Overview

A deep learning based customized service for Fast and accurate Time Series data prediction.

Deep Learning and Prediction Process

Features

Time series data preprocessing

Process raw time series data for more effective deep learning.

Future phenomena forecasting

Predict future outcomes of a system without real experiments.

Optimal deep learning method

Apply the most appropriate state-of-the-art deep learning methods for time series prediction.

Time series prediction tools for easy application

 · Provide customers with customized tools for easy application.
· Prediction simply requires executing the tool and entering a design of interest.

Effects

Enhance time series prediction accuracy

Supports fast and accurate time series prediction using archived data of various patterns.

Enhance practical applicability

Customized prediction tool allows easier application to practical problems.

Applications

Remaining useful life prediction for Li-ion Batteries

Objective
State of health (SOH) and remaining useful life (RUL) prediction using a deep learning model trained on corresponding archived time series data.

Input Data
Time series data before reaching end of life (EOL)

BruceTS Predictions
· State of Health(SOH)
· Remaining Useful Life(RUL)

Deep Learning Method
TCN(Temporal Convolutional Networks)