KR100195090B1 - Heredity algo-fuzzy theory possess refrigerator temperature controlling and its method - Google Patents

Abstract

The present invention relates to a temperature control device of a refrigerator. Specifically, the temperature distribution in the refrigerator is sensed by using a plurality of temperature sensors or infrared temperature sensors, and the genetic algorithm-fuzzy inference is used in a refrigerator (or freezer). The present invention relates to a refrigerator temperature control apparatus and method using a genetic algorithm-FUZZY inference that quickly adjusts the temperature distribution in the refrigerator to a set temperature by controlling the discharge direction and the amount of cold air supplied.

That is, the temperature control device and method of the refrigerator to which the genetic algorithm-fuzzy inference according to the present invention is applied, the temperature sensor is installed in the left and 1H / 3 right name and the 3H / 4 right and 1H / 3 left name respectively in the refrigerator compartment By controlling the cold air discharge direction while comparing the temperature distribution by learning with the data by inferring the temperature by the genetic algorithm-purge function, the cooling speed is increased and the temperature distribution of the whole refrigerating chamber according to the temperature difference of the refrigerated food is evened. As well as giving effect, there is an outside temperature compensation effect by opening and closing the door.

Description

Refrigerator temperature control device and method using genetic algorithm-FUZZY inference

1 is a side cross-sectional view of a typical refrigerator.

2 is a side cross-sectional view of a refrigerator according to the present invention.

3 is an enlarged cross-sectional view of a cold air duct according to the present invention.

4 is an exploded perspective view of a cold air duct according to the present invention.

5 is an enlarged perspective view of the cold air dispersing means according to the present invention.

6 is a cross-sectional view taken along the line A-A 'of FIG. 2, showing the flow of cold air

7 is a block diagram of a cold storage temperature control device using a conventional genetic algorithm-fuzzy inference.

8 is a block diagram of a temperature control device of a refrigerator to which genetic algorithm-fuzzy inference according to the present invention is applied.

9 is a general flow diagram of genetic algorithm-purge control.

10 is a flowchart of the genetic algorithm-fuzzy control routine of FIG.

11 is an explanatory diagram of the TSK (Takagi-Sugeno-Kang) method, which is one of genetic algorithm-fuzzy inference methods.

12 is a refrigerator obtained by a genetic algorithm-purge control of a cold air discharge control plate during cold air discharge according to the present invention and a characteristic curve showing a temperature drop according to a cooling time by a simple rotation method of a conventional cold air discharge control plate in the cold air discharge method. Curve of temperature drop with cooling time of.

13 is a schematic perspective view showing a temperature measuring point of each shelf for measuring the temperature distribution of the refrigerating compartment.

FIG. 14 is a graph showing a temperature distribution at each measurement point shown in FIG. 13 in the refrigerator to which the conventional genetic algorithm-fuzzy inference is applied.

FIG. 15 is a graph showing the temperature distribution at each measurement point shown in FIG. 13 in the refrigerator to which the genetic algorithm-fuzzy inference according to the present invention is applied. FIG.

16 and 18 are graphs each showing a fuzzy partition structure of an input variable.

17, 19 and 20 are graphs showing the fuzzy partitions of the actual respective membership functions in the TSK-fuzzy inference of FIG.

* Explanation of symbols for main parts of the drawings

1 body 2 intermediate wall

3: freezer 4: cold storage room

5: evaporator 6: cold duct

6a: cold air inlet 6b: grill

6b ', 6b, 6b': cold air discharge hole 6c: diaphragm

6c ', 6c, 6c': cold air outlet hole 7: cold air dispersion means

7a: rotational extraction 7b, 7b ', 7b: cold air dispersion member

7c, 7c ', 7c: disc member 7d, 7d', 7d: braid

7e: groove 8: drive means

8a: drive motor

The present invention relates to a temperature control device of a refrigerator. Specifically, the temperature distribution in the refrigerator is sensed by using a plurality of temperature sensors or infrared temperature sensors, and the genetic algorithm-fuzzy inference is used in a refrigerator (or freezer). The present invention relates to a refrigerator temperature control apparatus and method using a genetic algorithm-FUZZY inference to quickly and evenly distribute the temperature distribution in the refrigerator to a set temperature by controlling the discharge direction and the amount of cold air supplied.

In general, as shown in FIG. 1, the refrigerator has a structure including a main body 10 forming a food storage chamber and a door 20 opening and closing the food storage chamber.

The main body 10 includes a cabinet 11 forming an entire frame, a liner 12 disposed inside the cabinet 11, and a foam 13 filled and foamed between the cabinet 11 and the liner 12. Is made of.

The food storage chamber 32 is separated by the intermediate partition 30 and is divided into a freezing chamber 31 and a refrigerating chamber 32. The evaporator 40 is installed at the rear of the freezing chamber 31 (or the middle partition wall), and a blowing device 50 including a cold fan and a fan motor 52 is installed above the evaporator 40.

In addition, a cold air duct 60 is provided on the rear walls of the freezing chamber 31 and the refrigerating chamber 32 to guide the flow of cold air by the blower 50, and the cold air duct 60 includes a plurality of cold air discharge holes 61. This is formed, and the cold air is appropriately supplied to each compartment of the refrigerating compartment divided into several compartments by the shelf 70.

In the refrigerator configured as described above, cold air is generated by the evaporation action of the evaporator 40, and the generated cold air is forcedly circulated through the freezing chamber 31 or the refrigerating chamber 32 by the blower 50.

That is, as shown in FIG. 1, the cold air generated in the evaporator 40 passes through the cold air duct 60 by the blower 50, and the cold air discharge holes 621 formed in the cold air duct 60. It is supplied to the freezing chamber 41 and the refrigerating chamber 32 through. In addition, the cold air after the heat exchange is sucked and circulated to the evaporator 40 through the front side (door side) of the intermediate partition 30 again.

Therefore, the frozen food in the freezer compartment, the refrigerated food can be stored and stored at an appropriate temperature.

However, in such a refrigerator, in particular, in the refrigerating chamber, since the cold air flows from the upper side to the lower side and passes through the several cold air discharge holes 61 formed in the cold air duct 60, the cold air is discharged only in a predetermined direction. In addition, there is a problem in that the temperature distribution difference between each end of the refrigerating chamber 32 is not large because it can not be discharged simultaneously and evenly supplied to the refrigerating compartment divided into several compartments. That is, in the cold air duct 60, the cold air discharge holes 61 are simply arranged up and down, so that the difference in the amount of cold air flow out in the upper side and the lower side causes a large temperature difference between the upper and lower compartments of the refrigerating chamber, and the cold air is not dispersed. Since it is discharged while simply passing through each cold air discharge hole 61, the left and right temperature difference and the front and rear temperature difference of the refrigerating chamber occurs severely.

Therefore, part of the refrigerating chamber 32 is excessively supplied with cold air to cause overcooling, and some corners have a problem in that the cold air is undersupplied to become relatively high in temperature, thus deteriorating stored foods or preserving luck.

In order to improve such a problem, the applicant's patent application No. 94-12401 shows a new cold air duct. This will be described with reference to FIGS. 2 to 6.

2 is a side cross-sectional view of the refrigerator according to the present invention, FIG. 3 is an enlarged cross-sectional view of the cold air duct according to the present invention, and FIG. 4 is an exploded perspective view of the cold air duct according to the present invention.

The new cold air duct 6 is part of the cold air produced in the evaporator 5 via a distribution hole 2c whose lower side is in communication with the rear of the refrigerating compartment 4, as shown in FIGS. 2 and 3. It has a cold air inflow path 6a at the upper end so that it flows in, and the grille 6b which was detachable was provided in the front. The cold air duct 6 has a predetermined thickness and is immersed so that only the surface thereof is exposed to the rear wall of the refrigerating chamber 4 as shown in FIG.

3 and 4, the grill 6b has several discharge holes 6b ', 6b, 6b' which are spaced apart at predetermined intervals, and before and after the refrigerating chamber 4 described later. It has a hemispherical shape to spread the cold air evenly from side to side.

The front of the cold air duct 6 is sealed by the diaphragm 6c, and the front side of the diaphragm 6c has cold air outlet holes 6c ', 6c, 6c' so that the cold air can flow out of the grill 6b. The discharge holes 6b ', 6b, 6b' are provided so as to communicate with each other.

In addition, the refrigerating compartment 4 is located in front of the diaphragm 6c of the cold air duct 6 through the discharge holes 6b ', 6b and 6b' of the grill 6b of the cold air outlet holes 6c ', 6c and 6c'. The cold air dispersion means 7 which distributes evenly to the side is provided.

The cold air dispersing means 7 is formed in a long rod-like rotating shaft 7a which is pivotally arranged between the diaphragm 6c and the grille 6b of the cold air duct 6, and a plurality of cold air dispersing portions are formed in place of the rotating shaft to inflow cold air. Cold air dispersion members 7b ', 7b, 7b for dispersing cold air flowing out through the cold air outlet holes 6c', 6c, 6c 'communicating with the diaphragm 6c forming the furnace 6a toward the refrigerating chamber 4. Has')

The cold air dispersion members 7b ', 7b, 7b' are arranged on the rotational shaft 7a, and some sections of the disc members 7c, which are located inside the cold air outlet holes 6c ', 6c, 6c' of the diaphragm 6c, 7c ', 7c) and a plurality of grooves 7e through which the cold braids 7d, 7d', 7d of correspondingly pivoted thin plates and the cold air introduced into the outer periphery of the disc member are passed downward with respect to the center of the upper surface of the disc member. Has The groove portion 7e has a groove portion 7e formed in the disc members 7e 'and 7e at the lower side than the groove portion 7e formed in the upper disc member 7c so as to have the same size or to facilitate the downward passage of cold air. It can also be relatively large. In addition, although the groove 7e is shown to be curved in a semicircle, the present invention is not limited thereto, and the groove 7e may be used in various forms.

On the other hand, under the cold air dispersing means 7, a driving means 8 for rotating the cold air dispersing means is provided. This drive means 8 has a drive motor 8a immersed except for the motor shaft 8b below the cold air duct 6. The drive motor 8a is driven when the refrigerator operates and is composed of a stepping motor or a geared motor which can rotate in one direction or control the rotation angle and speed from side to side.

In addition, the driving motor 8a should be rotated at a low speed of 4-8 rpm to evenly distribute the cold air so as to prevent the generation of cold air therein when distributing the cold air to the refrigerating chamber.

In the case of operating the refrigerator configured as described above, cold air is generated in the same manner as the conventional evaporator 5 installed behind the freezer compartment 3, and the generated cold air is passed through the freezer compartment 3 and the refrigerating compartment by the blower 9. Forced circulation

That is, the cold air generated in the evaporator 5 as in the direction shown by the arrows shown in FIGS. 2 and 3 is introduced into the bet formed by the diaphragm 6c of the cold air duct 6 by the blower 9. It flows into the furnace 6a. As shown in FIG. 6, when a part of the cold air flowing in reaches the disc members 7c, 7c ', 7c located relatively above the rotating shaft 7a which is built up on the motor shaft 8b and rotates, Evenly distributed into the refrigerating chamber 4 through the discharge holes 6b ', 6b, 6b' of the grill 6b by the braids 7d, 7d ', 7d of the rotating disk members 7c, 7c', 7c. do. The evenly distributed cold air is heat-exchanged with the refrigerated food and then circulated to the evaporator 5 via the inlet hole 2b of the intermediate wall 2 again.

In order to quickly adjust the internal temperature of the refrigerator as described above and to evenly adjust the temperature distribution, a conventional temperature control apparatus of the refrigerator to which the user algorithm-fuzzy inference, as shown in FIG. 7, is applied. As shown in this figure, the temperature of the refrigerator, particularly the refrigerating chamber, which changes according to factors such as the frequency of the door opening and closing of the user, the outside temperature, the temperature change rate according to the door opening and closing, etc. In the control of the lungs, the genetic algorithm-FUZZY inference is applied to allow the cold air to be supplied quickly and to evenly distribute the temperature while considering the factors of temperature change in the refrigerator according to the outdoor temperature compensation and the food condition. .

However, the apparatus and method for controlling temperature in the refrigerator as described above properly compensates the actual outside air temperature by applying only the genetic algorithm-FUZZY rule as a method of analyzing the input according to each unspecific temperature change factor. Not only did it fail, but it did not satisfy the temperature control in the refrigerator. In other words, the cold air discharge system is forged, so that even if hot food comes in, it cannot concentrate the cold air in the direction, or the temperature difference between each compartment in the refrigerating chamber and the front, rear, left, and right sides of each compartment due to the opening and closing of the door. There was a problem that can not properly discharge cold air.

The present invention was devised to improve the above problems, using genetic algorithm-fuzzy inference, which quickly spreads the temperature distribution in a refrigerator (especially the refrigerating chamber) to a set temperature using genetic algorithm-fuzzy inference. It is an object of the present invention to provide a refrigerator temperature control device and method.

In order to achieve the above object, a refrigerator temperature control apparatus using a genetic algorithm-fuzzy inference according to the present invention, the compressor control means; Damper opening and closing control means; Cold air discharge orientation control means for determining a cold air discharge direction by rotating or swinging the cold air discharge control plate of the disc member; Cold air discharge control plate position sensing means for setting a cold air discharge direction with information applied from the cold air discharge direction control means; Refrigerator compartment temperature sensing means for sensing a temperature of at least four points or more so as to sense a temperature distribution of the entire refrigerator compartment; Temperature change rate calculating means for calculating a temperature change rate of the refrigerating compartment temperature detecting means at a predetermined number or more points of the refrigerating compartment temperature detecting means, respectively; Storing temperature change data detected by the respective refrigerating compartment temperature sensing means along the cold air discharge direction; The compressor control means, the damper opening and closing control means, and the cold air are inputted by genetic algorithm-purge inference by receiving predetermined information from the refrigerating chamber temperature sensing means, the temperature change rate calculating means, the temperature change data storing means, and the cold air discharge control plate position detecting means. Control means for controlling the temperature in the refrigerator by genetic algorithm-fuzzy inference that infers the control amount of the discharge direction control means; Characterized in that it has been provided.

In the present invention, the refrigerating compartment temperature sensing means is preferably an infrared temperature sensor, the infrared temperature sensing sensor is preferably disposed in the center of the grill located behind the refrigerating compartment.

In the present invention, it is preferable that the refrigerating compartment temperature detecting means is arranged at least four points or more at predetermined intervals so as to be evenly arranged at each end of the at least one compartment of the refrigerating compartment, and the control means is the refrigerating compartment temperature sensing means, temperature A genetic algorithm-fuzzy control routine for inputting predetermined information from the change rate calculating means and the temperature change data storing means and inferring the control amounts of the compressor control means and the damper opening and closing control means by genetic algorithm-fuzzy inference; A calculation routine which receives the inference information and the position information of the cold air discharge control plate from the cold air discharge control plate position sensing means, and receives the calculation output information of the operation routine and the output information of the genetic algorithm-purge control routine, Control means, damper opening and closing control means and cold earth Preferably, the genetic algorithm-purge control routine comprises an ambiguous state of the temperature of the stored refrigerated food obtained from the refrigerating chamber temperature sensing means and inferred prior to the load control routine for determining the control amount of the direction control means. Fuzzy model identification means, which is a fuzzy membership function, which obtains and determines reference learning data, which is an ambiguous state of the temperature of the stored refrigerated food, and a refrigerator temperature after a predetermined time for a temperature state that changes frequently in the refrigerator. It is preferable to have a genetic algorithm for inferring the objective function having the correlation coefficient of the inference value and providing the inference information to the fuzzy model identification means, wherein the genetic algorithm is expressed by the following equation. Change frequently in the fridge using the no-river method Inferred temperature is the derivation of a set of condition determination and execution (If ----, then ---) rules, and the refrigerating chamber temperature evolved every predetermined time is deduced as the objective function having the maximum correlation coefficient. It is preferable.

Where L _n ⁱ denotes a conditional expression, ₁ x ~x _m are the various condition variables such as sensed temperatures (R1, R2, R3, R4), the outdoor temperature, the temperature of the chilled food in the refrigerating chamber ₁ ~A A _m is a map It is a bounce value, y ⁱ _n is the linear function corresponding to the conclusion as the objective function, C _{0 to} C _m are the parameters of the conclusion, and the subscripts i, m and n are positive integers.

In the present invention, the calculation routine calculates the output information of the genetic algorithm-purge control routine and the output information of the cold air discharge control plate position sensing means to obtain a control amount of the cold air discharge direction control means and provide it to the load control routine. It is preferable.

In addition, in order to achieve the above object and the method of controlling the temperature of the refrigerator using the genetic algorithm-fuzzy inference according to the present invention, in the method of controlling the temperature of the refrigerator using the genetic algorithm-fuzzy inference, determining whether the initial input ; Sensing the temperature of each part by means of temperature sensing means arranged to grasp the temperature distribution of the entire refrigerating compartment if it is an initial input; Determining an optimal position of a cold air discharge control plate for determining a cold air discharge direction by a genetic algorithm-fuzzy function using the sensed temperatures of each part; Moving the position of the cold air discharge control plate according to the optimum positioning; Determining whether a predetermined time has elapsed from the initial input step; If a predetermined time has elapsed, determining the position of the cold air discharge control plate by the genetic algorithm-purge function using the temperature of the refrigerating compartment temperature sensing means currently input; Determining whether the position of the determined cold air discharge control plate is optimal; Moving a position of the cold air discharge control plate by driving a swing motor if the position is optimal; Characterized in that it comprises a.

In the present invention, when the temperature of the predetermined position is R1, R2, R3, R4, the genetic algorithm-fuzzy function is

(Wherein MIN is a function for selecting the minimum argument among the arguments in (), and MAX is a function for selecting the maximum argument among the arguments in ()).

Preferably, R1 is a temperature value measured by 3H / 4 left temperature sensing means of the refrigerating compartment, R2 is a temperature value measured by 1H / 3 right temperature sensing means of the refrigerating compartment, and R3 is a refrigerating compartment It is preferable that the temperature value measured by the 3H / 4 right temperature sensing means of, and R / 4 is the temperature value measured by the 1H / 3 left temperature sensing means of the refrigerating chamber.

Hereinafter, an apparatus and method for controlling a temperature of a refrigerator to which a genetic algorithm-fuzzy inference according to the present invention is applied will be described with reference to the accompanying drawings.

Before entering this description, we will briefly explain the differences between the genetic algorithm and the purge compared to the PID (proportion integrate differential) control system commonly used in control systems.

PID control is applied in a wide range of fields because the configuration of the device itself is simple and the control parameters can be easily adjusted. However, it is difficult to obtain sufficiently satisfactory control results when the control target is a system having many control variables and many unpredictable elements such as a dynamic system, and in particular, a mathematically undifferentiated characteristic. The adaptive control method is introduced for such a dynamic system, but specifying the model of the control target becomes very difficult when the characteristic changes frequently.

On the contrary, a knowledge engineering method (expert system) has been proposed, which does not rely on a mathematical model of a control object and constructs a control rule by analyzing a shaping operation method obtained by an experienced driver's experience.

Genetic algorithm-fuzzy control includes the subjective ambiguity of human beings in the control rules of this knowledge engineering technique. Therefore, the system considered as an application field of genetic algorithm-fuzzy control is 1) a complex system to be controlled, and 2) a system. 3) systems in which the internal model cannot be specified strictly;

In PID control, control results are determined in advance, evaluated using two evaluation criteria, and parameter tuning is performed. On the other hand, even in genetic algorithm-fuzzy control, the control rule is changed in response to the control result to improve the characteristics, but in the case of genetic algorithm-fuzzy control, the tuning procedure is undeniable.

As such, it is 'gene algorithm-fuzzy' to actively introduce the ambiguity that humans originally had in science and technology. If some degree is expressed using a value between 0 and 1, there are many intermediate grades, such as 0.8, 0.5, 0.2, etc. The function that maps these classes is called the membership function because it represents the degree of membership in the set. Compared with the conventional set of zeros and tens, the genetic algorithm-fuzzy set immediately arrives at reality.

On the other hand, the use of the membership function, because it is possible to replace the ambiguous with a numerical value, it is the 'gene algorithm-fuzzy inference' that was able to use the ambiguous information on the computer by using this.

Genetic algorithm-fuzzy inference describes rules in the form of conditional judgment and execution (if ---, then ---; if ---, ---), and describes the degree to which the state of reality fits this rule. Investigation is carried out to output an operation amount suitable for it. The biggest feature of this inference is that genetic algorithm-fuzzy sets can be used to keep information containing ambiguity in the if --- part (preconditional) or then --- part (end) of the rule.

However, the membership function may be determined based on subjectivity. However, if the determined person is an expert (expert) in the field, the person's experience or intuition (sixth) is put in an ambiguous expression, and is equivalent to the expert. There is a big advantage to being able to manipulate the computer. Thus, genetic algorithm-fuzzy control is executed by genetic algorithm-fuzzy inference.

Based on the above concept, it will be easy to understand the temperature control device and method of the refrigerator according to the present invention.

8 is a block diagram of a temperature control apparatus of a refrigerator to which genetic algorithm-fuzzy inference according to the present invention is applied. Referring to this figure, a temperature control apparatus of a refrigerator using a genetic algorithm-fuzzy inference according to the present invention will be described.

The temperature control apparatus of the refrigerator in which genetic algorithm-fuzzy inference is used includes temperature sensors such as an outside temperature sensor 24, a freezer compartment temperature sensor 23, and a refrigerator compartment temperature sensor R1 to R4, and a temperature from the temperature sensor. A temperature change rate calculation unit 25 for calculating a temperature change rate by receiving information of a variable, a data storage unit 26 for storing temperature change data of the refrigerating compartment temperature sensors of R1 to R4 according to the cold air discharge direction, and a compressor control unit ( 28, the damper control unit 29, the cold air discharge direction control unit 30, the cold air discharge control plate position sensor 31 for setting the cold air discharge direction, and the temperature factors varying according to the environment unspecified. In controlling the temperature in the refrigerator, particularly in the refrigerating chamber, by using the compressor on / off control unit 28, the damper opening / closing control unit 29, etc., by applying a genetic algorithm-fuzzy inference, the outside air temperature compensation and storage It is comprised by the controller 27 which controls cold so that the temperature dispersion | distribution in a refrigerator may be made evenly by supplying cold air, taking into account the temperature change factor in the refrigerator according to the state of food.

Here, the R1 temperature sensor is a temperature sensor installed on the left wall of the refrigerating compartment 3H / 4, the R2 temperature sensor is a temperature sensor installed on the right wall of the 1H / 3, and the R3 temperature sensor is a temperature sensor installed on the right wall of the 3H / 4, R4 is a temperature sensor installed on the left wall of 1H / 3. The controller 27 is composed of a microprocessor TMP87C840AN, in which a genetic algorithm-fuzzy control routine 271, an operation routine 272 and a load control routine 273 are performed. These routines are described together with the following FIG.

FIG. 9 is a general flow diagram of the genetic algorithm-fuzzy control applied to the present invention, and shows the genetic algorithm-fuzzy control process performed in the genetic algorithm-fuzzy control routine 271 of the controller 27 of FIG. Fig. 8 is a flowchart of the genetic algorithm-fuzzy control routine of Fig. 8.

First, the values of R1 to R4 are sensed, and then the optimal cold air discharge direction is selected by applying a genetic algorithm-purge function to the next field. This is performed every minute to adjust the position of the cold air discharge control plate by the swing motor as shown in the flowchart of FIG. 10 to make the cold air discharge direction the optimal position.

Here, Tr is a value stored in the data storage unit 26 at the temperature change of R1 to R4 according to the cold air discharge direction of FIG. In particular, these values are data obtained through various experiments on changes in outdoor temperature, temperature distribution of refrigerated foods stored in the refrigerating chamber, and the rate of change of temperature.

In addition, the genetic algorithm-fuzzy model identifier 51 has a built-in genetic algorithm-fuge membership function that determines an ambiguous temperature state such as hotness, hotness, moderateness and coldness of the load (refrigerated food) put into the refrigerating compartment.

That is, when the genetic algorithm-fuzzy function is MIN, a function of selecting a minimum argument among the arguments in (), MAX is a function of selecting a maximum argument of the arguments in (),

here,

P1, P2, P3, and P4 represent positions of the cold air discharge control plate for cold air discharge direction control determined by the genetic algorithm described below, and w represents the weight of the position P of the cold air discharge control plate at the optimum position.

And the genetic algorithm-fuzzy function is applied to the optimal positioning calculation process of the cold air discharge control plate of FIG.

Genetic Algorithm (TKS) is a process that creates a simple solution set of IF --- THEN --- rules, such as the evolution, breeding, mutation, and reproduction of ecosystems. The refrigerator temperature is inferred as an objective function with the largest correlation coefficient (correlation).

The Takagi-Sugeno-Kang (TSK) fuzzy inference method consists of the following fuzzy rules.

Where L ⁱ _n denotes that the fuzzy rule is the i-th fuzzy rule of the equation of the n-th unit, xj denotes the input variable Aj _l is the fuzzy junction, and Y ⁱ _n is the output of the i-th fuzzy rule of the n-th step, C ⁱ _j is the parameter of the conclusion. The inference result TY from any fuzzy model for any one input is given by

Where n is the number of fuzzy rules, if W is expressed as the i-th rule, the entire portion fit (i.e., weight) the fuzzy set A ⁱ _j of the X ⁱ _{j j} ⁱ A membership value in the W represent are the following:

The fuzzy model, which consists of fuzzy rules as shown in Equation (2-2), is a form in which each fuzzy subspace is represented by one linear equation by fuzzy division of the input space, and has an excellent ability to express complex nonlinear systems.

In the above fuzzy model, if the membership function of the whole fuzzy set is complex, the calculation time is long. However, if we take only three types of distinct linear functions as shown in Fig. 11 as the membership function of the predicate fuzzy set, the fuzzy model can be expressed by one relatively simple expression. Can be shortened.

In other words, denotes the membership values of the fuzzy sets A1, A2, and A3 of FIG. 11 with respect to the input x as A (x) _1, A (x) ₂ , and A (x) ₃ .

Since the equation (2-2) to express the entire model is simplified, the storage space for storing the model and the calculation time can be significantly shortened.

The norms for verifying the structure of the fuzzy model as described above can be described as follows.

What is important in the fuzzy model is the recognition of the structure, and therefore, the norm for evaluating the created structure is needed. In other words, there is a need for a norm evaluating whether the rules recognized from the finite data for recognition are not suitable for finite data used for recognition, but are universal rules that accurately represent the actual system. As the norm, unbiasedness criterion (U.C.) is used.

UC of Equation 2-4 is defined by assuming that A and B are fuzzy models created by dividing I / O data into two groups, A and B, and processed through rule recognition.

Next, conclusion recognition is made as follows.

In conclusion recognition, the output y is represented by one linear expression of the conclusion parameters C ¹ ₀ , C ¹ ₁ ,..., C ⁿ _m , so that a conventional method in a linear system can be used. In other words, the structural recognition of the conclusion part is to find out which variables are related to the conclusion part by using the variable reduction method, and the recognition of the conclusion part by using the least-squares method.

The inference equation (Equation 2-2) of the output from the fuzzy model can be expressed by the following Equation 2-5.

The inference equation y of the output in Equation 2-5 is expressed as one linear form of the conclusion parameters (C ¹ ₀ , C ¹ ₁ ,…, C ⁿ _m ).

Next, the premise recognition consists of the following steps.

In prerequisite recognition, prerequisite structure recognition is to find fuzzy division of variable space, and prerequisite parameter recognition is to find a parameter that characterizes the membership function of fuzzy set, and to minimize the sum of squared output errors or to maximize the correlation coefficient It is obtained using the nonlinear programming method as the objective function. The structure of the premise is recognized at the following levels.

Step 1: Recognize the evaluation norm U. C values by recognizing one linear model.

Step 2: Create a fuzzy model with two rules by dividing the space of one predicate variable into two. Then, as many fuzzy models as the number of variables can be made in the formulation. Calculate the U.C value for each fuzzy model, and predicate variables of the fuzzy model whose U.C values are less than the U.C values of the linear model will not be included in the preamble.

Step 3: In step 2, three fuzzy models having a regular number are created from the preamble structure of the fuzzy model having the smallest U.C value, and the U.C values for each fuzzy model are calculated. If the smallest of the U.C values is larger than the smallest U.C value of the previous step, the fuzzy model structure of the previous step is set to the optimal structure. Otherwise, go to the next step and continue the same work.

Referring to the process of the fuzzy inference about the optimum position of the cold air discharge control plate in the manner described above as follows.

In GA-FUZZY inference for setting the cold air discharge direction according to the internal temperature distribution, input data (variable) X ₁ is 3H / 4 left high temperature (R1), X ₂ is 3H / 4 right high temperature (R3), X Inference is performed for each step by setting ₃ as 1H / 3 left inner temperature R4, X ₄ as 1H / 3 right inner temperature R2, and output data (variable) Y as the cold air discharge direction P.

Stage 1

First, using the least-squares method from the entire data, including the appendix X ₁ , X ₂ , X ₃ , and X ₄ , a linear equation of the following equation (3-6) corresponding to the conclusion of TSK-fuzzy inference is produced. Variable reduction method to reduce the variable as much as possible.)

In addition, the data is divided into two arbitrary groups into groups A and B, and the U.C value of stage 1 is obtained by using a function of the formula (3-7).

The U.C value at this time is called U.C1 (= 1.25).

Stage 2

Set up a fuzzy model with two plant laws. This is where the preconfiguration problem comes. In setting up the structure, the choice of variables and the fuzzy division are considered at the same time.

First, consider a structure having only one of X1, X2, X3, and X4 as a predicate variable, and divide the space into two. Therefore, four types of structures of the premise part are considered.

Obtain each U.C value for the above four.

In order to find the UC value, we need to find the fuzzy partition (preliminary parameter setting) for each. In this system, we applied GENETIC ALGOFITHM instead of COMPLAEX method of preliminary parameter determination. .

For example, (1) The predicate parameter of the structure is obtained as follows.

The control range of high temperature -10 ℃ ~ 20 ℃ is subdivided into 0.1 ℃ to make 300 strings, and the string strings corresponding to them are randomly generated. Of the 300 strings of this data, four values of 1 corresponding to P1, P2, P3, and P4 are created, and the rest are filled with zeros.

We then calculate the correlation coefficient for each of the data, duplicate the top 10% of the good (high) candidates, cut off the bottom 10% of the bad ones, and cross the middle 10 with the top 10% to produce new data. Obtain the neutral correlation coefficient using the recreated data and repeat the above steps.

Continuously comparing the neutral correlation coefficient of the data produced repeatedly, if the higher value no longer occurs, the string data at that time becomes the final parameters as the final P1, P2, P3, and P4.

When the predicate parameter is determined, U.C value is calculated accordingly. At this time, U.C value becomes U.C value of (1) structure. The UC values UC (2-1), UC (2-2), UC (2-3), and UC (2-4) of the structures (1), (2), (3), and (4) are obtained. We choose the structure with the smallest UC value and use fuzzy partitioning to make three rules based on this structure.

In the current system, the magnitude of the U.C value is

Therefore, we create a new three-part structure based on the structure that divides the variable X ₁ .

Stage 3

To create a three-part structure, we need to add new variables to make the space X ₁ to X _i . In this case, since X _i = X ₂ , X ₃ , and X ₄ , many structures are generated. In order to eliminate unnecessary structures, the structure of the stage 2, which has a UC value larger than the value of UC ₁ , is omitted by simplifying the structure.

Therefore, the current system uses fuzzy partitioning of stage 3, which divides the space of X ₁ and X ₄ .

The structure at this time is as shown in FIG.

The purge partition is as shown in FIG.

(1) As for the method of obtaining the premise parameters of the structure, the genetic algorithm was applied as in the stage 2. First, the control range of high temperature -10 ℃ ~ 20 ℃ is subdivided into 0.1 ℃ to make 300 strings, and a large number of corresponding string data are randomly generated. Of the 300 strings of this data, six values of 1 corresponding to P1, P2, P3, P4, P5, and P6 are created, and the rest are filled with zeros. Then, the correlation coefficient for each data is calculated, and the top 10% of the high correlation coefficient is replicated, the low 10% of the bad one is removed, and the intermediate data is crossed with the top 10% to create new data. Obtain the neutral correlation coefficient using the recreated data and repeat the above steps.

Continuously comparing the neutral correlation coefficient of the data produced repeatedly, if the higher value no longer occurs, the string data at that time becomes the final parameter as the final P1, P3, P2, P4, P5, P6.

When the precondition parameter is determined, U.C value is calculated accordingly. At this time, U.C value becomes U.C value of (1) structure. In this way, the UC values UC (3-1), UC (3-2), and UC (3-3) of each of the structures (1), (2), and (3) are obtained, and the smallest UC value is selected. We do the fuzzy division that makes four RULE based on. At this time, UC (3-1), UC (3-2), and UC (3-3) values are better (smaller) than UC (2-1), so they are divided into four and all three UC values are UC (2- If it is worse than 1), the division of 4 is stopped, and the final RULE with the UC (2-1) value is defined as the final RULE.

In the current system, the magnitude of the U.C value is

UC _(3-1) (1.029) UC _(2-1) (1.14) UC _(3-2) (2.88) UC _(3-3) (3.08)... (3-9)

Therefore, a new quadrant structure is created based on the structure (1) of stage 3.

Stage 4

In this stage, the prerequisite structure of the stage 3 model is further subdivided, and a model consisting of four plant laws is set. If the UC value is smaller than UC (2-1) in Stage 3, it should be regarded as the starting point when the quadrant is divided into three, but in order to reduce the search method, the structure of the quadrant of the UC is shown in FIG. As you can see, there are five things you can do.

Find the U.C value for each structure.

Here, genetic algorithm is used to calculate U.C value like stage 2 and stage 3.

Structure (1), (2), (3). The UC values for (4) and (5) are UC (4-1), UC (4-2), UC (4-3), UC (4-4), and UC (4-5). The magnitude of UC at is

Therefore, based on the structure of U.C (4-3), five divisions of the stage 5 are performed. However, in this system, each U.C value was obtained by 5 divisions, and the results were worse than U.C (4-3) value. Therefore, the application of the cold air discharge direction setting GA-FUZZY according to the high temperature distribution has a third preliminary structure of four divisions of FUZZY.

Lastly, the preliminary structure and parameter setting, the structure of the conclusion part, and the parameter setting for the finally applied GA-FUZZY are summarized as follows (this is the third FUZZY partition structure of stage 4).

Here, substituting R1, R3, R4, and R2 in place of X ₁ , X ₂ , X _3, and X ₄ , respectively, yields a formula (1-3).

In order to find W ₁ and W _2, we must first find the parameters of the premise.

The parameters of the premise are the same as those in FIG.

In this system, as shown in FIG. 20, RULE of P1 = P3 = 1.33, P2 = P4 = 3.49, P5 = P7 = 1.9, P6 = P8 = 4.39 was applied, and P1, P2, P3, P4, P5, P6, P7 and P8 values are as follows.

With this value, use the following equation (3-11) to find the value of W ₁ , W ₂ .

Equation (1-1) is derived by substituting the corresponding R1 and R2 in place of X ₁ and X _{4 in the} formula (3-11).

Using W ₁ , W ₂ and y ₁ (P1), y ₂ (P2), y ₃ (P3) and y ₄ (P4), the final ejection direction is determined by the following equation.

The effects of the apparatus and method for controlling the temperature of the refrigerator to which the genetic algorithm-fuzzy inference described above are applied will be described below with reference to FIGS. 12 to 15.

12 is a characteristic curve representing a temperature drop of a refrigerating compartment according to a cooling time according to a simple rotation method of a conventional cold air discharging control plate in a cold air discharge method of a refrigerating compartment, and a cold air discharge control plate during cold air discharge according to the present invention. It is shown by comparing the temperature drop characteristic curve with the cooling time of the refrigerating chamber obtained by purge control.

Here, (A) to (D) is a temperature line when the refrigerator compartment is cooled while simply rotating the cold air discharge control plate, (A) is the lowest shelf temperature line, (B) is the highest shelf temperature line, (C) is the lowest temperature line of the shelf, and (D) is the highest temperature line of the shelf. (A ') to (D') are temperature lines when the refrigerator compartment is cooled by the cold air discharge control plate by position control of the cold air discharge control plate by genetic algorithm-purge function inference. Is the upper shelf temperature line, and (D ') and (D') are the lower half temperature lines.

As such, when cooling the refrigerating chamber while controlling the cold air extraction direction by the genetic algorithm-fuzzy inference, the difference between the highest temperature and the lowest temperature is almost eliminated, and the temperature gap between the shelf and the shelf is much smaller. It works.

FIG. 13 is a schematic perspective view showing a temperature measuring point of each shelf for measuring a temperature distribution of each point of a refrigerating compartment of a refrigerator based on a conventional refrigerator and genetic algorithm-fuzzy inference, and FIG. 14 is a conventional genetic algorithm-fuzzy inference In the refrigerator to which is applied, a graph showing a temperature distribution at each measuring point shown in FIG. 13, and FIG. 15 is a refrigerator to which genetic algorithm-fuzzy inference according to the present invention is applied, at each measuring point shown in FIG. It is a graph showing the temperature distribution of.

Here, the measurement conditions were performed at 30 ° C and 75% in a constant temperature and humidity chamber.

Temperature knitting of the refrigerating chamber at the time of conventional cooling T, as shown in Fig. 14, is 2.5 deg. C, but the temperature deviation of the refrigerating chamber during cold air discharge direction controlled cooling by genetic algorithm-fuzzy inference. T becomes a temperature distribution with much less variation at 0.9 ° C. as shown in FIG.

As described above, the apparatus and method for controlling the temperature of the refrigerator to which the genetic algorithm-fuzzy inference according to the present invention is applied are respectively provided on the 3H / 4 left wall and the 1H / 3 right wall and the 3H / 4 right wall and the 1H / 3 left wall of the refrigerator compartment, respectively. By installing a temperature sensor and inferring the temperature by the genetic algorithm-purge function to compare the temperature distribution by learning with the data and controlling the cold air discharge direction, the cooling speed is increased and the temperature of the whole refrigerator compartment according to the temperature difference of the refrigerated food is increased. In addition to the uniform temperature distribution, it also has the effect of compensating the outside temperature by opening and closing the door.

Claims (11)

Compressor control means; Damper opening and closing control means; Cold air discharge direction control means for determining the cold air discharge direction by power or switch control of the cold air discharge control plate of the disc member; Cold air discharge control plate position sensing means for setting a cold air discharge direction with information applied from the cold air discharge direction control means; Refrigerator compartment temperature sensing means for sensing a temperature of at least four points or more so as to sense a temperature distribution of the entire refrigerator compartment; Temperature change rate calculating means for calculating a temperature change rate of the refrigerating compartment temperature detecting means at a predetermined number or more points of the refrigerating compartment temperature monitoring means, respectively; Temperature change data storage means detected by each of the refrigerating compartment temperature sensing means along the cold air discharge direction; The compressor control means, the damper opening and closing control means, and the cold air are inputted by genetic algorithm-purge inference by receiving predetermined information from the refrigerating chamber temperature sensing means, the temperature change rate calculating means, the temperature change data storing means, and the cold air discharge control plate position detecting means. Control means for controlling the temperature in the refrigerator by genetic algorithm-purge inference that infers the control amount of the discharge direction control means; Refrigerator temperature control apparatus using a genetic algorithm-fuzzy inference, characterized in that provided with. The apparatus of claim 1, wherein the refrigerating compartment temperature sensing means comprises an infrared temperature sensing sensor. The apparatus of claim 2, wherein the infrared temperature sensor is disposed at the center of the grill located behind the refrigerating compartment. 4. The refrigerator temperature control according to claim 1, wherein the refrigerating compartment temperature sensing means is arranged at least four points at predetermined intervals so as to be evenly disposed at each end of the divided refrigerating compartment. Device. The control means according to claim 1, wherein the control means receives predetermined information from the refrigerating compartment temperature sensing means, the temperature change rate calculating means, and the temperature change data storing means, and performs the compressor control means and the damper opening and closing control means by a genetic algorithm-fuzzy inference. A genetic algorithm-fuzzy control routine for inferring the control amount of the control routine, an inference information of the control routine and a calculation routine for receiving and calculating the position information of the cold air discharge control plate from the cold air discharge control plate position sensing means, and the calculation output of the operation routine And a microprocessor configured to perform a load control routine for determining a control amount of the compressor control means, the damper opening and closing control means, and the cold air discharge direction control means by receiving information and output information of the genetic algorithm-purge control routine. Algorithm-Fuzzy Inference Refrigerator temperature control device. The reference learning data according to claim 5, wherein the genetic algorithm-purge control routine is information storing an obscure state of the temperature of the stored refrigerated food obtained from the refrigerating compartment temperature sensing means and an obscure state of the temperature of the refrigerated food previously deduced. The fuzzy model identification means, which is a fuzzy membership function, which is obtained by determining a function of the digital signal, and the temperature state that changes from time to time in the refrigerating chamber. The refrigerating chamber temperature after a predetermined time is inferred as an objective function having a correlation between the experimental value and the inferred value. And a genetic algorithm for providing the inference information to the model identification means, wherein the refrigerator temperature control device using the genetic algorithm-fuzzy inference. (Correction) The method according to claim 6, wherein the genetic algorithm obtains a solution set of condition determination and an execution rule for a temperature which changes frequently in the refrigerating chamber by using the Tagagi-Sugeno-Steel method represented by the following conditional expression L ⁱ _n . And inferring the refrigerator compartment temperature evolved every predetermined time as an objective function having a maximum number of the above-described phases of correlation. Where L _n ⁱ denotes a conditional expression, ₁ x ~x ₂ is a variety of condition variables, such as temperature sensing (R1, R2, R3, R4), the outdoor temperature, the temperature of the chilled food in the refrigerating compartment A ₁ ~A _m are members Shh, Y ⁱ _n is the objective function, the linear equation corresponding to the conclusion, C _{0 to} C _m are the parameters of the conclusion, and the subscripts, i, m and n are positive integers. And said calculation routine calculates output information of said genetic algorithm-purge control routine and output information of said cold air discharge control plate position sensing means to obtain a control amount of said cold air discharge direction control means and provide it to said load control routine. Refrigerator temperature control device using algorithm-fuzzy inference. CLAIMS What is claimed is: 1. A method of controlling a temperature of a refrigerator to which genetic algorithm-fuzzy inference is applied, the method comprising: determining an initial input; Sensing the temperature of each part by means of temperature sensing means arranged to grasp the temperature distribution of the entire refrigerating compartment if it is an initial input; Determining an optimal position of a cold air discharge control wave for determining a cold air discharge direction by a genetic algorithm-fuzzy function using the sensed temperatures of each part; Moving the position of the cold air discharge control plate according to the optimum positioning; Determining whether a predetermined time has elapsed from the initial input step; If a predetermined time has elapsed, determining the position of the cold air discharge control plate by the genetic algorithm-purge function using the temperature of the refrigerating compartment temperature sensing means currently input; Determining whether the position of the determined cold air discharge control plate is optimal; Moving a position of the cold air discharge control plate by driving a swing motor if the position is optimal; Genetic algorithm-temperature control method of the refrigerator using a fuzzy inference comprising a. 10. The method of claim 9, wherein when the temperature at the predetermined position is R1, R2, R3, R4, the genetic algorithm-fuzzy function is (Wherein MIN is a function for selecting the minimum argument among the arguments in (), MAX is a function for selecting the maximum argument among the arguments in ()). Temperature control method of the refrigerator using. 11. The method of claim 10, wherein R1 is the temperature value measured by the 3H / 4 left temperature sensing means of the refrigerating compartment, R2 is the temperature value measured by the 1H / 3 right temperature sensing means of the refrigerating compartment, R3 is 3H / 4 is a temperature value measured by the right temperature sensing means, and R4 is a temperature value measured by 1H / 3 left temperature sensing means of the refrigerating compartment.