Engine Fluid Flow Estimation Using Hydraulic Loss Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining fluid flow rates in vehicle engine systems, particularly when sensors are unavailable, are inefficient and require tedious calibration table construction due to changes in injector flow cross section caused by deposits or wear.
Innovation Solution
A method utilizing a supervised-learning loss estimation module, such as a neural network, to calculate a hydraulic loss coefficient based on pressure and temperature parameters, allowing for accurate estimation of fluid flow rates without direct measurement, and enabling predictive maintenance through alert thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional algorithmic logic with calibration tables is used to estimate flow rate, then the system can account for injector wear and deposits, but the process becomes tedious and complex due to the need to construct and maintain calibration tables
Solution Approach 1:
The patent replaces the mechanical/calibration-based system with a neural network-based intelligent system. The neural network learns the complex relationship between injection parameters and actual flow rates automatically, eliminating the need for manual calibration table construction and maintenance while maintaining high estimation accuracy
Solution Approach 2:
The neural network performs self-learning and self-adjustment by processing injection data automatically. It continuously improves its estimation capability without requiring external calibration table construction, making the system self-sufficient and reducing operational complexity
2Measurement precision
If direct sensor measurement of flow rate is implemented, then accurate real-time flow rate data is obtained, but the system complexity and cost increase due to additional hardware requirements
Solution Approach 1:
The patent introduces a neural network as an intermediary computational model that indirectly estimates flow rate from easily measurable parameters (injection pressure, temperature, duration). This intermediary approach avoids the need for direct flow rate sensors while achieving comparable accuracy through intelligent inference
Solution Approach 2:
The patent substitutes physical measurement hardware (flow rate sensors) with an intelligent software-based estimation system. The neural network processes readily available sensor data to infer flow rate, replacing complex hardware measurement systems with a computationally efficient software solution
3Adaptability or versatility
If calibration tables are constructed to account for injector changes over time, then the estimation adapts to wear and deposits, but the process requires significant time and resources for data collection and table construction
Solution Approach 1:
The neural network is pre-trained with comprehensive data covering various injector conditions (new, worn, with deposits). This preliminary training action equips the system with adaptive capability beforehand, eliminating the need for time-consuming calibration table construction when injector changes occur during operation
Solution Approach 2:
The neural network provides dynamic adaptation capability that automatically adjusts to changing injector conditions in real-time. Unlike static calibration tables that require periodic reconstruction, the neural network continuously adapts its predictions based on current operating parameters, providing ongoing versatility without additional time investment
Data Source
AI summary
A system and method for determining a value of a flow rate of a liquid in a vehicle engine system comprising a fluid tank (3), a pump (2), a fluid injector (1), with a fluid flow path from the pump to an injected zone (4), and an electronic control unit (5) for controlling opening of the injector, the method comprising:—providing a loss estimation module (52), supplying as output a hydraulic loss coefficient (CP),—carrying out a plurality of sequences of fluid injection, with values of a plurality of parameters (dP, P1, P0, T, X) being collected,—calculating a theoretical quantity (QTH) of fluid injected during these injection sequences, with the aid of the values of the parameters (P1, P0, T, X),—calculating an estimated actual quantity (QRE) of fluid injected during the injection sequences, by applying the loss coefficient (CP) to the calculation of the theoretical quantity of fluid.