The paper discusses the challenge of evaluating the prognosis quality of machine health index (HI) data. Many existing solutions in machine health forecasting involve visually assessing the quality of predictions to roughly gauge the similarity between predicted and actual samples, lacking precise measures or decisions. In this paper, we introduce a universal procedure with multiple variants and criteria. The overarching concept involves comparing predicted data with true HI time series, but each procedure variant has a specific pattern determined through statistical analysis. Additionally, a statistically established threshold is employed to classify the result as either a reliable or non-reliable prognosis. The criteria include both simple measures (MSE, MAPE) and more advanced ones (Space quantiles-inclusion factor, Kupiec's POF, and TUFF statistics). Depending on the criterion chosen, the pattern and decision-making process vary. To illustrate effectiveness, we apply the proposed procedure to HI data sourced from the literature, covering both warning (linear degradation) and critical (exponential degradation) stages. While the method yields a binary output, there is potential for extension to a multi-class classification. Furthermore, experienced users can use the quality measure expressed in percentage for more in-depth analysis.
This paper proposes \textit{layer fusion} - a model compression technique that discovers which weights to combine and then fuses weights of similar fully-connected, convolutional and attention layers. Layer fusion can significantly reduce the number of layers of the original network with little additional computation overhead, while maintaining competitive performance. From experiments on CIFAR-10, we find that various deep convolution neural networks can remain within 2\% accuracy points of the original networks up to a compression ratio of 3.33 when iteratively retrained with layer fusion. For experiments on the WikiText-2 language modelling dataset where pretrained transformer models are used, we achieve compression that leads to a network that is 20\% of its original size while being within 5 perplexity points of the original network. We also find that other well-established compression techniques can achieve competitive performance when compared to their original networks given a sufficient number of retraining steps. Generally, we observe a clear inflection point in performance as the amount of compression increases, suggesting a bound on the amount of compression that can be achieved before an exponential degradation in performance.
Many approaches for estimation of Remaining Useful Life (RUL) of a machine, using its operational sensor data, make assumptions about how a system degrades or a fault evolves, e.g., exponential degradation. However, in many domains degradation may not follow a pattern. We propose a Long Short Term Memory based Encoder-Decoder (LSTM-ED) scheme to obtain an unsupervised health index (HI) for a system using multi-sensor time-series data. LSTM-ED is trained to reconstruct the time-series corresponding to healthy state of a system. The reconstruction error is used to compute HI which is then used for RUL estimation. We evaluate our approach on publicly available Turbofan Engine and Milling Machine datasets. We also present results on a real-world industry dataset from a pulverizer mill where we find significant correlation between LSTM-ED based HI and maintenance costs.