Hemp Decortication Control Using Sensor Feedback and Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The decortication process of biomass products, such as hemp straw, is inefficient due to reliance on manual quality assessments and subjective selection methods, leading to high processing costs and variability in product quality.
Innovation Solution
A system and method that utilizes machine learning models trained on sensor data from decorticating, cutting, blending, and pulverizing processes to predict the yield and quality of bast, hurd, and dust fractions, allowing for automated selection and processing adjustments to achieve desired product properties, including the use of thermoplastic polymers and spectrometry analysis for material valuation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual quality assessment and subjective selection methods are used, then the decortication process can be operated with simple equipment, but the processing efficiency is low and product quality varies
Solution Approach 1:
The system performs preliminary analysis of biomass characteristics using sensors and spectrometry before decortication processing. This allows the machine learning model to predict optimal processing parameters and select appropriate input materials in advance, improving processing efficiency while maintaining controlled system complexity through automated decision-making
Solution Approach 2:
The system implements closed-loop feedback by continuously monitoring processing parameters and product quality, using sensor data to train and refine the machine learning model. This feedback mechanism enables real-time optimization of processing conditions, increasing productivity while the system learns to operate more efficiently without proportionally increasing complexity
2Manufacturing precision
If automated machine learning-based selection and processing adjustments are implemented, then product quality consistency improves, but system complexity and initial processing costs increase
Solution Approach 1:
The system replaces manual quality assessment and subjective selection with automated sensor-based detection and machine learning algorithms. Spectrometry sensors and other detection devices objectively measure biomass characteristics, eliminating human subjectivity and achieving consistent product quality through data-driven decision-making rather than mechanical or manual processes
Solution Approach 2:
The system dynamically adjusts processing parameters based on real-time sensor data and machine learning predictions. By changing operational parameters such as cutting dimensions, opening intensity, and decortication settings according to measured input material properties, the system maintains high product quality consistency while adapting to variations in biomass feedstock
3Productivity
If comprehensive sensor data collection and machine learning training are performed, then processing optimization improves, but data processing time and computational resources increase
Solution Approach 1:
The system collects and pre-processes sensor data during the decortication process itself, training and updating the machine learning model with real-time information. This preliminary data collection and continuous learning approach allows the system to optimize processing parameters without requiring extensive separate data processing time, as the learning occurs concurrently with production
Solution Approach 2:
The machine learning model operates continuously, constantly receiving sensor data and adjusting processing parameters in real-time throughout the decortication process. This continuous optimization eliminates idle data processing time, as the system continuously learns and adapts without interruption to production flow, maintaining both high productivity and processing optimization
Data Source
AI summary
A system, controller, and method for decortication processing on one or more input units of hemp into one or more resultant products. The method includes: analyzing one or more characteristics of the input units; cutting the input units into a predetermined size; opening the cut input units; performing decortication on the opened input units to separate the hemp into components, the components including bast, fibre, and hurd; densifying the fibre into bales; pulverizing the hurd and bast; combining the pulverized hurd and bast with thermoplastic polymers into a resultant product; receiving analyzer data from at least one of the decortication, the densifying, the pulverizing, and the combining; training a machine learning model based on the analyzer data; using the trained machine learning model to adjust one or more aspects to achieve a desired resultant product.