Study In Control Algorithm For Fuel Heater Based On Neural Network Model

Vehicle Heater is one kind of detached system of mini internal-combustion engine, used for carrying out the fuel oil preheating on the engine and driver’s cab working in low temperature condition.It is composed of the several major components such as fuel supplying, lighting, burning and heat exchange system, water circulation, warm wind circulation, and electronic control unit (ECU).The ECU design is a foremost technology in the entire heater design specifically for the project" YJHQ11.6/1fuel heater’s ECU designing ".The hardware system of heater ECU is made up of five parts:data collecting、drive of implement、communication、self-inspection and a central processor. In the course of work, ECU’s major task includes:real-time sampling of parameters of water temperature and combustion chamber temperature by sensor, driving windmotor, oilpump, waterpump, spark according to predetermined order; Communication with Host control system such as remote control, timer, checking device, accept control command, upload current information together. Checking wind motor and other implement in the process of starting and operation to correct response in time. This design adopt ATmega16as MCU, adopt inside ADC for water temperature and combustion chamber temperature ADC conversion; adopt the BTS441R power switch to drive implements, adopt the MAX232to exchange data with Host. Under hardware system support, not only ECU being able to ensure that the heater work regularly, and can protect itself under improper condition such as implements or sensor having open circuit, short circuit, over current, over voltage and other abnormality event, avoiding a fortuitous accident thereby.Software designed for ECU major include step sequence, sampling, communication, water pump and Sparkplug’s open-loop control, windmotor and oilpump’s closed-loop control, and self-checking, ignition distinguish arithmetic, flameout protection procedures modules. The core technology are self-checking, ignition distinguish arithmetic, flameout protection procedures. This work, using combustion chamber temperature, combustion chamber temperature gradient, cylinder temperature parameters as input parameters based on neural network and the single neuron perceptron discriminant function to constructe ignition distinguish; using combustion chamber temperature, combustion chamber temperature gradient as input parameters based on single neuron perceptron discriminant function to constructe flameout discrimination protection classification algorithm, using late Henan Construction of the self-checking algorithm, Using PWM mode to adjust the windmotor and sparkplug’s voltage, using PFM mode to adjust the frequency of oilpump; PD algorithm is based on the closed-loop control of the oil pump. On this basis, the heater control functions to meet the similar foreign products level of control.The laboratory tests, Vehicle testing result indicates the Vehicle Heater achieves the capability as the same product abroad. As we know, it’s the first time to use self-checking arithmetic, ignition distinguish arithmetic and flameout protection arithmetic to design Vehicle Heater in our country.This work imbursed by "YJHQ11.6/1fuel heater’s ECU designing" of LLYCO Vehicle Heater(ChangChun), Ltd.