Abstract:Extraction rate and energy consumption are two important indicators of the sugarcane milling system, which are calculated by offline experiments and inspection methods. Soft sensors are inferential models, which are enabling online prediction of these indicators. Selection of input variables is one of the most critical issues in soft sensors design. So a deep feature extraction method is proposed by combining mutual information theory and hybrid chicken swamp algorithm to determine the input and parameters of model. Then, input the obtained variables into the data-driven model of deep kernal extreme learning machine (DK-ELM) with optimized parameters to realize the prediction of extraction and hourly energy consumption. It combines the kernel method with the multi-layer ELM. Experiments are conducted to verify the effectiveness and superiority of the modeling method. Comparing the proposed framework with the four experiments designed on a sugarcane milling system, the proposed framework for predicting extraction and energy consumption with better feasibility and effectiveness, and provided a good reference for end-point control and judgment of quick direct tapping. Practical Applications Extraction and energy consumption are two important indicators of the milling process and whether they meet standards will affect the smooth operation and economic benefits of sugar production. This study presents a novel modeling method to predict these indicators based on the deep feature extraction and deep kernal extreme learning machine. These methods provided a good reference for end-point control and judgment of quick direct tapping. In addition, it a new concept for the design of soft sensors for such complex industrial processes.

Dynamic Feature Extractor for Industrial Key Variable Prediction Based on Enhanced Gating Mechanism

Gated Quality-Driven Feature Extraction with A Hybrid SIAE Model for Industrial Soft Sensor

Long-short Memory with Feature Representation by Latent-variable Model for Industrial Soft Sensor

DAMPNN: Dynamic Adaptive Message Passing Neural Network for Industrial Soft Sensor

Automatic Deep Extraction of Robust Dynamic Features for Industrial Big Data Modeling and Soft Sensor Application

Stacked Spatial–Temporal Autoencoder for Quality Prediction in Industrial Processes

Gated Stacked Target-Related Autoencoder: A Novel Deep Feature Extraction and Layerwise Ensemble Method for Industrial Soft Sensor Application

Cooperative Deep Dynamic Feature Extraction and Variable Time-Delay Estimation for Industrial Quality Prediction

A Faster Dynamic Feature Extractor and Its Application to Industrial Quality Prediction

Dynamic Historical Information Incorporated Attention Deep Learning Model for Industrial Soft Sensor Modeling

Soft-Sensor Development Through Deep Learning with Spatial and Temporal Feature Extraction of Complex Processes

Dynamic Time Features Expanding and Extracting Method for Prediction Model of Sintering Process Quality Index

Dual Temporal Attention Mechanism-based Convolutional LSTM Model for Industrial Dynamic Soft Sensor

Weighted Nonlinear Dynamic System for Deep Extraction of Nonlinear Dynamic Latent Variables and Industrial Application

Deep Learning for Industrial KPI Prediction: when Ensemble Learning Meets Semi-Supervised Data

Dynamic Features Incorporated Locally Weighted Deep Learning Model for Soft Sensor Development.

Soft sensor modeling based on DICA-SVR

Soft Sensor with Deep Feature Extraction for a Sugarcane Milling System

Dynamic Global Feature Extraction and Importance‐correlation Selection for the Prediction of Concentrate Copper Grade and Recovery Rate

Deep Nonlinear Dynamic Feature Extraction for Quality Prediction Based on Spatiotemporal Neighborhood Preserving SAE

A Graph-Based Time-Frequency Two-Stream Network for Multistep Prediction of Key Performance Indicators in Industrial Processes