Q2. What future works have the authors mentioned in the paper "Remaining useful life estimation in prognostics using deep convolution neural networks" ?
While good experimental results have been obtained by the proposed method, further architecture optimization is still necessary, since the current training time is longer than most shallow networks in the literature. Deep learning methods generally suffer from high computing load, and that will be focused on in further research. Efforts should be made on including the score function into the loss function of the neural network in the future.
Q3. What are the two metrics used for evaluating the performance of the proposed prognostic method?
In this study, 2 metrics have been used for evaluating the performance of the proposed prognostic method, i.e. scoring function and root mean square error.
Q4. What is the advantage of applying neural networks on prognostic and health management?
The advantage of applying neural networks on prognostic and health management lies in that highly nonlinear, complex, multi-dimensional system can be well modeled without prior expertise on the system physical behavior.
Q5. What is the way to improve the prognostic performance of the proposed method?
To further improve the prognostic performance, a fine-tuning process using the back-propagation (BP) algorithm is applied [39], where the parameters of the proposed model are updated to minimize the training error.
Q6. Who built a deep convolution neural network to predict the RUL of system?
Babu et al. [31] built a 2-dimensional (2D) deep convolution neural network to predict the RUL of system based on normalized variate time series from sensor signals, where one dimension of the 2D input is the number of sensors.
Q7. What is the main reason why the engine is not evaluated in the late period?
A good evaluation of the engine status in the late period is able to enhance operation reliability and safety, reduce maintenance costs and improve the whole system performance.
Q8. What are the main areas of engineering maintenance and prognostics?
Engineering maintenance and prognostics are very crucial in many industry areas such as aerospace, manufacturing, automotive, heavy industry and so forth.
Q9. What is the main reason why the RUL estimation is so promising?
As an improvement of the traditional RNN, a long short term memory (LSTM) based neural network scheme was proposed by Yuan et al. [22] for RUL estimation of aero-engines in the cases of complicated operations, hybrid faults and strong noises.
Q10. How can the deep CNN structure be used to estimate the RUL of a machine?
High-level abstract features can be successfully extracted by the deep CNN architecture, and the associated RUL value can be estimated based on the learned representations.
Q11. Why is long-short term memory preferred by many researchers?
as a variant of RNN, long-short term memory method is prefered by many researchers to prevent backpropagated errors from vanishing or exploding [46].
Q12. Why is the engine unit near failure enhanced?
That is because when the engine unit is working close to failure, the fault feature is enhanced and that can be captured by the proposed network for better prognostics.