Smart Leather Speed ​​Body Drives a New Threading System

In the sliding grinding stage, the target sliding stage is controlled by the throttle opening to reflect the intention of the driver, and the driver intends to quantify the engine target speed. The principle is to select the engine speed as the engine target speed eon when the maximum torque at a throttle opening is selected. The difference between the actual engine speed and the target speed is the clutch engagement speed v as the output. If the actual speed is higher than the target speed, the engagement is increased to increase the engine load and the actual speed is decreased; otherwise, the engagement is reduced, the engine load is reduced, and the actual speed is increased. Through the performance test of the 7101 Antelope sedan engine, the engine speed is selected as the target speed when the maximum torque of one throttle is selected.

AMT vehicle fuzzy neural network control model control system describes the use of fuzzy neural network control engine actual speed and target speed difference eonΔ and its rate of change dtndeo /) (Δ to control the clutch engagement speed v, to meet the requirements of vehicle start, control system structure As shown, where ne is the actual engine speed and nc is the driven disk speed.

At this point, the fuzzy control model for starting the vehicle has been established. However, in practical application, the selection of fuzzy membership parameters ia and ib is relatively large, resulting in the fact that the control law cannot objectively reflect the actual situation, and the membership function must be optimized. With self-adaptive self-adaptive technology of neural network, test data (training samples) are obtained from the sensor of the vehicle start control system, and the initially determined fuzzy controller is trained with data, and fuzzy rules and membership functions are revised according to the training results; The shape of the membership function of the fuzzy system is adjusted according to changes in the external environment so as to eliminate human influence and enable it to have a true ability to adapt to the external environment.

Design of fuzzy neural network controller Fuzzy neural network controller structure The fuzzy neural network system is based on the Takagi-Sugeno model of fuzzy inference system. It formulates the fuzzy logic rules and membership function parameters in the fuzzy inference system through neural network self-learning. And adjustment, and automatically modify the membership function, effectively overcome the fuzzy control system membership function of human factors, lack of self-learning ability, so that the system has a fast convergence speed, small error, the required training sample less features. Fuzzy neural network system is very suitable for controlling complex time-varying and non-pure system. This article uses a 5-layer dual-input, single-output fuzzy neural network system, as shown.

eonΔ first layer second layer third layer fourth layer fifth layer dtndeo/) (Δv fuzzy neural network structure is the input layer, the speed difference eonΔ and speed difference change rate dtndeo/) (Δ as input, The number of nodes in the layer is 2.

The second layer is the fuzzification layer, and the selection error and error change language domain is {negative, negative, negative, zero, positive, median, and positive) denoted by the symbol {NB,NM,NS,0,PS , PM, PB}. Gaussian function as a membership function. The third layer is the rule layer, which corresponds to fuzzy reasoning and calculates the fitness of each rule. The number of nodes corresponding to this layer is 49. The fourth layer is normalized. The fifth layer uses the center method to defuzzify the output v.

Because of the slow learning speed of the BP algorithm and the fact that it easily falls into local minimum, the network training can use the improved BP algorithm with the momentum item as the learning algorithm to train the network weights. Through training and learning, the membership function parameters are continuously optimized.

The neural network model is constructed according to the above form, and the ANFIS network model is trained and learned using the fuzzy control data as a training sample. Using MATLAB to perform network training operations, the training accuracy is 0.002, and the learning rate is 0.01. Finally, the mapping relationship between the input and output is the mapping relationship of the fuzzy control rule base.

Simulation Analysis Membership Function Optimization In the above model training process, the membership function we chose was a Gaussian function. After training, the input variable speed difference eonΔ and speed difference change rate dtndeo/) (the membership functions of Δ are shown respectively. a is the membership function before and after eonΔ training, b is dtndeo/) (Δ membership function before and after training, dashed line The membership function before training, the solid line shows the membership function after training.

AMT vehicle simulation experiment In order to verify the effect of the optimization model, an AMT vehicle dynamics model was established using Simulink. The main structural parameters are as shown. The optimized fuzzy neural network system algorithm is put into the vehicle dynamics model for simulation and the simulation results are obtained.

SC7101 Sedan Parameters Parameter Value Clutch Pressure plate Effective radius (mm) 74 Friction material factor 0.25 Full load mass (kg) 1399 Tire working radius (mm) 1190 Transmission system speed ratio 0.274 Mechanical efficiency 0.90 When the oil degree opening β = 25% Fuzzy neural network system control and general fuzzy control engine speed simulation results. As can be seen, the fuzzy neural network controller adopts a better tracking characteristic than an ordinary fuzzy controller. Its overshoot is small, the adjustment time is short, and the system quickly enters a steady state, which can better reflect the driver's intention.

Conclusion (1) The starting control strategy of automatic transmission vehicle realizes that it can use fuzzy neural network control to solve the common fuzzy control of vehicle starting. It is influenced by human factors more, difficult to adjust and other difficulties, reflects the intention of the driver, and improves the stability of the starting process. .

(2) Using neural network to optimize the established fuzzy control system and carry out the whole vehicle dynamic simulation. The result shows that the fuzzy neural network can achieve good results through a small amount of debugging and training on the basis of ordinary fuzzy control. It has a good application prospect in the vehicle start control method.

Five furrow reversilbe plow.

Combination of force, color, and beauty.

Product name	Hydraulic Reversible Plow
Product model	1LFT-450	1LFT-550
Dimension(mm)	4300*1700*1800	5360*2000*1800
Structure mass(kg)	1430	2000
Auxiliary power	118-154	154-206
Operationg width(cm)	140-180(adjustable)	175-225(adjustable)
Number of plough bodies(Units)	4*2	(4+1)*2
Operationg width of plough body(cm)	500
Operating tillage depth(cm)	18-35
Longitudinal spacing of plough body(cm)	100
Operationg speed(km/h)	6--10
Production efficiency(Mu/h)	13-27	16-34
Products meet the implementation of the standard GB/T14225-2008 Moldboard Plough
Production efficiency is calculated according to the therory data of the operating speed 6-10km/h

Five Furrow Turnover Plough

Five Furrow Turnover Plough,Multifunctional Turnover Furrow Plough,Tractors Reversible Mouldboard Plough,Disc Reversible Rotary Plough

Shandong Gold Dafeng Machinery Co., Ltd , http://www.golddafengharvester.com