A Low-Code Edge Computing-Based Predicting Scheme for Aeroengine Components to Enable Intelligent Aeronautical Manufacture

Even if the Big Model/Big Data technologies give a chance that deals with large-scale industrial problems, it has brought a huge demand for computing power. This cannot be provided in most industrial computing scenarios, while low-code and low-cost computing schemes are indispensable, especially on account of the edge computing architecture. This article focuses on aeroengine parameters prediction and proposes an edge computing-based low-code predicting framework that divides the predicting framework into three functional layers. It supports hierarchical and scalable prediction. Based on the proposed framework, this article proposes a predicting scheme based on Bayesian Ridge Regression. In particular, it contains as well a data cleaning procedure based on local outlier factor, to enhance predicting accuracy. Compared with the normal RR, KNN, ELM, LSTM, and GRU approaches, the proposed predicting scheme achieves the highest predicting accuracy on an evaluated edge server. The purpose of this work aims at providing a low-code and low-cost approach to support timely and accurate prediction for aeroengine performance parameters.