JPH07332732A - Control device for air conditioner - Google Patents

Abstract

(57) [Summary] [Purpose] To improve comfort by enabling air conditioning operation suitable for environmental conditions. [Configuration] Controller (30) of air conditioner (20) and room temperature sensor (4
0) and the outside air temperature sensor (41) are provided, and the heat input q
Frequency sensor (42) for detecting the frequency input F of the air conditioner (20) corresponding to the above, and dead time detecting means (31) for the air conditioner (20)
And are provided. Further, the indoor temperature Ta, the outdoor temperature To, the dead time L, the variation dTa / dt of the indoor temperature Ta, the heat capacity Ca of the room (11), the heat transfer coefficient h between the inside and outside of the room (11), and the frequency input F. Parameter calculation means (32) for deriving the heat capacity Ca, the heat transfer coefficient h, and the coefficient k in the equation of state specified by the coefficient k of the frequency input F by the least square method is provided. In addition, the parameter calculation means (32)
Gain calculation means (33) for deriving a control gain based on the empirical formula of Chen et al. Is provided based on the heat capacity Ca, the heat transfer coefficient h and the coefficient k derived by the above and the dead time L detected by the dead time detection means (31). ing.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device for an air conditioner, and more particularly to measures for adapting to environmental conditions.

[0002]

2. Description of the Related Art Conventionally, various types of control devices for air conditioners, which are air conditioners, have been proposed. Generally, as shown in FIG.
Is input, the controller (90) controls the air conditioner (91), and the air conditioner (91) air-conditions the room (92) which is the environment. Then, the room temperature Ta is fed back to the controller (90), and the controller (90) controls the air conditioner (91) so that the room temperature Ta becomes the set temperature T. Further, the comfort of the occupant cannot be improved only by inputting the room temperature Ta, and therefore, JP-A-5-23
As disclosed in Japanese Patent No. 1693, there is a system to which individual difference learning adaptive control is applied, and this control device is a function of the environmental physical quantity of the room (92) such as the room temperature and the four parameters. While calculating the thermal sensation index of thermal sensation,
Have the residents report true thermal sensation, learn the above parameters, calculate the thermal sensation index from this parameter and the indoor temperature, etc., and control the compressor frequency based on this thermal sensation index. It is designed to satisfy the occupants' comfort.

[0003]

However, in the above-described control device for the air conditioner, the room environment to be controlled is controlled assuming a typical room environment, and the control gain of the controller has an optimum value. There was a problem that it was not set to. That is, conventionally, the room environment is set to be fixed between 6 tatami mats and 8 tatami mats, for example, and even if the rooms are of the same size, the environment differs depending on furniture and the like. is there. However, in the past, these room environments were controlled by fixing the environmental conditions, so that the indoor conditions were kept comfortable for the occupants.
It will take time to air-condition (92). Specifically, as shown in FIG. 5X, when the heating operation is started, the room temperature temporarily exceeds the comfortable temperature Z, overshoots, and then converges to the comfortable temperature Z, so that comfort cannot be improved. There was a problem.

The present invention has been made in view of the above points, and an object thereof is to improve comfort by enabling air conditioning operation suitable for environmental conditions.

[0005]

In order to achieve the above-mentioned object, means taken by the present invention is to set a control gain suitable for environmental conditions. Specifically, FIG.
As shown in FIG. 1, the means taken by the invention according to claim 1 is that an air conditioner (20) for air-conditioning an air-conditioned space (11),
It is premised on a controller for an air conditioner including a controller (30) for controlling the operation of the air conditioner (20) so that the internal air temperature of the air-conditioned space (11) becomes a predetermined value.
Then, an internal temperature detecting means (40) for detecting an air temperature inside the air-conditioned space (11) and an external temperature detecting means (41) for detecting an air temperature outside the air-conditioned space (11) are provided. Along with, heat input detection means (42) for detecting heat input from the air conditioner (20) to the air-conditioned space (11), and dead time detection means (31) for detecting dead time of the air conditioner (20). ) And are provided. Further, based on the internal air temperature and the external air temperature detected by the internal temperature detection means (40) and the external temperature detection means (41), and the heat input detected by the heat input detection means (42), the conditioned space ( A parameter calculation means (32) is provided for deriving a plurality of parameters of the thermal equation of state that specifies the environmental state of 11). In addition, the parameters derived by the parameter calculation means (32) and the dead time detection means
Gain calculation means (33) for deriving a control gain of the controller (30) from the dead time detected by (31) and outputting the control gain to the controller (30) is provided. According to the invention of claim 2, in the invention of claim 1, the heat input detecting means (42) detects the frequency input F of the air conditioner (20) corresponding to the heat input q. Is configured. Then, the parameter calculation means (32) calculates the internal air temperature Ta, the external air temperature To, the dead time L, the change amount dTa / dt of the internal air temperature Ta, the heat capacity Ca of the air-conditioned space (11), and the air-conditioned space (11). The heat capacity Ca and the heat transfer coefficient h and the coefficient k in the equation of state specified by the heat transfer coefficient h between the inside and outside, the frequency input F, and the coefficient k of the frequency input F are configured to be derived by the least square method. ing. Further, the means taken by the invention according to claim 3 is the invention according to claim 2, wherein the gain calculating means (33) is the parameter calculating means (32).
The control gain of the controller (30) is derived based on the empirical formula of Chen et al. Based on the heat capacity Ca, the heat transfer coefficient h and the coefficient k derived by the above, and the dead time L detected by the dead time detecting means (31). It was done.

[0006]

With the above construction, in the invention according to claim 1,
For example, when the air conditioner (20) is activated, the internal temperature detecting means (40) measures the internal air temperature of the air-conditioned space (11), and the external temperature detecting means (41) measures the external air temperature of the air-conditioned space (11). On the other hand, the heat input detecting means (42) detects the heat input to the air-conditioned space (11), and the dead time detecting means (31) detects the dead time.
The parameter calculation means (32) is configured to detect the internal air temperature and the external air temperature detected by the internal temperature detection means (40) and the external temperature detection means (41), and the heat input detected by the heat input detection means (42). Based on and, a plurality of parameters of the thermal equation of state that specify the environmental state of the air-conditioned space (11) are derived. Specifically, in the invention according to claim 2, the heat input detecting means (42)
Is the frequency input F of the air conditioner (20) corresponding to the heat input q
On the other hand, the parameter calculating means (32) detects the internal air temperature Ta, the external air temperature To, the dead time L, and the internal air temperature Ta.
Change amount dTa / dt and the heat capacity of the air-conditioned space (11) Ca and the air-conditioned space (1
To derive the heat capacity Ca, the heat transfer coefficient h, and the coefficient k in the equation of state specified by the heat transfer coefficient h between the inside and outside of 1), the frequency input F, and the coefficient k of the frequency input F by the method of least squares. become. In other words, the inventors of the present application set the environmental conditions of the air-conditioned space (11) such as the indoor temperature Ta, the outdoor temperature To, the dead time L, the change amount dTa / dt of the indoor temperature Ta, the heat capacity Ca of the air-conditioned space (11), and the air conditioning. Heat transfer coefficient h of space (11) and frequency input F
We find a point that can be specified by a simple equation of state by using the coefficient k and the coefficient k, and the heat capacity Ca, the heat transfer coefficient h, and the coefficient k.
And are derived by the method of least squares. Subsequently, the gain calculation means (33) determines the controller (3) by using the parameters derived by the parameter calculation means (32) and the dead time L detected by the dead time detection means (31).
0) Derivation of the control gain and the controller gain
It will be output to (30). Specifically, in the invention according to claim 3, the gain calculating means (33) includes the heat capacity Ca, the heat transfer coefficient h and the coefficient k derived by the parameter calculating means (32).
And the dead time L detected by the dead time detecting means (31), the control gain is derived based on the empirical formula of Chen et al. calculate. After that, the gain calculation means
After the control gain derived by (33) is output to the controller (30), the control gain of the controller (30) is set, and the controller (30) is system-identified in the environmental state, normal temperature control is performed by the controller ( 30), and operation control suitable for the environment of the air-conditioned space (11) is performed.

[0007]

Therefore, according to the inventions of claims 1 and 3, the environmental condition of the air-conditioned space (11) is specified by a simple thermal state equation, and the parameter calculating means (32) and the gain calculating means (3
By setting the control gain according to the environment of the air-conditioned space (11) by 3), set the control gain suitable for the room environment, for example, for 6 or 8 tatami mats. can do. As a result, since the internal temperature does not overshoot the set temperature or the internal temperature does not vibrate, stable temperature control can be performed,
The room temperature can reach the set temperature in a short time. And while improving comfort,
Energy loss can be reduced. Further, in the invention according to claim 2, since the parameters are derived by the least squares method, it is possible to derive the reliable parameters,
Accurate system identification can be performed.

[0008]

Embodiments of the present invention will now be described in detail with reference to the drawings. As shown in FIG. 2, (10) is an air-conditioning system to which the present invention is applied, in which a wall-mounted air conditioner (20) is installed in a room (11) in a closed space which is an air-conditioned space. Then, the air conditioner (20) air-conditions the room so that the room temperature of the room (11) becomes a set value. The air conditioner (20) is hereinafter referred to as an air conditioner. FIG. 3 shows the above air conditioner.
Shows the control block of (20), controller (30)
Is configured to output a control signal to the air conditioner (20) and control the air conditioner (20) when the set temperature is input from the setting unit (21). Then, the room temperature in which the air conditioner (20) is air-conditioned is fed back to the controller (30).

The air conditioning system (10) is provided with a room temperature sensor (40), an outside air temperature sensor (41) and a frequency sensor (42), while a dead time detecting means (31) and a parameter calculating means (31). 32) and a gain calculating means (33) are provided so as to system-identify the controller (30) in an environmental state. The room temperature sensor (40) constitutes an internal temperature detecting means for detecting an indoor temperature Ta which is an air temperature inside the room (11), and an outside air temperature sensor (41) is an air temperature outside the room (11). The external temperature detecting means for detecting the outdoor temperature To which is The frequency sensor (42) is connected to the room from the air conditioner (20).
The heat input detecting means for detecting the heat input q to (11) is configured, that is, the heat input q is proportional to the frequency input F input to the compressor of the air conditioner (20), Coefficient k in F
Since it is equal to the value multiplied by, the frequency sensor (42)
Are configured to detect the frequency input F upon which the heat input q is based. Further, the dead time detection means (31),
The dead time of the air conditioner (20) is detected. That is, the dead time detection means (31) detects the time interval of the output of the air conditioner (20) in response to the input of the air conditioner (20), and, for example, during the air conditioning operation, the change amount of the indoor temperature Ta. Becomes a predetermined value,
It is configured to detect a response time for an input such as temperature setting as dead time. Parameter calculation means
(32) receives the detection signals of the room temperature sensor (40), the outside air temperature sensor (41) and the frequency sensor (42), and determines the heat capacity Ca and heat in the thermal equation of state that specifies the environmental state of the room (11). It is configured to derive the transfer coefficient h and the coefficient k by the method of least squares. In other words, the inventors of the present application set the environmental conditions of the room (11) such that the indoor temperature Ta, the outdoor temperature To, the dead time L, the change amount dTa / dt of the indoor temperature Ta, the heat capacity Ca of the room (11), and the room (1
Finding a point that can be specified by a simple equation of state by the heat transfer coefficient h of 1), the frequency input F, and the coefficient k,
The heat capacity Ca, the heat transfer coefficient h, and the coefficient k are derived by the least square method. The gain calculation means
(33) is the heat capacity Ca derived by the parameter calculation means (32),
Based on the heat transfer coefficients h and k and the dead time L detected by the dead time detecting means (31), the control gain of the controller (30) is derived based on the empirical formula of Chen et al. And the control gain is given to the controller (30). It is configured to output.

Therefore, the basic principle of system identification of the controller (30) by the heat capacity Ca, the heat transfer coefficient h, the coefficient k, and the dead time L will be described. First, the environmental condition of the room (11) relating to the basic heat can be expressed as the following equation as a relatively simple differential equation.

[0011]

[Formula 1] This equation is the above-mentioned thermal equation of state, Ta (t)
Is the indoor temperature, To (t) is the outdoor temperature, and q (tL) is the heat input from the air conditioner (20). In addition, Ca is the heat capacity of the room (11), h is the heat transfer coefficient between the room and the room, and it is assumed that effects such as volume and area are included. L is dead time. In the above equation, the heat input q (tL) is the frequency input F (tL)
Since it is considered to be equal to the value obtained by multiplying by, a coefficient k is obtained.

[0012]

[Formula 2] Laplace transform of this equation gives the following equation.

[0013]

[Formula 3] In this equation, the first term on the right side is related to the room, the second term is related to the outdoors, and there is a time delay (temporary delay) Ca / h between the room temperature ta and the frequency input F. It can be seen that there is a relationship {(Ca / h) + L} in which time L is added. If the proportional coefficient k / h, the time delay Ca / h, and the dead time L are known, the PID gain or PI gain, which is the control gain of the controller (30), can be obtained from Chen's empirical formula.

[0014]

[Formula 4] This formula and the empirical formula of Chen et al. Shown in the formula show that when the controlled object can be treated as dead time + temporary delay system, P is calculated by the proportional coefficient K, the time constant T and the dead time L
The control gain of the ID controller (30) or the PI controller (30) can be derived. Substituting the proportionality coefficient k / h, the time delay Ca / h, and the dead time L into the above equation and the equation,

[0015]

[Formula 5] That is, if the coefficient k, the heat capacity Ca, the heat transfer coefficient h, and the dead time L are known, the PID gain or PI gain in the controller (30) can be determined.

On the other hand, in the above equation, the indoor temperature Ta, the outdoor temperature To, and the frequency input F are physical variables obtained by sensing, and the heat capacity Ca, the heat transfer coefficient h, and the coefficient k are collected as time series data. Can be obtained by the method of least squares. Further, the dead time L can be detected by the temperature fluctuation of the room temperature Ta at the time of starting the air conditioner (20) as described above. Therefore, the least squares method for obtaining the heat capacity Ca, the heat transfer coefficient h, and the coefficient k may be efficiently learned with a small memory by using a sequential learning type algorithm. Mathematically expressing this iterative learning type least squares algorithm is as follows. First, if the above formula is transformed,

[0017]

[Formula 6] And the sensing vector ZN of the detected value to be detected and the parameter vector Cha of the desired parameter.
tN is defined as follows.

[0018]

[Formula 7] The parameter vector ChatN of the above equation is sequentially updated according to the following equation, learning and adaptation are performed, and the proportional coefficient k / h and the time delay Ca / h are derived by this parameter vector ChatN to obtain the optimum control gain of the controller (30). Is set.

[0019]

[Formula 8] The parameter vector ChatN can be updated each time sensing is performed, but when sensing data includes a vibration error, it is preferable to perform smoothing on several time-series data.

-System Identification Operation- Next, the system identification operation of the air conditioning system (10) will be described with reference to FIG. First, when the air conditioner (20) is started, in step ST1, after setting the initial value, the process proceeds to step ST2, and after resetting the timer for counting the sensing time, the process proceeds to step ST3 and the room (11) physical Sensing a variable. That is, the indoor temperature Ta detected by the indoor sensor and the outdoor temperature detected by the outdoor air temperature sensor (41).
After taking To and the frequency input F of the compressor in the air conditioner (20) detected by the frequency sensor (42), the process proceeds to step ST4, and it is determined whether or not the dead time L needs to be measured. Whether or not the dead time L needs to be measured is determined by, for example, whether or not the variation dTa / dt of the room temperature Ta detected by the room temperature sensor (40) is larger than a set value.
If the change amount Ta of Ta is larger than the set value, the process proceeds from step ST4 to step ST5, and the dead time detection means (31)
Detects the response time to the input such as temperature setting as the dead time L, and moves to step ST6. On the other hand, when the change amount dTa / dt of the room temperature Ta is smaller than the set value, it is determined that the dead time L does not need to be measured, and the step ST
It moves from 4 to step ST6.

In step ST6, it is determined whether or not the shift to the learning mode is required. Since system identification is performed during normal air conditioning operation, step ST6
When the determination at 6 is NO, the process proceeds to step ST7, and the controller (30) controls the temperature of the air conditioner (20) normally. On the other hand, during the initial operation of the air conditioner (20) and the room temperature during installation.
If Ta is abnormal or the like, it is determined that the system has not been identified, and the determination in step ST6 is YES, and the process moves to step ST8 to perform system identification by the learning algorithm that is a feature of the present invention. Specifically, the indoor temperature Ta detected by the room temperature sensor (40), the outdoor temperature To detected by the outdoor air temperature sensor (41), and the frequency input F detected by the frequency sensor (42) are sequentially taken in, and the indoor temperature Ta and Parameter calculation means from the outdoor temperature To and the frequency input F (32)
Update the parameter vector ChatN of the equation by the method of least squares based on the equation (10) to calculate the proportional coefficient k / h and the time delay Ca.
/ h will be derived. After that, the process proceeds from step ST8 to step ST9, and the gain is calculated from the proportional coefficient k / h and the time delay Ca / h derived by the parameter calculating means (32) and the dead time L detected by the dead time detecting means (31). Computing means
(33) is based on the empirical formula of Chen et al.
The derivative gain Td is derived as necessary. After that, the control gain derived by the gain calculating means (33) is changed to the controller.
It is output to (30) and the control gain of the controller (30) is set to identify the system. Then, the process proceeds from step ST9 to step ST7, and the normal temperature control is performed by the controller (30). At that time, since the system identification of the controller (30) is performed, the operation control suitable for the environment of the room (11) is performed.

Then, from step ST7 to step ST
Moving to 10, it is determined whether the system identification is completed, for example, the step ST10 to step ST until the control gain calculated by the gain calculating means (33) converges to a predetermined value.
Go to step 11, advance the sensing time of the timer, return to step ST3 above, repeat the above operation, and repeat the above steps.
System identification will be performed by updating the parameter vector ChatN of the equation in ST8. After that, when the control gain converges to a predetermined value, the determination in step ST10 becomes YES, the system identification is completed, and the air conditioning operation is controlled by the identified control gain. According to this system identification, as shown in FIG. 5X, when the heating operation is started, the indoor temperature once overshoots the comfortable temperature Z and then converges to the comfortable temperature Z. In the example, as shown in FIG. 5Y, the room temperature Ta
Will converge to the comfortable temperature Z stably and quickly.

-Effects of Embodiment-As described above, according to this embodiment, the environmental condition of the room (11) is specified by a simple thermal state equation, and the parameter calculating means is specified.
Since the control gain is set according to the environmental condition of the room (11) by the (32) and the gain calculating means (33), for example, the room ( A control gain suitable for the environment of 11) can be set. As a result,
Since the indoor temperature Ta does not overshoot the set temperature and the room temperature Ta does not vibrate, stable temperature control can be performed, so that the indoor temperature can reach the set temperature in a short time. it can. In addition, it is possible to improve comfort and reduce energy loss. Also, the above parameter calculation means (32)
However, since the parameters are derived by the method of least squares, reliable parameters can be derived and accurate system identification can be performed.

-Other Modifications- In this embodiment, the parameter calculating means (32) is
Although the parameters are derived by the least squares method, in the invention of claim 1, the parameters may be derived by applying a neural network such as a back propagation algorithm. Further, the gain calculation means (33) has a proportional coefficient k / h, a time delay Ca / h, and a dead time L.
The optimum PID gain or PI gain for the above three factors may be approximated by simulation in advance, or a program for obtaining a rule base may be created to derive the control gain. Also, the controller (30)
In addition to PID control or PI control, fuzzy control may be performed, and in that case, the control gain of fuzzy control is system-identified.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of the present invention.

FIG. 2 is a system schematic diagram showing an air conditioning system.

FIG. 3 is a control block diagram of the air conditioning system.

FIG. 4 is a control flow chart of system identification.

FIG. 5 is a change characteristic diagram of room temperature.

FIG. 6 is a control block diagram of a conventional air conditioning system.

[Explanation of symbols]

 11 Room (air conditioning space) 20 Air conditioner (air conditioner) 30 Controller 31 Dead time detection means 32 Parameter calculation means 33 Gain calculation means 40 Room temperature sensor (internal temperature detection means) 41 Outside air temperature sensor (external temperature detection means) 42 Frequency Sensor (heat input detection means)

Claims (3)

[Claims] 1. An air conditioner (20) for air-conditioning an air-conditioned space (11), and a controller for controlling the operation of the air-conditioner (20) such that the internal air temperature of the air-conditioned space (11) reaches a predetermined value. In the control device of the air conditioner comprising (30), an internal temperature detecting means (40) for detecting an air temperature inside the air-conditioned space (11), and an air temperature outside the air-conditioned space (11). An external temperature detecting means (41) for detecting the heat input, a heat input detecting means (42) for detecting a heat input from the air conditioner (20) to the air-conditioned space (11), and a waste of the air conditioner (20). Dead time detection means (31) for detecting time, internal air temperature and external air temperature detected by the internal temperature detection means (40) and external temperature detection means (41), and heat input detection means (42) The air-conditioned space based on the heat input (11)
Parameter calculating means (32) for deriving a plurality of parameters of the thermal state equation that specifies the environmental state of, the parameter derived by the parameter calculating means (32) and the dead time detected by the dead time detecting means (31) And a gain calculation means (33) for deriving a control gain of the controller (30) and outputting the control gain to the controller (30). 2. The control device for an air conditioner according to claim 1, wherein the heat input detection means (42) is configured to detect the frequency input F of the air conditioner (20) corresponding to the heat input q. On the other hand, the parameter calculation means (32) uses the internal air temperature Ta, the external air temperature To, the dead time L, the change amount dTa / dt of the internal air temperature Ta, the heat capacity Ca of the air-conditioned space (11), and the air-conditioned space (11). The heat capacity Ca and the heat transfer coefficient h and the coefficient k in the equation of state specified by the heat transfer coefficient h between the inside and the outside, the frequency input F, and the coefficient k of the frequency input F are derived by the least square method. An air conditioner control device characterized by being provided. 3. The control device for an air conditioner according to claim 2, wherein the gain calculation means (33) detects the heat capacity Ca, the heat transfer coefficient h and the coefficient k derived by the parameter calculation means (32), and dead time detection. A control device for an air conditioner, which is configured to derive a control gain of a controller (30) based on the empirical formula of Chen et al. Based on the dead time L detected by the means (31).