JPS61220909A - Air-conditioning device for automobile - Google Patents

Abstract

PURPOSE:To enhance the responsiveness of control, by using an addition integration type optimum regulator as an air-condition control means which carries out the feed- back control of a blow-out air control means so that internal variables including the optimum feed-back gain of the feed-back control may be changed in accordance with various values of environmental conditions. CONSTITUTION:In an air-conditioning device for automobiles, an air-condition control means M4 carries out the feed-back control of an blow-out air control means M2 for controlling various values including at least the temperature and air volume of air blown into the passenger's compartment M1 so that the internal temperature detected by an internal temperature detecting means M3 reaches a desired temperature. In this arrangement, the air-condition control means is composed of an addition integration type optimum regulator for determining a feed-back value in accordance with an optimum feed-back gain which is set in accordance with a dynamic model for an air-conditioning system. Further, there is provided an arrangement such that internal variables including the optimum feed-back gain are changed over, corresponding to a change in the above-mentioned model, in accordance with environmental conditions including the quantity of solar radiation, detected by an air-condition detecting means M5.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to an air conditioner for an automobile, and more specifically, it adjusts the temperature inside the vehicle to a set target temperature based on a dynamic model of the system that performs the air conditioner. The present invention relates to an air conditioner for an automobile that performs feedback control suitable for controlling the air conditioner.

[Prior art] Air conditioners have been used to control the temperature, humidity, cleanliness, etc. inside the vehicle interior in order to make the environment comfortable for the passengers. Devices that control the temperature inside the vehicle interior are now widely used. In these automotive air conditioners, in order to control the temperature of the blown air over a wide range from low to high temperatures, a cooler (such as an evaporator) is placed upstream of the ventilation passage to cool the blown air, and then It is further heated by a heater (heater core, etc.) to obtain the required temperature of the blown air. A series of devices that perform such blowing, cooling, and heating are referred to as blown air control means, and the whole is called an air conditioning unit. In recent years, air conditioning units for automobile air conditioners that have been widely used include a reheat type that varies the amount of heat supplied to a heater and an air mix type that varies the proportion of air that passes through the heater.

In any case, in these air conditioners for automobiles, the temperature inside the vehicle interior is controlled by the amount of heat that the blown air has, that is, the volume and temperature of the blown air.

The volume of the blown air is determined by the blowing capacity of the blower motor, etc., and the temperature is determined by the cooling capacity of the cooler (evaporator), or in other words, the capacity of the cooling system including the compressor, etc., and the heating capacity of the heater, that is, the reheat type. For the type, it is determined by the amount of hot water circulated, and for the air mix type, it is determined by the damper opening degree of the air mix damper.

When air conditioning is started, the air conditioner detects the temperature in the vehicle interior and performs feedback control on the temperature and air volume of the blown air based on the deviation from the set target temperature. Therefore, the temperature inside the vehicle interior (hereinafter referred to as interior air temperature) gradually approaches the set target temperature due to the amount of heat of the blown air.

Such control is disclosed in Japanese Patent Application Laid-open No. 55-47914, Japanese Patent Application Laid-Open No. 55-77659, and the like.

[Problems to be Solved by the Invention] The conventional device described above is based on feedback control based on the deviation between the inside air temperature and the target temperature so that the inside air temperature approaches and is maintained at the set temperature, and furthermore, the outside air temperature is This system employs predictive control in which the control amount is set to satisfy preset thermal parallelism conditions, taking into account the amount of heat and solar radiation. or,
The amount of air blown has been simply controlled by increasing the amount of air when the above-mentioned temperature deviation is large, and decreasing the amount of air flow Jjt as the deviation becomes smaller.

Therefore, the transient response when changing the set value of the target temperature is not necessarily sufficient, and the transient response may vary depending on the set target temperature, the internal air temperature at that point, the capacity of the air conditioning unit, etc. There have been problems in that the performance may be insufficient, and it may be difficult to maintain a comfortable environment for the passengers. Furthermore, it has been considered that control of internal air temperature may not be optimal when environmental conditions such as the amount of solar radiation in the air conditioning system change.

In addition, the capacity of an air conditioning unit is determined by a combination of airflow volume, cooling capacity of the cooler, heating capacity of the heater, etc., but it is not clear how to combine these to optimally control indoor air temperature. Conventionally, the amount of air flow has been determined by simple combinations such as the above-mentioned control of the amount of air blown, based on the designer's experience. Therefore, it has not always been possible to fully utilize the capacity of the air conditioning unit.

Therefore, the present invention was made with the aim of solving these problems, and it is an object of the present invention to provide an air conditioner for an automobile that optimally controls the temperature inside the vehicle interior (inside air temperature) by maximizing the ability of the air conditioning unit. purpose.

[Means for solving the problems] In order to achieve the above object, the present invention has the following configuration as a means for solving the problems. That is, as shown in FIG. 1, there are: a blown air control means M2 that controls various quantities of air blown into the vehicle interior M1, including at least the temperature and air volume; and an inside air temperature detection means that detects the temperature of the vehicle interior M1. M3
and an air conditioning control means M4 that performs feedback control of the blown air control means M2 so that the detected inside air temperature becomes a set target temperature. The air conditioning condition detection means M5 detects various quantities including at least the amount of solar radiation as the air conditioning environmental conditions that change the dynamic behavior, and the air conditioning control means M4 includes: a dynamic model of the air conditioning system; The regulator is configured as an additional integral type optimal regulator that performs the feedback control based on a predetermined optimal feedback gain according to the above, and further, according to various environmental conditions of the air conditioning system detected by the air conditioning condition detection means M5, This is the configuration of an air conditioner for an automobile, characterized in that it is configured to switch internal variables including at least an optimal feedback gain of the additional integral type optimal regulator in response to changes in the dynamic model.

Here, the blown air control means M2 substantially corresponds to the air conditioning unit described in the section [Prior Art], and is composed of at least means for controlling the temperature and air volume of the blown air. For example, if we take the air volume as one of the various quantities of Suita air, we will consider the blower motor, sirocco fan, etc. that control the air volume by the number of rotations, the opening of the throttle, etc., and if we consider the temperature of the blown air, we will consider the air cooler, etc. There are actuators that control the cooling capacity of the evaporator, and actuators that control the opening degree of the air mix damper or the amount of heat supplied to the heater (heater core). Actuators that control the capacity of a cooler include those that change the capacity of a compressor to vary its capacity, and actuators that control the flow rate of refrigerant.

The air conditioning control means M4 usually uses a microprocessor and controls the RO
M3 is realized as a logic operation circuit configured with peripheral elements such as RAM and input/output circuits, and the target temperature and inside air temperature detection means M3 are set according to pre-stored processing procedures.
Based on the detected internal air temperature, the blowout air control means M2 is controlled by a feedback amount determined from an optimum feedback gain determined in advance according to a dynamic model of the air conditioning system.

That is, the air conditioning control means M4 is configured as an additional integral type optimum regulator that determines the optimum feedback amount of various quantities of the blown air controlled by the blown air control means M2 so as to bring the inside air temperature closer to the target temperature.

The air conditioning condition detection means M5 detects various quantities including at least the amount of solar radiation, such as outside temperature, humidity, vehicle speed, rotational speed of the internal combustion engine, or efficiency of the heat exchanger. The various quantities to be detected may be appropriately selected and configured according to the mode of the harmonizing device. These various quantities as environmental conditions of the air conditioning system are factors that change the dynamic behavior of the air conditioning system, and are variables that cannot be controlled by the air conditioning control means M4, that is, they cannot be controlled by feedback control. These are various quantities that become disturbances for the system in which this occurs.

The air conditioning control means M4 is configured to switch internal variables including at least the optimum feedback gain of the additional integral type optimum regulator according to various amounts of air conditioning environmental conditions detected by the air conditioning condition detection means M5. . The internal variables of the additive integral type optimal regulator are:
There are parameters of the state observation unit (observer), etc., and these take various forms depending on the configuration of the additive integral type optimal regulator. Therefore, next, the configuration of the additional integral type optimal regulator will be explained.

The method of configuring such an additive-integral optimal regulator is well known, for example, in Katsuhisa Furuta's [Linear System Control Theory] (1976), Shokodo, etc., so here we will give a general perspective on the actual configuring method. However, we will provide some supplementary information regarding internal variable switching. In addition, in the following explanation, F,
X, A, [3, C.

It, J, V, Ll, L, G, Q, IR, P indicate a vector (matrix), the subscript T such as AT indicates the transpose of the matrix, the subscript °1 such as A-1 indicates the inverse matrix, Furthermore, a subscript such as X indicates that it is an estimated value, and a symbol such as a shadow indicates that it is another system generated by transformation etc. from the system to be controlled, here a state observer (hereinafter referred to as an observer). The symbol *, such as y*, indicates the quantity being treated, and the symbol *, such as y*, indicates the target value.

In the control of a system related to the controlled object, here the internal temperature, the dynamic behavior of this controlled object is expressed as x (k) = A-x (k-1) + 1B-u (k-
1)-(1)V(k-1>=C-X(k-1>
It is known from modern control theory that it can be described as −(2>.Here, equation (1) is expressed as state equation
is called an output equation, X(k) is a state variable representing the internal state of this system, u(k) is a vector consisting of various quantities of the blown air controlled by the blown air control means M2, and 1( k) is a vector consisting of various quantities indicating the output of this system. Note that in the air conditioning system handled by the present invention, this output vector V (k ) is only the indoor air temperature, so it will be treated as a scalar quantity y (k ) below.

Also, formula (1).

(2) is written in a discrete system, and the subscript indicates the current value, and k-1 indicates the value at the previous sampling time.

The state variable quantity X(k) indicating the internal state of a system that controls air conditioning, in this case indoor air temperature, indicates information about the history of the system that is necessary and sufficient to predict future effects on the control system. There is. Therefore, the temperature inside the vehicle cabin where air conditioning is performed by the blowing air control means M2 (inside air temperature)
If we can clarify the dynamic model of the system in which
), the indoor air temperature can be optimally controlled. Note that in the servo system, it is necessary to expand the system, but this will be described later.

However, for complex objects such as air conditioning, it is difficult to accurately obtain a dynamic model theoretically, and it is necessary to determine it experimentally in some way. This is a model construction method called system identification, and when an automobile air conditioner is operated in a predetermined state, it is assumed that a linear approximation holds in the vicinity of that state, and equations (1) and (2) are used. The model is constructed according to the state equation of It is possible to make accurate approximations and define individual dynamic models.

If the controlled object is one for which a physical model can be constructed relatively easily, system identification is performed using methods such as frequency response method or spectral analysis method, and a dynamic system model (here, vector A, B and C), but in a multi-component controlled system such as the air conditioning system discussed here, it is difficult to create a physical model that approximates to a certain extent, and in this case, the minimum Dynamic models are constructed using multiplication, auxiliary variable methods, online identification methods, etc.

In this case, if the environment of the air conditioning system, such as the amount of solar radiation or humidity, changes, the dynamic behavior of the system will change. Therefore, the system identification described above must be performed separately for each range of environmental conditions that can maintain the same dynamic behavior of the system, and for each range a dynamic system model, that is, vectors A and B
, C will be determined.

Once the dynamic model is determined, the amount of feedback is determined from the state variable IX(k), the internal air temperature y(k), and its target temperature y''(k), and the control amounts of the various Mu(k) of the blown air are theoretically determined. is optimally determined.

Normally, in an automobile air conditioner, various quantities directly involved in controlling the inside air temperature include, for example, the amount by which the amount of air blown by a blower motor affects the inside air temperature, that is, the amount of air blown that contributes to the inside air temperature, converted into temperature. For example, we can use the amount by which the opening degree of the air mix damper affects the inside air temperature4 and treat this as the state variable quantity X(k).
Most of these quantities cannot be directly observed.

Therefore, in such a case, a means called a state observer is configured in the air conditioning control means M4, and the state variables of the air conditioning system are X(k) can be estimated.

This is the so-called observer in modern control theory,
Various observers and their design methods are known. These include, for example, ``Mechanical System Control'' by Katsuhisa Furuta et al.
(1982), etc., and can be designed as a minimum dimension observer or a finite-settling observer in accordance with the aspect of the applicable control object, here an air conditioner for an automobile.

The observer is originally intended to estimate the state variable amount representing the internal state of the air conditioning system, so it is a dynamic model of the system (here, vector A).

If B and C) change depending on the environmental conditions in which air conditioning is performed, the internal parameters of the observer must be changed in the same way. Therefore, when the air conditioning control means M4 is configured as an additional integral type optimal regulator having a state observation section (observer), the above parameters are determined based on various air conditioning environmental conditions detected by the air conditioning condition detection means M5. It may be configured to switch between the two.

In addition to the observed state variable quantity or the state variable quantity X(k) estimated by the above-mentioned observer, the air conditioning control means M4 generates a cumulative value of the deviations between the set target temperature and the actual inside air temperature. In the system expanded using the above, an optimum feedback amount is determined from both of them and a predetermined optimum feedback gain, and the blowing air control means M2 is controlled.

The cumulative value is a necessary amount because the set target temperature changes depending on the driver's operation and requests from the automatic air conditioner. In general, servo system control requires control that eliminates the steady-state deviation between the target value and the actual control value, and it is said that this requires including 1/S (9th order integral) in the transfer function. . In addition, in the case where the transfer function of the system is determined by system identification as described above and the equation of state is established from this, it is desirable to include such an integral quantity from the viewpoint of stability against noise.

In addition, since the internal variables of the additive integral type optimal regulator are switched within a certain range in response to changes in environmental conditions that change the dynamic behavior of the air conditioning system, control due to errors caused by this is changed. From the point of view of absorbing disturbances,
It is preferable to obtain such an integral amount and configure the air conditioning control means M4 as an additional integral type optimal regulator.

In the present invention, it is sufficient to consider beam=1, that is, linear type integral. Therefore, if we expand the system by adding this cumulative value to the state variable IX(k) mentioned above, and determine the amount of feedback by both of them and a predetermined optimal feedback gain F, we can obtain an additive integral type optimal regulator. The control amount to be controlled, that is, the various operating conditions of the blown air controlled by the blown air control means M2 are determined.

Next, the optimal feedback gain will be explained. In the optimal regulator with the integral amount added as described above, it has been clarified how to obtain the control input (here, various amounts of air blown from the air conditioning system) that minimizes the evaluation function J, and the optimal feedback The gain is also the solution of Rikatsuchi equation, state equation (1), output equation (2) A, IB,
It is known that it can be obtained from the C matrix and the weight parameter matrix used for the evaluation function (see above). Here, the weight parameter is initially given arbitrarily, and the evaluation function J changes the weight that restricts the behavior of the various amounts of air blown out of the system in which air conditioning is performed. Optimum values can be determined by performing a simulation using a large-scale computer by giving arbitrary weight parameters, and repeating the simulation by changing the weight parameters by a predetermined amount based on the behavior of the various quantities of blown air obtained. As a result, the optimal feedback gain ``is also determined.

Of course, since this optimal feedback gain takes a different value depending on the dynamic model of the system (vectors A, [B, C), it can be assumed that the dynamic model of the air conditioning system maintains the same identity. The optimum feedback gain F is determined for each predetermined range of environmental conditions that affect air conditioning, and the optimum feedback gain is switched in accordance with various quantities detected by the air conditioning condition detection means M5.

As explained above, the air conditioning control means M4 of the automotive air conditioner of the present invention is configured as an additive integral type optimal regulator using a dynamic model of the air conditioning system determined in advance by system identification etc. However, the observer parameters, optimal feedback gain, etc. within the system are all determined in advance by simulation for each predetermined range of air conditioning conditions.

In the above explanation, the state variable quantity X(k) was explained as a quantity representing the internal state of the air conditioning system, but it is also a variable quantity corresponding to an actual physical quantity, such as the rotation speed and opening degree of a blower motor. etc., or it can be designed as a vector quantity consisting of various quantities converted as quantities directly related to the internal air temperature as described above.

[Function] The automotive air conditioner of the present invention having the above configuration is configured as an additive integral type optimum regulator based on the set target temperature inside the vehicle interior and the inside air temperature detected by the inside air temperature detection means M3. The air conditioning control means M4 calculates the optimum feedback amount so that the internal air temperature reaches the target temperature, and the blown air control means M2 functions to perform feedback control of various amounts of blown air. However, in the automotive air conditioner of the present invention, the various environmental conditions of the air conditioning system are detected, and the internal variables (for example, the observer parameters, etc.).

Therefore, the inside air temperature is not controlled by simple feedback control or predictive control based on the deviation from the target temperature, but is controlled to approach the target temperature by optimally controlling the state of the blown air. It works to achieve optimal control regardless of changes in the environmental conditions in which air conditioning is performed, such as the amount of sunlight or humidity.

[Example] Next, an example of the present invention will be described in detail based on the drawings. Fig. 2 is a schematic configuration diagram showing an automotive air conditioner according to an embodiment of the present invention, Fig. 3 is a control system diagram showing a control model of the air conditioning system, and Fig. 4 is a block diagram used to explain system identification. 5 is a signal flow diagram thereof, FIG. 6 is a block diagram showing the configuration of the observer, and FIG. 7 is a flow chart showing an example of control executed in the electronic control circuit. I will explain them in order.

In FIG. 2, 1 is a blower motor 3. Evaporator 5
．． The air conditioning unit is configured as an air mix type mainly including a heater core 7, an air mix damper 9, etc., and 10 is an inside air temperature sensor 12 for detecting the inside air temperature TR. A temperature setting device 142; a passenger compartment equipped with a solar radiation sensor 16 for detecting the amount of solar radiation; 20, an air conditioning control means M for controlling the air conditioning unit 1;
An electronic control circuit as 4 is shown in each figure.

In the air conditioning unit 1, the air sucked in by the pro-armotor 3 via the inside/outside air switching damper 21 is cooled once by passing through the evaporator 5, and then a part of it passes through the heater core 7 and is heated again. 7
It is mixed with air that does not pass through and is blown into the passenger compartment 10. The ratio of air passing through heater core 7 to air not passing through heater core 7 is controlled by the opening degree of air mix damper 9.

The evaporator 3 includes a compressor 22 and a pipe line for circulating refrigerant.
By controlling the capacity of No. 2, the cooling capacity is controlled. Control of the capacity of the compressor 22, which is powered by an on-vehicle engine (not shown), is built into the compressor 22.
This is performed by an actuator (not shown) that controls the opening area of a passage that communicates the high-pressure chamber and low-pressure chamber.

The electronic control circuit 20 controls the drive voltage of this actuator to control the cooling capacity; hereinafter, the drive voltage of the built-in actuator will be simply referred to as the drive signal (drive voltage) of the compressor 22.

The heater core 7 is configured so that engine cooling water (hot water) (not shown) circulates therethrough, and a certain amount of heat is supplied to the heater core 7 when the engine has finished warming up. Further, the air mix damper 9 is configured such that its damper opening degree is controlled by a damper actuator 24.

The electronic control circuit 20 is a well-known CPU 30. ROM32, R
Mainly AM34 etc., input port 36° output port 38
etc. are connected to each other by a common bus 40 to form a logic operation circuit. The input port 36 receives the inside air temperature TR from the inside air temperature sensor 12, the target temperature TR* from the temperature setting device 14, and the solar radiation fiQ from the solar radiation sensor 16.
Input as an electrical signal corresponding to the amount. Output port 38
outputs a drive signal VB for driving the pro-armotor 3, a drive signal VC for the compressor 22, a drive signal VD for the damper actuator 24, and the like.

The electronic control circuit 20 operates based on signals (TR'', TR, Q, etc.) input from the temperature setting device 14, the inside air temperature sensor 12, etc. according to a program stored in advance in the ROM 32.
Pro motor 3. Compressor 22. The damper actuator 24 and the like are feedback-controlled by drive signals (VB, VC, VD, etc.), and the control model used for feedback control at this time will be described next. In particular, the state equation (1) and the output equation (,
2) etc., how to design an observer based on this, how to find the feedback gain, etc. will be explained based on actual conditions. Fig. 3 is a diagram showing the control system. The control system shown in Figure 3 is actually realized by executing a series of programs shown in the flowchart of Figure 7, and is not realized as a discrete system. There is.

As shown in FIG. 3, the target temperature TRI is first set by the target temperature setting section P1. In this embodiment, the temperature setting device 14 corresponds to the target temperature setting section P1. The integrator P2 accumulates the deviation between the target temperature TR'' and the actual inside air temperature TR to obtain a cumulative value ZTR(k).

P3 indicates a perturbation extracting section that extracts a perturbation component from the inside air temperature TR in a state where steady air conditioning is performed. As mentioned above, in order to perform linear approximation to a nonlinear model, this is the range in which linear approximation holds in the vicinity of multiple steady air conditioning conditions. This is due to the fact that we constructed a dynamic model for this system by considering it as a series of . Therefore, once the inside air temperature TR is changed from the predetermined nearest steady state by the amount of perturbation δTR (=TR-T
It is treated as Ra>. The operating conditions of the air conditioning unit 1 determined by the integrator P2, the observer P4 described later, and the feedback amount determination unit P6, that is, the drive voltage VB of the blower motor 3, the drive voltage VC of the compressor 22, and the air Driving voltage V of the damper actuator 24 that determines the opening degree of the mix damper 9
D is also treated as perturbation valves δVB, δVC2δVD.

The observer P4 estimates the state variable amount X(k) expressing the internal state of the air conditioner from the perturbation valve δTR of the inside air temperature and the perturbation valves δVB, δvC1δVD of the above operating conditions, and obtains the state estimated amount X(k). It is something to seek. As shown in the figure, observer P4 monitors the environmental conditions of the air conditioning system.
Here, in order to switch internal parameters according to the amount of solar radiation Q detected by the air conditioning condition detection means P5, a plurality of sets of parameters are stored in advance.

The air conditioning condition detection means P5 corresponds to the solar radiation sensor 16 that detects the solar radiation amount Q in this embodiment. In this embodiment, only the solar radiation IQ will be considered as the various quantities that change the environmental conditions of air conditioning, but if necessary, the humidity of the blown air detected by various sensors, the outside temperature, etc. can also be detected. , the parameters in the observer P4 and the optimum feedback gain in the feedback amount determination unit P6 may be switched based on these quantities.

State estimate X(k) estimated by observer P4
and the above-mentioned cumulative value ZTR(k''), the optimum feedback gain F is integrated in the feedback amount determination unit P6 to obtain the control amount (δVB, δVC1δVD).This set of control amounts (δVB, δ■ C2δVD)L
Since this is a perturbed valve based on operating conditions corresponding to a steady operating condition selected by the perturbation component extracting section P3, the reference setting values VBa, VCa corresponding to the steady operating conditions are added to this by the reference setting value addition section P7. , VDa, various operating conditions for the air conditioner, VB, VC.

It determines the VD.

The configuration of this control system has been briefly explained above, but the driving voltage VB of the blower motor, the driving voltage ■C of the compressor, and the driving voltage VD of the damper actuator 1 were taken up as an example as the operating conditions of the air conditioner. This is because these quantities are fundamental quantities related to the control of the inside air temperature TR in the automobile air conditioner having the air mix type air conditioning unit 1. Therefore, in this embodiment, the air conditioner is treated as a multicomponent system with three human power and one output. If the automotive air conditioner is a reheat type, you can create other multi-component control models as necessary, such as replacing the water valve control that varies the flow rate of hot water circulating in the heater core with one of the inputs. good.

The hardware configuration of the automotive air conditioner and the configuration of the control system for a three-person, one-output system that controls the output have been described above. Next, the construction of a dynamic model through actual system identification, the design of the observer P4, and how to provide the optimal feedback gain E will be explained.

First, we will build a dynamic model of an automotive air conditioner.

FIG. 4 is a diagram showing a system of an air conditioner that is operated steadily as a three-manpower one-output system using transfer functions 01(2) to G3(Z).

In addition, l indicates 7 conversions of sample values of input/output signals, and Gl
It is assumed that (z) to G3(z) have appropriate orders. Therefore, the entire transfer function matrix G(z) is 6 (z
) = [G1(z) G2(z) G3(z
) is represented by ko.

As with the air conditioner in this example, when the control system is a three-man power system with one output, and there is interference in input and output quantities, it is extremely difficult to define a physical model. becomes. In such cases, the transfer function can be determined by a type of simulation called system identification.

The method of system identification is explained in detail in, for example, "System Identification" by Setsuo Sagara et al. (1981), Institute of Instrument and Control Engineers, etc., but here, identification is performed using the method of least squares.

The air conditioner is operated steadily under a predetermined condition, and the amount of change δ in the compressor 22, damper actuator 24, and drive voltage is calculated.
Both VC1δVD are set to O, and the variation δVB of the drive voltage of the book motor 3 is controlled by an appropriate test signal. At this time, the input δVB and the change in internal air temperature δ as the output
TR data is sampled N times. This is expressed as an input data series (u(i>)=(δVBi) and an output data series (y(i))=(δTRi) (where I=1.2, 3...N). At this time, the system can be regarded as one person with one output, and the system's transfer function G1 (
Z) is Gl (z) = B (z')/A (z °
' ) - (3) That is, Gl (Z) = (bo + bl - Z' +bnZ ) / (1 + a1
-z-f+a2-z +・+an -z )...
(4) It can be found as follows. Note that here, 2'' is a unit transition operator and means z-1-X (k)=X (k-1).

Input/output data series (u (+ >), (V(+
>) to the parameters a1~an, bo~bn of equation (4)
By determining , the transfer function Gl(Z) of the system can be found. In system identification using the least squares method, these parameters a1~
an, bO~bn as Jo=Σ[(y(k) + al −y (k−1)
+-...10an-y(k-n))-(bo-u(k)10bl ・u (k-1)10...10bn-u (k-n)) ]2... (5) is determined to be the minimum. In this example, n=1,
Each parameter was determined. In this case, the signal flow diagram of the system becomes as shown in Figure 5, and the state variable quantity is T[xl
(k) x2 (k)]” and calculate its state・
The output equation is xl (k-1>=z-x1 (k)=-a-x
l (k) +bl −u (k)
...(6) 'j (k)=xl (k)
...It can be expressed as (7). Therefore, if the system parameters A, B, and C when considered as a system with one human power and one output are 'AI −, B1 ', and CI =, then AI
"=-al 131 -=bl...(
8) CI”=1.

Using a similar method, the transfer functions G2(z), G3(2) and the system parameters ^2', A3 for each
", lB2 +, B3 =, C2-, C3" are obtained. Therefore, from these system parameters, the system parameters of the original multicomponent system with three human power and one output, namely the state equation (1), the output equation (2) vectors A, IB,
C can be determined.

If the amount of solar radiation Q is different, the transfer coefficient G (
Z) is also different, so the dynamic model of the system also changes. Therefore, the range in which the identity of the dynamic model of the system can be maintained, for example, the solar radiation Q is less than 250 kcal/h, 250 kcal/h
More than 750kcal/h, less than 1250kcal/h, 1250kcal/h
The above-mentioned system identification is performed by dividing into the above ranges, and the state equation (1) and output equation (
2) Find each set of vectors A, B, and C in advance.

In this way, the dynamic model of this example was obtained by system identification, and when the air conditioner is operated in a predetermined state, a linear approximation is established in the vicinity of this state. It is determined in the form of one. Therefore, for a plurality of steady air conditioning states, each of the transfer functions G1 (z) to G3 (z) is obtained using the above method, and each state equation (1) and output equation (2), that is, the vector A .

[B and C are obtained, and the relationship between their input and output is established during the perturbation amount δ.

Next, a method of designing the observer P4 will be explained. There are Gopinath's design methods for designing observers, which are detailed in ``Basic System Theory'' (1973) by Katsuhisa Furuta and Akira Sano, published by Corona Publishing, etc., but in this embodiment, it is designed as a minimum order observer.

Observer P4 detects the perturbation of the air temperature inside the air conditioner (δTR) and the perturbation of various quantities of operating conditions (δVB, δV
C1δVD) and the state variable amount X inside the air conditioner
The reason why the state estimate X(k) obtained by observer P4 can be treated as the actual state variable X(k) in the control of this system is as follows. Assume that the output X(k) of the observer P4 is configured as shown in the following equation (9) based on the state equation (1) and the output equation (2).

X (k)= (A-L-C)X (k-1>10B-
u (k-1)+IL-v (k-1) (9) In equation (9), [ is an arbitrarily given matrix. Transforming from equations (1), (2>, and (9), [X(k)−
X(k) co=(A-L・C) [X (k-1> -X (k-1
>Ko...(10) Obtain. Therefore, if we select the matrix [ so that the eigenvalues of the matrix (A-[・C) are within the unit circle, we can get X(k) by 1+■
→X(k), and the amount of state variables inside the controlled object X(k
) is input to the control vector u (k ) (that is, the drive voltage of the blower motor 3, etc. [VB (k ) VC (k ) VD
(k > ] ) Output vector 7L/11(k) (
In other words, here, the inside air temperature TR(k) '') as the space ff1y(k) is a series from the past *),
It can be estimated correctly using V (*).

FIG. 6 is a block diagram showing the configuration of the minimum dimension observer. If the observer is configured in this way and the state variable amount inside the observer is assumed to be W(k), then W(k)=P−W(k−1)+M−y(k−1>+J−
u (k-1) ... (11)X (k-1
> =C-W (k-1> +[)-V (k-1>・
It is understood that the state estimate X (k-1) can be obtained as (12). Vector J can be arbitrarily selected under specific conditions, and the speed at which X(k) → X(k) converges can be changed. Here, vector J.

The vector that integrates M is redefined as vector M,
Expression (11) is written as W(k) −P−W(k−1> 10M・[V (k−1)10u (k−1) ]T...
Let it be (13).

As already mentioned, the Gobinas design method is known as a specific design method for such a minimum dimension observer, and this example is used to set the solar radiation amount Q to 1000 kcal.
For a certain steady operating condition of the air conditioner at /h, the following was obtained.

Similarly, when the amount of solar radiation is 0 kcal/h, it is 500 kcal/h.
h, 1500 kcal/h, etc., calculate each vector IP, M, C, [) in a predetermined steady state.

In the observer configured as above, δTB(k), δTC(
k), δTD(k). δT8 (k
δ
Similarly, TC(k) is the drive voltage V of the compressor 22
The perturbation of the actual temperature inside the vehicle that is affected by C is δT
D (k) respectively means a perturbation of the actual temperature inside the vehicle which is also affected by the damper actuator 24. That is, the state estimate X(k) is k(k)=[δTB(k) δT C(k) δTD(k
)]”...(18)

Next, we will explain how to find the optimal feedback gain.
For example, since I am familiar with "linear system control theory" (mentioned above), I will omit the detailed explanation here and only show the results.

Control human power of air conditioning unit 1 u(k) = [VB(k)
VC(k) VD(k)]” and its output y(k)
= TR (k), around a certain stationary point, δu (k) = u (k) −u (k-1) δV
(k)='l/(k kV (k-1)), and finding the optimal control input that minimizes the following evaluation function J, that is, the operating condition u''(k), is an addition regarding the control system of the air conditioner. This will solve the control problem as an integral type optimal regulator.

J=Σ[δy” (k) ・Q・δy (k ) 10 δI
J"(k)・lR−δtJ(k)]...(19) Here, C>, IR indicates the weight parameter matrix, and k indicates the number of samples with O as the control start time, respectively. , the right side of equation (19) is the so-called 2 with G and IR as diagonal matrices.
This is the following formal representation.

At this time, the optimal feedback gain ``F=-(IR
117-1p-flood)-1・蔀T,l)-λ...(20
) is obtained as. Note that A in formula (20).

B is respectively, and P is the solution of the Rikatsuchi Manbaku-style % formula % ( Here, the meaning of the evaluation function J in equation (19) is the various Mu of operating conditions as a control input to the air conditioner. (k) = [VB (k) VC(k)
VD (k)] while restricting the control output y(k), here the target value TR of the internal air temperature TR(k).
The intention is to minimize the deviation from (k)*. The weighting of constraints on various quantities u (k ) of operating conditions can be changed by the value of the weight parameter matrix 0°R. Therefore, the dynamic model of the air conditioner that has already been obtained, that is, the matrices A, B, C (here, A
, B, C), and an arbitrary weight parameter matrix G, IR
is selected, P is obtained by solving equation (23), and the optimal feedback gain is obtained by equation (20). Then, state variable amount X(k) is expressed as state estimation amount X(k) by equation (12>, (
13), so 1k)=F-[x(k) ZTR(k)]”...(
24) The control human power u (k ) for the air conditioner can be determined by the following. Weight parameter matrix Q.

By changing R and repeating the above simulation until the optimal control characteristics are obtained, the amount of solar radiation can be increased to 1000 kc.
The optimal feedback gain for al /h was determined.

was asked like this.

Vector P as a parameter in the observer.

Similar to M, C, and D, the optimal feedback gain is also
Find several types depending on the amount of solar radiation Q.

Above, we have explained how to construct a dynamic model of the control system of an air conditioner using system identification using the least squares method, design an observer with the minimum dimension, and calculate the optimal feedback gain. Then, inside the electronic control circuit 20, actual control is performed using only the results.

Next, the control actually performed by the electronic control circuit 20 will be explained with reference to the flowchart shown in FIG. still,
In the following explanation, quantities handled in actual processing will be indicated with a subscript (k), and quantities handled last time will be indicated with a subscript (k-1).

CPLJ30 starts CPtJ after the air conditioner is started.
After performing initialization processing such as clearing the internal registers of 30 and setting initial control values, the processing of steps 100 to 230, which will be described later, is repeatedly executed according to the procedure stored in the ROM 32 in advance. In this vehicle interior temperature control routine, the above-mentioned vectors P and M are stored in the ROM 32 in advance.

First, in step 100, the output signal of the solar radiation sensor 16 is input through the input port 36, and a process of reading the solar radiation amount Q is performed.

In the subsequent step 110, the output signal of the inside air temperature sensor 12 is inputted through the input port 36, and the temperature inside the vehicle interior, that is, the inside air temperature TR(k) is read. In step 120, the output signal of the temperature setting device 14 is similarly input to read the target temperature TR*(k).

In the following step 130, the deviation between the inside air temperature TR (k) read in step 110 and the target temperature TR'' (k) read in step 120 is calculated as e(k) = T.
R*(k)-TR(k), and in the next step 140, the cumulative value ZTR(k) of this deviation e(k) from the past is calculated. That is, using the repetition time of the process shown in FIG. 7 as the king, ZTR(k)=Z(k-1>+T-e(k)・
...(26) The cumulative value ZTR(k) is obtained. The above steps 130 and 140 correspond to the integrator P2 in FIG.

In the following step 150, when a dynamic model of the air conditioner is constructed from the inside air temperature TR(k) read in step 110, the steady operating state of the air conditioner, which is taken as the range where linear approximation holds, is determined. the closest state (hereinafter, this will be referred to as the stationary point TRa, VBa, VCa, VDa)
). In step 160,
Regarding the inside air temperature TR(k) read in step 110, the perturbation force δT from the steady point determined in step 150
Processing to obtain R(k) is performed. As for this perturbation force, it is assumed that the values, including δTR(k-1), from the previous execution of this control routine are stored. The processing of steps 150 and 160 corresponds to the perturbation extraction section P3 in FIG.

In the following step 170, parameters P, M, and C in the observer corresponding to the current operating state of the air conditioner are determined based on the solar radiation IQ read in step 100 and the steady point selected in step 150.

A process of selecting b, optimal feedback gain F, etc. is performed.

Following step 180. Step 190 is the state estimator X(
k), and the equation (12) %() δTD(k)] is calculated. That is, the variable W(k) in the observer = [Wl (k) W2 (k)]
In step 180, Wl (k).

W2(k), Wl(k)=Pll・Wl (k-1)×P12・W2 (k-
1>10M11-δVB (k-1>10M12-δVC
(k-1>10M13・δVD (k-1) 10M14・
δTR(k-1>W2(k)=P21・Wl (k-1>+P22・W2 (k-
1>10M21-δVB (k-1>10M22-75V
C(k-1>10M23・δVD (k-1>+M24
・δTR(k−1>), and in the subsequent step 190, the result of step 180 is used to estimate the state as δTB(k)=W1(k)+D1・δTR(k)δTC(
k) = W2 (k) +D2 ・δTR(k)δT
Processing is performed to obtain D (k) = δTR (k) - δTB (k) - δTC (k3. Here, δVB (k-1), δVC (k-1), δV used in step 180
D (k-1>, δTR(k-1>, etc.) are the values when this control routine was executed last time, as mentioned above. Also, δTD(, which is one of the state estimation IIX(k)
k), that is, the perturbation valve δVD (k
), the temperature perturbation valve δTD (k) influences the inside air temperature perturbation valve δTR(k) by δTR(k)
-δTB(k)-δTC(k) is calculated because the perturbation valve δTR(k) of the inside air temperature is measured (step 160), so the calculation is simplified in consideration of improving the processing speed. The aim is to

In the following step 200, step 180. Step 1
State estimation amount X(k) = [δT
B (k) δTC(k') δTD(k)]
T and the cumulative value ZTR (k
), using the optimal feedback gain E, the perturbation valve δVB (k) of the drive voltage of the blower motor 3,
Perturbation valve δVC(k) of the drive voltage of the compressor 22, perturbation valve δVD of the drive voltage of the damper actuator 24 (
k) is performed. Figure 7 Step 20
If the formula shown in 0 is expressed as a vector, [δVB(k) δVC(k) δVD(k )]
"="・[δTB (k') δTC(k)δ
TD(k) ZTR(k)]T. This is the feedback amount determination section P6 in FIG.
This is the process equivalent to .

In the following step 210, the perturbation valves δVB(k), δVC(k), δVD(
Adding the values VBa, VCa, and VDa at the steady point to k),
Actual driving voltage VB (k), VC (k >, V
Processing to obtain D (k) is performed. This is the process corresponding to the reference value addition section P7 in FIG.

In the following step 220, each drive voltage VB (k >, vc (k >, vo (k
) through the output boat 38 to the blower motor 3° compressor 22 . Control is performed to output to each of the damper actuators 24. In step 230, the value of the subscript indicating the number of times of sampling, calculation, and control is incremented (updated) by 1, and the process returns to step 100 to repeat the processes of steps 100 to 230 described above.

An example of control performed using the present control routine configured as described above is shown in FIG. 8 in comparison with an example of conventional simple feedback control. As an example of control, from a state where air conditioning is performed and the inside air temperature is in thermal equilibrium at 15°C,
The case where the target temperature of the vehicle interior temperature is changed by 20 degrees Celsius, that is, +5 degrees Celsius, is set. This change in target temperature is shown by a dashed line P in FIG. 8, and the solid line G and broken line F are plots of changes in the inside air temperature based on the output signal of the inside air temperature sensor 12. A solid line G shows an example of control of the internal air temperature according to this embodiment, and a broken line F shows an example of control using conventional control.

As is clear from FIG. 8, this embodiment achieves faster response (start-up) than the conventional control example, has almost no overshoot or undershoot, and brings the inside air temperature to the target temperature. I am able to do that. Comparing the time required for the air conditioning system to stabilize, it can be seen that this example achieves an improvement of more than one order of magnitude despite the rapid start-up. As a result, not only can the temperature inside the vehicle compartment be controlled to the target temperature with good response, but also the blower motor 3. Compressor 22. Damper actuator 2
4 is optimally controlled, no wasted energy is consumed, fuel efficiency is achieved, and since the compressor 22 is not controlled on and off, fluctuations in the output torque of the internal combustion engine can also be reduced.

This is because, in the control of this embodiment, instead of simple feedback control that predicts thermal balance, the control device using the electronic control circuit 20 is configured as an additive integral type optimal regulator.
In other words, the state of the controlled object, the state of the controlled object, and
That is, this is achieved by estimating information regarding the past Iil history of the system that is necessary and sufficient to predict future effects, and using this information to perform control.

Next, control characteristics when the solar radiation IQ changes will be explained. Figure 9 shows that the solar radiation IQ is 0kcal/h at time t1.
It is a graph showing the change in indoor temperature when the temperature suddenly increases from 1000 kcal/h to 1000 kcal/h. In the figure, the solid line Q shows the control characteristics in this embodiment, and the broken line f shows the control characteristics in conventional control.

In this embodiment, it can be seen that the internal air temperature TR of the passenger compartment 10 hardly changes because the dynamic model is immediately switched and control is performed due to an increase in the amount of solar radiation Q.

It can be seen that in conventional control, overshoots and undershoots occur several times before the equilibrium state that is once disrupted due to a sudden change in the amount of solar radiation Q is restored.

Next, in Fig. 10 (Δ>, (B), solar radiation mQ is 0kc).
In the case of al/h and in the case of 1000kcal/h, the target temperature of the passenger compartment 10 is set from 15℃ to 20℃.
It is a graph showing how the inside air temperature TR changes when changed to °C. A solid line G and a broken line F in the figure indicate the control characteristics of this embodiment and those of the conventional example, respectively, as in the above example. In the conventional example, solar radiationff1okcal/h
(for example, at night), the delay in response is noticeable, and when the solar radiation is 11,000 kCat/h, overcontrol such as overshoot is noticeable. In contrast, in this embodiment, since the dynamic model is switched, the responsiveness and
Control with excellent stability has been achieved.

Furthermore, in the automotive air conditioner of this embodiment, the feedback gain in the electronic control circuit 20 that controls the inside air temperature is designed very logically and is optimally determined. Therefore, unlike conventional control devices, there is no need to design based on the designer's experience, make actual adjustments as necessary, and set the feedback gain deemed appropriate. Development man-hours and costs can be reduced.

Although one embodiment of the present invention has been described above, the present invention is not limited to this embodiment, and may be applied to a reheat type air conditioner or use other variables as the state variable X(k). It goes without saying that the invention can be implemented in various ways without departing from the spirit of the invention.

As detailed above, according to the air conditioner for automobiles of the present invention, the temperature inside the vehicle interior (inside air temperature) can be controlled with extremely high responsiveness and followability. It has excellent effects. In controlling the inside air temperature, overcontrol (m) (overshoot or undershoot) hardly occurs, and various quantities of air blown into the vehicle interior, including at least the air volume and temperature, can be optimally controlled. Therefore, the burden on the power source of the automobile air conditioner can be minimized, and when an internal combustion engine is used as the power source, the effect of improving fuel efficiency when the air conditioner is operated can be obtained.

Moreover, the automotive air conditioner of the present invention detects the environmental conditions of the air conditioning system including at least the amount of solar radiation,
As a result, the internal variables of the additional integral type optimal regulator including at least the optimal feedback gain are switched, so that the excellent effect of stably controlling the temperature inside the vehicle even if the environment such as the amount of solar radiation changes is achieved.

Therefore, suitable control characteristics can be achieved over a wide range of environmental conditions for air conditioning.

Additionally, a secondary effect has been obtained in that the number of man-hours required for designing and developing an air conditioner can be reduced.

[Brief explanation of drawings]

FIG. 1 is a basic configuration diagram of the present invention, FIG. 2 is a schematic configuration diagram of an automobile air conditioner as an embodiment of the present invention, and FIG. 3 is a control system diagram of the air conditioning system in the embodiment. Figure 4 is a block diagram used to identify the system model of the example, and Figure 5 is a signal 70- for determining the transfer function.
FIG. 6 is a block diagram showing the configuration of the minimum dimension observer, FIG. 7 is a flowchart showing control as an additive integral type optimal regulator in the embodiment, and FIG. 9 and 10 (A) and (B) are graphs comparing the control characteristics of the embodiment and an example of conventional control, respectively. 1...Air conditioning unit 3...Blower motor 5...
-Evaporator 7...Heater core 10...Crew compartment 12...Inside air temperature sensor 14...Temperature setter 20...Electronic control circuit 22...Compressor 24...Damper actuator 30...CPU 32...ROM

Claims (1)

[Scope of Claims] 1. Blowing air control means for controlling various quantities of air blown into the vehicle interior, including at least the temperature and air volume; and an inside air temperature detection means for detecting the temperature inside the vehicle interior; an air conditioning control means for feedback controlling the blowing air control means so that the inside air temperature reaches a set target temperature; air conditioning condition detection means for detecting various quantities including at least the amount of solar radiation as environmental conditions for air conditioning that change the air conditioning behavior; The regulator is configured as an additional integral type optimal regulator that performs the feedback control based on the optimal feedback gain determined by the air conditioner, and further controls the operation according to the environmental conditions of the air conditioning system detected by the air conditioning condition detection means. 1. An air conditioner for an automobile, characterized in that said internal variable including at least an optimum feedback gain of said additional integral type optimum regulator is switched in response to a change in a model. 2. The control means, using one or more parameters set in advance based on a dynamic model of a system related to air conditioning of the automotive air conditioner, adjusts various amounts of the blown air and the temperature inside the vehicle interior. a state observation unit that estimates a state variable quantity of an appropriate order representing the dynamic internal state of the system; and an accumulation unit that accumulates the deviation between the set target temperature and the detected vehicle interior temperature. and various quantities controlled by the blowing air control means from one or more optimal feedback gains preset based on a dynamic model of the system, the estimated state variable quantity, and the cumulative value. a feedback amount determination section that determines each control amount; and an additive integral type optimal regulator. parameters,
The air conditioner for an automobile according to claim 1, which is configured to switch the optimum feedback gain of the feedback amount determining section. 3. The various amounts of the blown air controlled by the blown air control means include at least the amount of air blown by the blower motor that blows the blown air, the cooling capacity of the blower motor to once cool the air blown, and the amount of air blown by the blower motor. The air conditioner for an automobile according to claim 1 or 2, further comprising: a control amount of an actuator that heats the air again to bring the temperature of the blown air to a predetermined temperature. 4. A patent in which the air conditioning condition detection means detects one or more of the following: outside temperature, humidity, vehicle speed, internal combustion engine rotational speed, or heat exchanger efficiency, in addition to the amount of solar radiation. An air conditioner for an automobile according to any one of claims 1 to 3.