Turnout jamming fault intelligent early warning method based on power numerical analysis

Abstract

The invention provides a turnout jamming fault intelligent early warning method based on power numerical analysis. The turnout jamming fault intelligent early warning method comprises the following specific steps: S1, collecting a large number of non-fault characteristic power numerical values in advance to establish a sample analysis set {F} with the capacity of K2; S2, sequentially solving a reference early warning upper limit value, an early warning lower limit value and a reference early warning intermediate value; S3, taking a limited number of characteristic power values generated by continuous movement at present as an input sequence during actual operation; S4, mining and fitting a change rule of the input sequence and predicting a power value sequence of a next time period; S5, calculating the probability that the predicted power value of the next section is in a three-level abnormal value interval; and S6, constructing a probability early warning decision tree according to the obtained probability value and the size relationship, and outputting multi-stage early warning. According to the method, feature calculation does not need to be carried out on the turnout curve, theeasily-obtained power numerical value serves as an analysis object, the change rule of the power numerical value is mined and fitted, the power numerical value in a subsequent period is effectively predicted, a multi-stage probability early warning model is established, and the problems existing in an existing method are relieved.

Description

technical field [0001] The invention belongs to the field of early warning of turnout faults, and in particular relates to an intelligent early warning method for blockage faults of turnouts based on power numerical analysis, and is especially suitable for electric turnouts such as urban rail transit S700K. Background technique [0002] Turnouts and their operating devices are the most basic equipment used for track switching, and play an important role in maintaining the safety of rail transit. Whether it is a high-speed railway or an urban rail transit line, the frequency of turnout movements in important areas or stations is very high. The long-term moving back and forth and the vibration caused by the train passing the turnout can easily cause wear, displacement, and deformation of the joint parts of the turnout to occur imperceptibly. Strong hidden dangers, resulting in mechanical jamming failure. [0003] At present, the common real-time turnout jamming fault warning ...

AI Technical Summary

Problems solved by technology

The mainstream schemes of intelligent early warning that have been proposed at home and abroad are based on algorithms such as expert systems, support vector machines, and artificial neural networks. There are mainly two methods based on learning and statistics, both of which have a serious imbalance in the proportion of positive and negative samples (the actual number of one-way action failures of a turnout is only a small probability event compared with the number of normal actions) and the problem that the samples are difficult to meet strict distribution conditions

The learning-based method still has problems such as the construction selection of core control parameters such as kernel function and penalty factor, which needs to be determined subjectively, and the complexity of large-scale sample training is high.

(2) Most of the features are analyzed and compared with an action curve that appears at the current moment, without comprehensive consideration of the changes in the previous continuous action curves, which may easily lead to misjudgment and missed judgment

(3) It is impossible to predict the state of the subsequent period of time, and give early warning according to priority

(4) It is often necessary to perform feature extraction and calculation for the turnout curve. Unreasonable feature extraction will easily lead to large judgment errors, and more and more complex feature selection may lead to low calculation efficiency and poor practicality.

Images