Electronic expansion valve control algorithm for evaporator superheat

The constraints of conditions such as complexity, practical application is not much.

For systems where the model of the evaporator is susceptible to changes in the outside world, good control effects must be obtained, in addition to adaptive algorithms, input or output domain optimization. In online optimization, people are charged with the dual task of identifying system models and optimizing control. These two tasks are directly the contradictions that control theories are difficult to coordinate. Taking into account the safety and real-time requirements of the refrigeration system control, an off-line optimization method is used. 2.1 Optimization objectives and optimization parameters As with other optimization methods, genetic algorithms also have optimization goals, but in genetic algorithms it also has a proprietary name, performance evaluation index, which is used to calculate the adaptation values ​​for each sample. The performance evaluation index is determined according to the solved questions. In the evaporator superheat control system. The selected optimization 1 is marked as the inter-valued deviation, 7 is the sampling period as the sampling times, ie the first sampling, and the engraved; the force=2, 2, 1 is the change of the valve opening; the ice 2 is the weighting coefficient.

The sum of the squares of the superheat deviations can be used as an optimization goal, which can reflect the speed of the superheat response and can reflect the superheat steady-state leftward.

The reason why the square of the valve opening change is used as a performance index is to make the adjustment process smooth by reducing the change in the opening degree of the electronic expansion valve. At 1 o'clock, it helps to increase the life of the electronic expansion valve and its drive circuit.

In recent years, many scholars have used genetic algorithms to automatically generate fuzzy controls. Many methods have been proposed, such as the spoon control rules and the degree of inflammation.

These methods can be categorized into several types of sets of fuzzy control rules set by people. Only the membership functions that generate the fuzzy sets are learned. 2 From the public setting, the numbers, only the concept of learning begins to gain more and more attention. However, the determination of fuzzy control rules and membership degree depends on human experience. When there are many input variables with fuzzy draw or more fuzzy variables defined by the input domain, the number of rule sets that fuzzy control can list is very large. Finding the optimal rules is very difficult and sometimes impossible. The optimization of the current wing fuzzy rules mainly adopts the neural network method. However, due to the network structure and network size of the neural network and the poor convergence of learning, no satisfactory results have been obtained so far. In this paper, the genetic algorithm is introduced into the fuzzy control of the evaporator superheat, so that the fuzzy control has a certain self-learning ability.

1 Fuzzy control of the evaporator superheat fuzzy control system 1. Where, the charge is the degree of deviation of the deviation of the superheat degree deviation of the superheating degree and the output of the actuator, and the degree of superheat which is the most controlled parameter of the corresponding soak paste change respectively. 1 is. 1 The shape of the membership function, such as the angular bell, has little effect on the control performance. The size of the width, that is, the degree of overlap of the membership function, has a greater impact on performance. If the membership function holds the input to these intervals, the fuzzy twist does not output. Cough 1 is not good in convergence; if the membership function is too large, the rules influence each other. Symmetrically fully overlapping angular functions are used as membership numbers. The input input domain and the output domain are respectively selected to have different scale factor centers = 3.5, along = 70, and 沁 = 0.5, and the input domain and output domain are still respectively divided into 7 ranges, and 51 fuzzy operators are defined as negative users. Small, small, factory, and small to find the heart, their affiliation, degree of function, such as Court 2, which, in order to affect the shape of the membership function to be optimized parameters. Fuzzy Control Rule 1. Adopt the Big-Fetch-Small 1 algorithm as the fuzzy reasoning method and the center-of-gravity method as the method of solving the fuzzy.

2 Genetic algorithm optimization Fuzzy control of membership function Genetic algorithm Human is an optimization algorithm that simulates the genetic and genetic mechanisms of biological populations. It is essentially an evolutionary mathematical model along the 9-magnetic inverted line, 笕, regular phase-by-phase 4 fuzzy rules and fuzzy The set-constrained constitutive degree function simultaneously learns fuzzy control rules and fuzzy set membership functions.

For the refrigeration system of a fixed speed compressor, the superheat control of the evaporator can be simplified to a single input/output system. At this time, the corresponding relationship between the superheat degree and the opening degree of the electronic expansion valve is relatively simple, and the fuzzy control rules are also relatively simple. What we care about is only The quantitative relationship between superheat and valve opening. To this end, the selection method is to set the set of fuzzy control rules by the person, and only learn 4 to generate the deformity function of the fuzzy set.

The number of degrees of membership is chosen as the number of degrees of membership of the over-alleged angular overlap. For the membership function of the input domain 0 and the output domain, there are three parameters that can be optimized, namely, 3.

2.2 The step code of genetic algorithm. The key to the application of genetic algorithms is genetic coding, the structure of chromosomes. There are only 3 parameters optimized by the genetic algorithm, ie. The value of the 131 parameter is all 07, and an integer between 07 can be randomly generated, and they are converted into a binary system, ie, the genetic algorithm coding is completed. For example, more parameters are optimized. It is necessary to fully consider the compositional structure among various parameters, as well as the mutual agreement and conditions, and whether or not the optimization process is particularly effective.

Generate initial sample groups. A random number of chromosomes is selected as the initial sample, and the maximum number of iterations is 86610. Here, take the initial sample and find the matching value for each copy. Determining the optimal number of indicators The experiment was performed using the corresponding group optimization parameters for each sample, and the corresponding index values ​​were found. And according to the order of preference and inferior times. Then breed. Finding the Propagation Probability of Each Sample For each sample in the population, select the different number of Dongs as its parent value as the parent sample, and pair them randomly, using crossover and mutation methods to breed offspring.

Crossover operation. A break point is randomly generated in the string of the parental sample 4 and the substrings after the break point are interchanged to generate two child samples, 15 being mutated. Select 7 randomly in the string of each child sample. The newly generated sample of the child is added to the original population and the value of the index function of the new sample is found. Select a good sample from the expanded population to form a new population.

Go to step 3 until the iterative termination condition is reached.

The best sample wood code was converted into an optimization parameter 3 experimental apparatus experimental system 3 in which the components of the refrigeration apparatus and its main parameters were as follows.

The use of voltage 22050, nominal power 745., the working fluid is only 12, the cylinder stroke volume 32.71 (2) condenser fins or air-cooled Zhan diameter, while the outer 3 evaporator is light bare spiral type, immersed in the refrigerant water. The evaporation coil is a light pipe, its inner diameter 13, outer diameter 15 length is 131, the refrigerant water phase is cylindrical, diameter 700 is selected, and 215 direct-acting electronic expansion valve. DC12, power supply, phase stepping motor drive, 12-phase excitation, drive frequency 35 pulse 8, fully closed to the full open number of 320 pulses.

In order to make the evaporator load variable, an electric heater was placed at the bottom of the evaporator to adjust the input of the heater to the regulator to change the heating amount. Heating power is read out by power.

In order to increase the efficiency of the evaporator and maintain the temperature of the refrigerant water evenly and evaporate, a small bar stirrer is placed in the middle of the tank. Start the stirrer before the refrigeration system starts working. The speed of the stirrer motor is not adjustable.

71 In order to prevent the migration of refrigerant during shutdown.

There is an electromagnetic before expansion.

Controller parameters, a proportional coefficient magic Xuan 7 integral coefficient sub coefficients 60 differential coefficient is 1 = 15 found in the experiment, by the control system, a variety of interference effect, superheat sampling with noise. The control does not work properly, so the system actually uses rot control. Control results 5.

2 The dynamic characteristic parameters of the available evaporator are = 8; 3, 8 = 0.28 pulses. Based on the experimental results obtained by the experiment, the use of the empirical formula proposed by Office 5 can be obtained by turning on and off the electric heater of the evaporator side of the evaporator without sudden change of the water temperature. In the experiment, the effect of 1 control was verified by changing the superheat setting. Before the change, the superheat is set at 7, the cumulative number of the expansion valve changes at 120130, the evaporator inlet temperature is 3, the outlet is at 10, right, at, 175, 8, the superheat setting is changed to 9, and the superheat setting is 9. Increased, 1 control to reduce the cumulative number of expansion ceramic pulse to the left and right, the evaporator inlet temperature dropped to 0.5, right, no significant change in the outlet temperature, about 8,8 superheat transition to a new set value. From the overshoot, there is a clear overshoot in the degree of superheat. 4.2 Fuzzy Control Optimized with Genetic Algorithm According to the Optimized Ligature Function. The resulting fuzzy control 2. It is used on the computer 刖 1; language fuzzy control 1 pods off-line. Control experimental results

r Accumulated pulse number 8 Evaporator inlet and outlet temperatures and refrigerant water temperature are the three values ​​necessary for the electronic expansion valve controller program. After they have been amplified by 0595 and 1324, they are input to the SCM, and then sent by the SCM through the mouthpiece. Machine notes from childhood.

4 Experimental Verification 4.1 Routine, 10 Control In order to compare the fuzzy control of other conventional algorithms with the fuzzy control optimized by the genetic algorithm, a conventional, 10 control experiment was performed first.

For easy access, parameters. The approximate number of the transfer number delayed by the response of the superheat of the evaporator and the response of the electron expansion, and the force can be obtained from the response curve of the superheat degree to the step change of the pulse of the electronic expansion valve.

Planting Qi Peng, by fuzzy control of the inlet exit superheat valve opening, a control serial port superheat setting value in the 2500 tree from 7 to 9 changes, the evaporator inlet temperature changed from 4.7 right down to 2.5 right, There is no obvious change in the superheat at the outlet of the evaporator, and the transitional process experiences approximately 100% of the superheat of the experimental system after the superheat of 100,6. The set value changes from 9 to 6 and the temperature of the evaporator port is 2. The right rose to 4.5, right, the evaporator outlet temperature dropped slightly, the transition process went through approximately 18, 8. The control obtained from 6 can be achieved, and the superheat can be quickly and stably controlled.

4.3 Comparison of Two Controls The control of pasting and pasting of a drawing system, for example, 17 is performed on a time coordinate. When the superheat setting value changes from 7 Ao to 9 the superheat of the control rises faster than that of the fuzzy control, and it seems that the overshoot is greater than the fuzzy control. Transition process, controlled valve. There is a clear overshoot in the jinage. The fuzzy control is smaller. After the superheat degree is stable, the superheat value is influenced by the noise of the control system. There is no difference between the 1 control and the fuzzy control, but the change of the opening degree of the electronic expansion valve of the fuzzy control is obviously much smaller, and the superheat degree of the fuzzy control should be changed. 1 smooth control.

5 Conclusions The genetic algorithm was introduced into the fuzzy control of the evaporator superheat degree to optimize the membership function. Get a good control effect. 1 The experiment showed that, due to various interference effects of the control system, the superheat sampling was noisy, making the 10 control not work properly, while the 1 control had better anti-down1 interference capability.

Genetic compatibility optimization fuzzy control. The comparison of the effect of drawing is obvious. 1 The response to control the degree of superheat is faster than the fuzzy control, but the overshoot is greater than the fuzzy control; in the adjustment process, there is obvious overshoot in the open of 1 control, and the fuzzy control is more Small; superheat degree is stable, fuzzy control of the electronic expansion valve opening change is much smaller, indicating the fuzzy control of the superheat change ratio, 1 control of the flat to simplify the genetic algorithm encoding and calculation, only to select the membership function more Many fuzzy control parameters, such as proportional factor control rules, are optimized to achieve better results.

Meng Jianjun. A Study on Adaptive Control of Evaporator Target Superheat in Refrigeration System 0. Xi'an Xi'an Jiaotong University Energy and Power Engineering Author brief introduction Chen Wenyong was born in 1969. He was in Xi'an in 1999 and is currently a postdoctoral fellow of the Institute of Refrigeration and Cryogenic Engineering at Shanghai Jiaotong University. The main research interest in refrigeration systems is automatic control.

Tiger 9 Electric Scooters

Eeve Electric Scooter,Hiboy Scooter,New Electric Scooter,Motorised Scooter

Ningbo Hiley Technology Co.,Ltd , https://www.hiley-mobility.com