JPS60223940A - Frost detecting device for heat pump type air conditioner - Google Patents

Abstract

PURPOSE:To eliminate mis-detection of transitional frosting condition of an outdoor side heat exchanger by a method wherein existence of frosting is decided based upon whether the time series of a difference between an operation system, estimating a room heating capacity during operation by employing a mathematical model obtained under non-frosting condition, and the other operating system computing an actual room heating capacity generates white noise or not. CONSTITUTION:The room heating capacity Q(K) can be estimated by the mathematical model (formula 1) in case the operation of the air conditioner is stable and the outdoor side heat exchanger is not being frosted. On the other hand, the actual room heating capacity Q(K) can be calculated by the formula 2 from the blow-off temperature TBL(K) of air blown off out of an indoor side heat exchanger, the suction temperature TIN(K) of air sucked into the same heat exchanger and the flow amount W(K) of air passing through the indoor side heat exchanger. Here, in case the operating air conditioner is not being frosted, the room heating capacity, calculated from the previously obtained mathematic model (formula 1) coincides with the actual room heating capacity when the formulas 1, 2 are compared, and the non-coinciding part or the time series of estimated error becomes the white noise while, on the contrary, it is judged that the air conditioner is being frosted in case the time series does not generate the white noise.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a frost detection device for a heat pump air conditioner that detects frost on an outdoor heat exchanger during heating operation of the heat pump air conditioner.

FIG. 1 is a conventional heat ponzo system capable of air conditioning and heating. FIG. 2 is a circuit diagram of the defrost device shown in FIG. Since the present invention, which will be described later, relates to frost detection, in explaining the conventional art, explanations regarding the cooling operation and the room temperature control section will be omitted.

First, in the refrigeration circuit shown in Fig. 1, in heating operation,
The refrigerant compressed by the compressor 1 is condensed in the indoor heat exchanger 4 via the four-way valve 2, adiabatically expanded in the expansion valve 7, evaporated in the outdoor heat exchanger 8, and then passed through the pipe 13 and returned to the indoor heat exchanger 4. It passes through the four-way valve 2 and returns to the compressor 1 via the accumulator 12.

In this case, the heating effect is obtained when the refrigerant is condensed in the indoor heat exchanger 4, but if the heating operation continues for a while, the heat exchange rate with the outside air will increase. As a result, heating capacity decreases.

Therefore, the temperature detector 11 detects whether frost has adhered to the outdoor heat exchanger 8, and the defrost device 15 drives the solenoid 3 to switch the four-way valve 2 to the cooling operation side.

In this way, when the refrigeration cycle is switched from heating operation to cooling operation, the high temperature and high pressure refrigerant gas compressed by the compressor 1 is sent to the outdoor heat exchanger 8 through the four-way valve 2.
The frost that has adhered will melt. In this way, the heat exchange rate that has decreased due to frost is returned to a predetermined value, and efficient heating operation can be continued again.

In addition, 5 in FIG. 1 is an indoor fan, and motor 6
It is driven by Further, 9 is an outdoor fan, which is driven by a motor 10.

Next, the function of the defrost device 15 will be described with reference to FIG. The defrost device 15 includes a solenoid drive power source 16. switching circuits 17 and 18, constant generator 19. The comparator 20 is also constructed. The switching circuit 17 is for switching between the heating side a and the cooling side, and the switching circuit 18 is for switching between the heating side a and the defrost side.

During heating operation, the switching circuit 17 is connected to the heating side a by the cooling/heating changeover switch 14, and when defrosting is not required, the output of the comparator 20 connects the switching circuit 18 to the heating side a, and the solenoid is driven. Power source 16 is applied to solenoid 3, and four-way valve 2 operates on the heating side.

In the case of cooling operation, the switching circuit 17 is connected to the cooling side by the cooling/heating changeover switch 14, so power is not applied to the solenoid 3 and the four-way valve 2 operates on the cooling side.

The problem occurs when the outdoor heat exchanger 8 is frosted, but in this case, the temperature detected by the temperature detector 11 is compared with the temperature predetermined by the constant generator 19 by the comparator 20, When the temperature of the former becomes υ lower than that of the latter, the output of the comparator 20 connects the switching circuit 18 to the defrost side. In this way, when frost forms, no power is applied to the solenoid 3, and the four-way valve 2 operates on the cooling side.

The outline of the conventional 70 stroke detection is as above, but its drawbacks will be described next. A drawback of the conventional system is what is called a "fake defrost." In other words, "fake def 0
"defrost" means that the defrost operation is performed even when the outdoor heat exchanger 8 is not frosted. This may be due to the following reasons.

(1) The temperature of the outdoor heat exchanger 8 may temporarily become the temperature when frost has formed even though no frost has formed.

(2) The temperature detected by the temperature detector 11 installed in the outdoor heat exchanger 8 is the temperature of a certain portion, and does not indicate the frosting state of the entire heat exchanger.

(3) Only one point (constant generator 1
9), and even if it is an appropriate value for a certain outside temperature, it is an inappropriate value for other outside temperatures. In other words, frost formation differs depending on the outside temperature.

(4) Since the state of frost formation is also affected by the humidity of the outside air, it is difficult to detect frost formation based only on the temperature of the outdoor heat exchanger 8.

This invention has been made to eliminate the above-mentioned drawbacks of the conventional technology, and uses a mathematical model established in advance for the non-frost state of the outdoor heat exchanger during heating operation of a heat pump type air conditioner. A calculation system that estimates the heating capacity during operation, a calculation system that calculates the actual heating capacity during operation, and whether or not the time series of the difference between the two calculation systems becomes white noise or not is used to determine the presence or absence of frost formation. To provide a 70 stroke detection device for a heat pump type air conditioner, which eliminates "false defrost", performs well-timed defrost operation, and performs efficient heating operation by being configured with a calculation system that performs With the goal.

Next, an embodiment of the frost detection device for a heat pump type air conditioner according to the present invention will be described based on the drawings. Third
The figure is a block diagram showing the configuration of one embodiment.

In this Fig. 3, 10 is the compressor rotational speed signal Ucp.
(K 1), 102 is the indoor fan rotation speed signal UPI (K
-1), 103 is the outdoor fan rotation speed signal UFO (K 1
), 104 is the expansion valve opening signal UEX (K 1), 105
indicates the outside air temperature signal U70 (K1). Here (K-
1) means time point (K-1), and 1 from point K
It represents a time in the past. This one point difference is the sampling period τ.

106 is a blowout temperature signal TBL (6) of the air blown out from the indoor heat exchanger, 107 is a suction temperature signal TIN (6) of the air sucked into the indoor heat exchanger, 300
respectively indicate the indoor fan rotation speed signal UPI (6),
These are the values at the time.

These signals may be detected by appropriate means, and the present invention is not sensitive to this detection method.

Compressor rotation speed signal 101, indoor fan rotation speed signal 102
, outdoor fan rotation speed signal 103, expansion valve opening signal 104
, the outside air temperature signal 105, the air outlet temperature signal 106, the suction temperature signal 107, and the indoor fan rotation speed signal SOO are inputted to the samples 108 to 113.301, respectively.

The output of sampler ios~112 is each zero-order holder 1
Arithmetic circuit B123 to arithmetic circuit F1 through 15 to 119
27. Also, sampler 1
The outputs of 20 and 121° 300 are input to the capacity calculation circuit 128.

Output signal 130 of arithmetic circuit B123 to arithmetic circuit E126
-133 are sent to adders 201-204, and output signal 134 of arithmetic circuit F127 is also sent to adder 204.

The output signal of adder 20 is sent to adder 200, the output signal of adder 202 is sent to adder 201, the output signal of adder 203 is sent to adder 202, and the output signal of adder 204 is sent to adder 200. The signal is sent to adder 203.

The output signal of the arithmetic circuit A122 and the output signal of the adder 201 are added by the adder 200, and the output signal is the estimated heating capacity Q(K)Jss is input to the arithmetic circuit A122 through the holder 500. It is also input to the subtractor 131.

On the other hand, the heating capacity Q calculated by the capacity calculation circuit 128
(F0136 is sent to the subtracter 137. The subtracter 137 subtracts this heating capacity Q(6) 136 and the sum QC) 137, and sends the result to the delay circuit 1.
38 to the now e-worth vector calculator 139.

The output signal of this Tsuyaworth vector arithmetic unit 139 is sent to an arithmetic circuit G.

A defrost start signal 141 is output from this arithmetic circuit G140.

Figure 4 shows the calculation circuit A122 to calculation circuit F in Figure 3.
127 is a block diagram showing the configuration of 127. FIG. In FIG. 4, the outputs of constant setters 142-145 are sent to multipliers 146-149, respectively.

The multiplier 146 multiplies the output signal of the constant setter 142 by the input signal, and this input signal is applied to the 1-sample delay calculator 15.
3 to be input to the multiplier 147.

The output signal of the 1-sample delay calculator 153 is input to the multiplier 148 through the 1-sample delay calculator 154, and is also input to the multiplier 149 through the 1-sample delay calculator 155.

The output signals of the multipliers 146 and 147 are input to an adder 150, and the output signals of the adder 150 and the output signal of the multiplier 148 are input to an adder 151. The output signal of the multiplier 149 is input to the adder 152.

Note that the arithmetic circuit A122 to the arithmetic circuit F1 shown in FIG.
27 have the same configuration as shown in FIG. 4, but the constant setters 142 to 145 have different constants set for each of the arithmetic circuits A122 to F127.

Moreover, FIG. 5 shows the configuration of the capability calculation circuit 128 of FIG. 3. In this FIG. It has become so. The output signal of this function generator 156 is sent to a multiplier 158.

The subtracter 157 also includes the sampler 120.1 in FIG.
21 output signals are input. The output signal of this subtracter 151 is sent to a multiplier 158. A constant setter 4 is connected to the input terminal of the multiplier 158.
The output terminal of 00 is connected. The output signal of this multiplier 158, that is, the actual heating capacity 136Q(6) is output.

FIG. 6 shows a detailed configuration of the delay circuit 138 in FIG.
The output signal is output from the connection point with 1.

Furthermore, FIG. 7 is a block diagram showing the detailed configuration of the arithmetic circuit G340 in FIG. The defrost start signal 141 is output from the output terminal of the AND circuit 166.

Next, the heat of this invention configured as described above, f?
This article describes the principle of the window air conditioner frost detection device.

The current heating capacity of the air conditioner is Q (
K), and this time series (Q(6); =..., -1, 0, 1, 2,...)
...Consider (1). The time difference between the time point (K-1) and the time point becomes the sampling period of the time series (1) (this is assumed to be τ).

This time series generally exhibits irregular fluctuations due to various causes. This irregular fluctuation is acted upon by the stochastic multiplication law and can be regarded as a stationary ergodic stochastic process.

Assuming that the time series of heating capacity is such a stochastic process, the following hexagonal model can be used as a mathematical model to describe it.

o Ucp(K) + UFI(K) + U
yo(K)*Uax(K)*Uto(K) are variables having the meanings described in the configuration column.

a (c) and lb (c) are constants determined by appropriate means, and serve as the parameters of equation (2).

Similarly, M is a constant determined by appropriate means and is equal to the order of equation (2). e(6) is the prediction error.

It is known that e(6) becomes a white noise sequence if the parameters &(e), lb) in equation (2) and the order M are appropriate.

In other words, equation (2) considers the air conditioner as a dynamic system, and its inputs are compressor rotation speed, indoor fan rotation speed, outdoor fan rotation speed, expansion valve opening degree,
It is modeled by taking the outside air temperature as the output and the heating capacity. (
6) is (2) if it is steady and the outdoor heat exchanger is frost-free.
It can be estimated using the first and second terms of the equation. That is, on the other hand, the actual heating capacity Q(6) is determined by the blowout temperature TntH of the air blown out from the indoor heat exchanger, the suction temperature TIN(6) of the air sucked in, and the temperature of the air passing through the indoor heat exchanger. It can be calculated from the air volume W(6) as follows.

Q(K)==C-W(6)・(TBL(6)-TIN(
6)) ...(6) Here, C indicates the specific heat of air,
It is a known constant.

Here, the following can be said by comparing equations (5) and (6). If the air conditioner actually in operation is frost-free, the heating capacity calculated based on the mathematical model (formula (5)) and the actual heating capacity (value calculated using formula (6)) indicates a match, and the part of the mismatch, that is, the time series of prediction errors, should be white noise.

Conversely, if the time series of prediction errors does not become white noise, then this is evidence that some state change has occurred in the operating air conditioner.

Since we are now comparing the model with no frost,
It is natural to judge this state change as frost formation.

Next, the concrete operation of the present invention will be explained based on the above principle.

(1) First, we will discuss the part of estimating heating capacity using a mathematical model.

Well, assuming that the heating capacity 135 indicates the heating capacity level <K-'1) at the time when the estimated heating capacity 135 was estimated (K-1), this is held at the same value for a sampling period of 7 hours by the holder SOO, and the calculation circuit A! l J 22 input. Constant setters 142, 143, 14 of arithmetic circuit A122
4.145 is the mother meter a(1) of equation (5), respectively.
l m(2) , , IQX, IMI are set.

The arithmetic circuit A122 first multiplies the constant setter 142 and the heating capacity (K-1) by the multiplier 146, and then a(
1)·e(K-1) becomes the output of the multiplier 146.

Also, by passing through the 1-sample delay calculator 153, the heating capacity G (K-1) is calculated as ◇(K-2) and Ka, multiplier 1
47, the constant setter 143 and the heating capacity stand (K-2) are multiplied, and a(2)·e(x-2) becomes the output of the multiplier 147. The adder 150 adds up the outputs of the multipliers 146 and 147, that is, a(1)·e(K-1) ten m(
2) Output -G(K-2)=fia&+tm=1 plan (K-m).

Thereafter, in the same manner, a (1) x (K71 Xi = 3.
4. . . , M) is added to Σa(E)・Noboru(K−m), and finally, the first term Σa(n・e(K -m) is obtained as m=1.

(11) Next, the output of the arithmetic circuit B123 is Σb1GT! in the second term of equation (5). m=1, indicating that l-U, , (Km). Compressor rotational speed signal UcP (K-1) at time (K-1)
) 101 is detected by a suitable detector and the sampler 10
8, the data is sampled every time at the sampling period, and the value at the time of sampling is held by the zero-order holder 115 and becomes an input to the arithmetic circuit B123.

The structure of the arithmetic circuit B123 is exactly the same as that of the arithmetic circuit A122, except that the constant setters 142, 143
, 144.145, the parameter b+(
IL bl(2)+b,(3),b,(goods) are set.

First, if the compressor rotational speed signal 101 is now the compressor rotational speed signal UcP (K-1) at time (K-1), the output signal of the constant setting device 142 and the compressor rotational speed signal Uc, (K-1)
) is multiplied by the multiplier 146, and b, (1)・U,
, (K-1) becomes the output of the multiplier 146.

Also, by passing through the 1-sample delay calculator 153, the compressor rotational speed signal Uc, (K-1) is converted to U, (K-2
), the multiplier 147 multiplies the output signal of the constant setter 143 and the compressor rotation speed signal UcP(K-2), and b, (2) ・UcP(K-2) is multiplied by the multiplier 14
The output will be 7.

The adder 150 sums the output signals of the multipliers 146 and 147, that is, b, (1).UcP(K-1)+, (2).UcP(K-2")=Σb, 6Tl )
-Uo, (K-m) is outputted as m=1.

Similarly, p b, (+) ucP (K-t) (t=
3.4. −..., M) is Σb 1@ ・Uc p (K
m) is added to m=1, and finally, the output signal 130 of the arithmetic circuit B123
As, Σb1(c)・UcPm= in the second term of equation (5)
1 (Km) is obtained.

GiDzO2 is the indoor fan rotation speed signal U, I (K-1)
, 103 as the outdoor fan rotation speed signal U, o (K-1), 1
04 is the expansion valve opening signal U. If X(K-1), 105 is the outside air temperature signal UT, (K-1), the above-mentioned arithmetic circuit B
123, arithmetic circuit C124 and arithmetic circuit D
125, arithmetic circuit E126, output 1 of arithmetic circuit F127
31, 132, 133° However, the constant setters 142, 143, 144, 145 of the arithmetic circuit C124 include the meters b2(1), b2(2), b in equation (5).
2(3), b2(goods), and those of the arithmetic circuit D125 include b5(1), b3(2), b3(3), b
3), those of the arithmetic circuit E126 include b4(1),
b4(2), b4(3), b4(goods).

Those of the arithmetic circuit F127 include b5(1) and b5(2).
).

b5(3) and b5(goods) are set respectively. Here, samplers 1011, 109, 110, 111゜1
12 shall work at the same timing.

ψ Arithmetic circuit A, B, C, D, E, FJ22202.
203.204 and finally calculated as.

[I[] Next, the heating capacity of the air conditioner during operation will be described.

The indoor fan rotational speed signal UFI (6) SOO at the moment detected by an appropriate detector is sampled by the sampler 301 at every sampling period, and the value at the time of resampling is held by the zero-order holder 302. This becomes the input to the function generator 156.

The function generator 156 generates a function W(6)-F(U,,(6)) (F: indicates a function)
...(7) is installed, and the indoor fan rotation speed UF
It functions to convert I(6) into air volume W(K). Therefore, the indoor fan rotation speed signal UFl (6) is generated by the function generator 15.
6, the air volume becomes W(6).

Also, a blowout temperature signal TBL(6) 106 and a suction temperature signal TBL(6) 106 of the air at the moment it is blown out from the indoor heat exchanger are provided.
IN(6) 107 are similarly detected by appropriate detectors, sampled at each sampling period of 1 hour by sunshields 113 and 114, and placed in the zero-order holder 1.
20 and 121 hold the value at the time of resampling and input it to the subtracter 152. The subtracter 157 is TBL
It functions to calculate (K)-TIN(6).

The samplers 113 and 114 operate at the same timing as the sampler 301, and furthermore, the samplers 108, 109, and 11 operate at the same timing as the sampler 301.
It is assumed that 0, 111, and 112 operate at the same timing.

The difference TBL between the air volume W(6) and the blowout temperature and suction temperature is
6) -TIN(6) and the specific heat C set by the constant setting device 400 are multiplied by the multiplier 158, and the heating capacity during operation Q(6) 136 is calculated as Q(6)=c-w(6)・(TBL (6) - Ding 1, (6
)) ...(6) can be calculated.

Next, the subtracter 137 calculates the difference between the heating capacity ◇ (6) 135 at the time point estimated from the value at the time (K-1) using the mathematical model and the actual heating capacity Q (6) 136 at the time point. , we will explain how to obtain the time-series spectrum of this difference.

i! ＋41r 箕I Q 组进中しh奔曙爾之过浙爾之一个这种电联之前进一步处从空间的空间的加热性Q(epsilon(6))，该输入银电路1380输入1样品预计计计器1。
59, 160, 161, the delay circuit 138 has ε(K
), ε(K-1), t(K-2),...
.

Output ε(K-N+1).

That is, the delay circuit 138 generates N prediction error series ε(K-4) (i=o, 1, 2) of the heating capacity.

..., N-1), and from this N time series data, the mean worth vector calculator 139 calculates the prediction error.
Calculates a thousand worth vector.

- The arm worth vector calculator 139 may be a standard one. For example, a power spectrum calculator using a correlation method calculates the t-th order (/=(N-1,
The autocorrelation function Rg(t) up to (t: natural number) is determined here as t=0, 1, 2,..., t, and this autocorrelation function is multiplied by an appropriate window f(t). By Fourier transforming , the power vector 56 (jω) is obtained. That is, here, ω=2πf f; Frequency Therefore, the output of the power spectrum calculator 139 has a time of twice the sampling period τ, that is, 2τ (
A four-worth vector divided into (t+1) frequency components as shown in equation 9) is obtained.

The method of this power spectrum calculator may be other than the one described above, and the present invention is not sensitive to this method.

[IV) Finally, the operation of the arithmetic circuit G140 that determines whether the power spectrum of the prediction error requested by the lower spectrum arithmetic unit 139 is white noise will be explained.

If the power spectrum S(jω) is a safe white noise, it assumes a constant value for all frequency components)1 and becomes S(jω)=K (K is a constant value)...Ql.

However, the time series of prediction errors for heating capacity that we are currently considering contains various types of noise, so even in non-frost conditions,
The 79worth vector does not take a constant value. Therefore, the whiteness is determined as follows: KO≦Sε(jω)...0]).

That is, the Wors vector Sg(jω) obtained by the power spectrum calculator 139 has a certain level K for any of the (t+1) frequency components. If it is above, it is regarded as white noise O. Accordingly, Ko is set in the constant setter 162, and the frequency components of each of (t+1) . If the equation 0]) is established, the output of the AND circuit 166 becomes high level, and this acts as the defrost start signal 141.

If any one of the (t+1) comparators does not satisfy the equation 0), the output of the AND circuit 166 becomes low level, and the defrost start signal 141 is not output.

Next, other embodiments of the invention will be described.

[A] Selection of variables in the mathematical model (1) As shown in Figure 3 or Equation (5), the input variables of the mathematical model include the compressor rotation speed, indoor fan rotation speed, outdoor fan rotation speed, We selected five variables: expansion valve opening and outside air temperature, but other variables, such as
It is also possible to use six variables, including the humidity of the outside air.

(11) If the mathematical model representing the heating capacity is a valid model, the selection of variables may be anything other than the item (1).

CB) Configuration of arithmetic circuit G The arithmetic circuit G140 may be configured as shown in FIG. This figure 8 shows constant setter 162, 167, comparator 163.
, 164, 165.168° 169.170 and AN
It is composed of D circuits 166, 171 and 172.

That is, the output signal of the lower spectrum calculator 139 is input to the comparators 163 to 170;
At step 65, the output signal is compared with the output signal of the constant setter 162, and the respective output signals are outputted to the AND circuit 166.

Further, the comparators 168 to 170 compare the output signal of the constant setter 167 and the output signal of the value vector calculator 139, and send the output signals to the AND circuit 171.

The AND circuit 166 ANDs the output signals of the comparators 163 to 165 and outputs it to the AND circuit 172, and the AND circuit 171 ANDs the output signals of the comparators 168 to 170 and outputs it to the AND circuit 172. Tep, A
The ND circuit 172 ANDs the output signals of the AND circuits 166 and 171 to output the defrost start signal 141.

The arithmetic circuit G140 in FIG. 8 determines that the requested worth vector Sε(Jω) is white noise if it is within a certain level range. That is, if Ko<Sε(jω) kuni, ...if αbori, then Sε(
jω) is white noise. (a) This is the first half of formula.

Ko<Sε(jω)...a Ka The part shown in FIG. 7 is the same as that explained in FIG.

The constant setter 167 is set to (Ko(K,), and the comparators 168, 169.170 calculate 8g(jω)
≦1... If the threshold is established, the AND circuit 171 outputs a high level.

On the other hand, if Equation 09 also holds true, the output of the AND circuit 166 is also at a high level, so the output signal of the AND circuit 172, ie, the defrost start signal 141, is also at a high level.

If any one of the comparators does not satisfy the α formula or the 61 formula, the output signal of the AND circuit 172, that is, the defrost start signal 141 becomes low level, and the defrost start signal is not output. Due to these effects, the whiteness of the Shiny Wirth vector is defined more strictly than in Figure 7, and the configuration in Figure 8, although more complex, has the effect of making a better judgment of whiteness. This is also fine.

As is clear from the above, the principle of this invention is completely different from that of conventional systems. This is due to the whiteness of the time series of the difference between the theoretically estimated heating capacity and the actual heating capacity.

As described above, according to the frost detection device for a heat pump type air conditioner of the present invention, a predetermined mathematical model is used for the frost-free state of the outdoor heat exchanger during heating operation of the heat pump type air conditioner. A calculation system that estimates the heating capacity during operation, a calculation system that calculates the actual heating capacity during operation, and whether or not the time series of the difference between the two calculation systems becomes white noise. Since it is configured with a calculation system that makes judgments, it eliminates false detection of transient conditions of frost formation on the outdoor heat exchanger, and also prevents partial frost formation from being judged as true frost formation. Furthermore, since the outside temperature is taken into consideration in the estimated heating capacity, it is possible to determine changes in the frosting state due to high or low outside air temperatures, and even if an event that causes a "false defrost" occurs, , this can be completely avoided, "false defrost" can be eliminated, and the defrost operation can be performed with good timing.

[Brief explanation of the drawing]

Figure 1 is a circuit diagram showing the refrigeration circuit of a conventional heat pump type air conditioner, Figure 2 is a circuit diagram showing the detailed configuration of the defrost device in the refrigeration circuit of the heat pump type air conditioner shown in Figure 1, and Figure 3. 4 is a block diagram showing the configuration of an embodiment of the defrost detection device for a heat pump type air conditioner according to the present invention, and FIG. FIG. 5 is a block diagram showing the detailed configuration of the capacity calculation circuit in the frost detection device of the heat pump air conditioner shown in FIG. 3, and FIG. FIGS. 7 and 8 are block diagrams showing the detailed configuration of the delay circuit in the frost detection device, respectively. FIGS. be. 108-114, 301... sampler, 115-1
21.302... Zero-order holder, 122-127°14
0...Arithmetic circuit, 128...Ability arithmetic circuit, 137
．． 157...Subtractor, 138...Delay circuit, 139
...power spectrum calculator, 146-149.20
0 to 204--Adder. Patent Attorney Patent Attorney Takehiko Suzue Figure 1 Figure 2 5 7--' Figure 5 128 Figure 6 38 Figure 7

Claims (1)

[Claims] A calculation system that estimates the heating capacity of a heat pump type air conditioner during heating operation using a preset mathematical model when the outdoor heat exchanger is in a non-frosting state, and a calculation system that estimates the actual heating capacity during operation. Frost detection in a heat pump air conditioner characterized by consisting of a calculation system that performs calculations and a calculation system that determines the presence or absence of frost formation by determining whether or not the time series of the difference between the two calculation systems becomes white noise. Device.