Fuzzy temperature compensation scheme for hot wire mass airflow sensor
Thermal flow measuring technology has come a long way since the introduction of thermocouple technology and early hot wire anemometers. Thermal technologies depend on heat transfer and traditionally operate on differential temperature measurements between two temperature sensitive materials to gener...
| Summary: | Thermal flow measuring technology has come a long way since the introduction of thermocouple technology and early hot wire anemometers. Thermal technologies depend on heat transfer and traditionally operate on differential temperature measurements between two temperature sensitive materials to generate a signal directly proportional to the temperature differential and mass flow rate. In this thesis, the development of an open-loop Fuzzy Temperature Compensation Scheme (FTCS) for Hot Wire Mass Air Flow (MAF) Sensor is presented. The FTCS for Hot Wire MAF Sensor is used in automotive application to measure the volume and density of air entering the engine at any given time. The Electronic Control Unit (ECU) uses this information in conjunction with input from other sensors to calculate the correct amount of fuel to deliver to the engine and also used indirectly to help calculate desired ignition timing and transmission operating strategies. This FTCS used to compensate the error occurred for the Hot Wire MAF Sensor measurement caused by the temperature variation in the air. The data collection for Hot Wire MAF Sensor inaccuracy analysis is done using NI PCI 6251 DAQ, NI Elvis Board and LABVIEW software. Based on the collected data, the absolute error and percentage error for the sensor output voltage have been calculated compared to the output voltage for the standard temperature value. Then, based on the offset error, six rules for Fuzzy Inference System (FIS) have been developed. The Sugeno type FIS is used for the FTCS design. In order to verify the performance of the proposed Hot Wire MAF Sensor temperature compensation scheme, first a simulation model is developed using Matlab/Simulink. The effectiveness of the proposed fuzzy compensation scheme is verified at different temperature variations compared with Radial Basis Function Neural Network (RBFNN) Temperature Compensation Scheme. Then, based on the Matlab/Simulink simulation, the FTCS has been implemented in real-time using Digital Signal Controllers, dsPIC30F4013 with the Programming C Language. In this regard, a performance comparison of the output voltage of the Hot Wire MAF Sensor after compensated using FTCS, RBFNN Temperature Compensation Scheme and without compensates is provided. These comparison results demonstrate the better improvement for the Hot Wire MAF Sensor measurement accuracy with the estimation percentage error after compensation is only within 0.8451 % of full-scale value. |
|---|