]:Power transformer is one of the most important equipments in power system. It is very important to ensure its safety and stable operation for people's production and life.In order to enable the transformer to get reasonable overhaul, the paper makes a deep research on the state evaluation and prediction method of power transformer.Not only a systematic evaluation index system is established, but also a large number of simulations and examples are made for the evaluation and prediction of power transformer status. Finally, it is determined that the improved evaluation and prediction method is true, effective and scientific.The main work is as follows: (1) the paper studies the power transformer condition evaluation index system.A scientific and effective evaluation index system is the basis of state evaluation. Based on the collection and arrangement of a large number of technical standards, rules and guidelines, expert experience and actual operation status data of transformers,The index system of power transformer condition evaluation is studied in depth.Finally, a complete power transformer condition evaluation index system is established from three aspects: oil chromatographic test, electric test and oil test.It ensures the accuracy of transformer state evaluation and prediction.In this paper, the state of the transformer is further refined and divided into five grades.On this basis, the paper not only applies the set pair analysis theory and fuzzy theory to the power transformer state evaluation, but also from each step of the transformer state evaluation, proposes a subjective and objective integrated weighting algorithm.This algorithm synthetically considers the subjective expert experience and objective facts, and makes the index weight of transformer condition evaluation more reliable and reasonable.The effect of expert experience on transformer condition evaluation is fully reflected by the five-grade scale method in subjective weighting method, and the function of the change of index data in determining the weight of transformer is fully reflected by the entropy value method in objective weighting method.In accordance with the objective facts, the principle of least square method is used to synthesize the subjective and objective weights, and the index weights are theoretically the most scientific.The result of the final evaluation is the most accurate compared with other methods.This improved evaluation algorithm provides a strong guarantee for the condition maintenance of transformers.) the paper studies the state prediction of power transformers.In this paper, data preprocessing and prediction model in the process of state prediction are studied.The innovation in data preprocessing is that the average weakening operator is used to process the characteristic gas data of power transformer, and the smoothness of the characteristic gas data of power transformer is more in line with the requirement of prediction model.The accuracy of the prediction is relatively improved.The improvement of the prediction model is that the multi-parameter grey prediction model MGM1N) is used to predict the characteristic gases of power transformers in groups.The grouping method is based on the three-ratio diagnosis method commonly used in fault diagnosis of power transformers. This method can avoid the correlation analysis needed for the prediction of multi-parameter grey prediction model and simplify the calculation steps of the model.Computational time is also saved.Through the analysis of the experimental results, it can be seen that the average weakening multi-parameter grey prediction model based on the three-ratio method is more accurate in forecasting and can grasp the regularity of the system better.Finally, the state of power transformer is evaluated by using the theory of ideal point solution and the predicted data. The results show that it is true and reliable.
【学位授予单位】：青岛科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM41

【参考文献】