Deep Learning Engine Knock Detection Using Low Fidelity Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High fidelity engine knock detection sensors are expensive and difficult to install due to their requirement for high temperatures and pressures, while low fidelity sensors sacrifice accuracy for ease of installation and cost-effectiveness.
Innovation Solution
A system and method using a deep learning system trained with data from both high fidelity pressure sensors in the combustion chamber and low fidelity vibration sensors on the engine block to accurately detect engine knock, allowing for the fidelity of high fidelity sensors without the high costs and installation complexities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high fidelity knock detection sensors are used in the combustion chamber, then measurement precision is improved, but device complexity and installation cost increase
Solution Approach 1:
The patent creates a virtual copy of the high fidelity sensor's measurement capability by training a deep learning model to map low fidelity sensor readings to the equivalent high fidelity sensor outputs. The model learns the relationship between engine block vibrations and combustion chamber pressure, enabling the low fidelity sensor to indirectly provide high fidelity measurement accuracy without physical installation in the combustion chamber.
Solution Approach 2:
The patent replaces the mechanical/physical high fidelity pressure sensor installation in the combustion chamber with an electronic/software-based deep learning system. Instead of physically installing complex sensors that require special handling and positioning, the solution uses computational algorithms running on existing electronic control units to achieve equivalent measurement precision through data processing and pattern recognition.
2Measurement precision
If high fidelity knock detection sensors are installed in the combustion chamber, then measurement precision is improved, but installation cost increases
Solution Approach 1:
The patent creates a virtual copy of the high fidelity sensor's measurement capability by training a deep learning model to map low fidelity sensor readings to the equivalent high fidelity sensor outputs. The model learns the relationship between engine block vibrations and combustion chamber pressure, enabling the low fidelity sensor to indirectly provide high fidelity measurement accuracy without physical installation in the combustion chamber.
Solution Approach 2:
The patent replaces expensive, difficult-to-install high fidelity sensors with inexpensive, easily-installed low fidelity sensors. The low fidelity sensors are described as 'cheap' and 'easy to install' compared to high fidelity sensors, and the patent accepts this trade-off by compensating through the deep learning processing system that extracts high-value information from the lower-cost sensor data.
3Ease of manufacture
If low fidelity knock detection sensors are used, then installation cost is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent performs preliminary action by training the deep learning model offline using paired data from both high fidelity and low fidelity sensors. This pre-training phase establishes the mapping relationship between the two sensor types before actual knock detection begins. The model learns to translate low fidelity readings into accurate knock detection signals in advance, so that during operation, the low fidelity sensor immediately provides high fidelity accuracy without real-time conversion delays.
Solution Approach 2:
The patent uses feedback by training the deep learning model with ground truth data from high fidelity sensors to continuously improve the accuracy of knock detection. The model learns from the difference between low fidelity sensor readings and actual high fidelity measurements, adjusting its internal parameters to minimize detection errors. This feedback mechanism ensures that the low fidelity sensor system converges to high fidelity performance levels.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate detection of engine knock with the precision of high fidelity sensors using low fidelity sensors, reducing installation costs and maintaining performance.
Implementation Method 1
The second data represents vibrations of an engine block
Implementation Method 2
The first data represents pressure in a combustion chamber of the engine
Data Source
AI summary
A system for training a deep learning system to detect engine knock with accuracy associated with high fidelity knock detection sensors despite using data from a low fidelity knock detection sensor. The system includes an engine, a high fidelity knock detection sensor, a low fidelity knock detection sensor, and an electronic processor. The electronic processor is configured to receive first data from the high fidelity knock detection sensor. The electronic processor is also configured to receive second data from the low fidelity knock detection sensor. The electronic processor is further configured to map the first data to the second data, train the deep learning system, using training data including the mapped data, to determine a predicted peak pressure using data from the low fidelity knock detection sensor, receive third data from the low fidelity knock detection sensor, and using the third data, determine the predicted peak pressure.