KR20150063184A - Fault diagnosis method of variable Speed Refrigeration System - Google Patents

Abstract

The present invention relates to a fault diagnosis method of a variable speed refrigeration system capable of more rapidly and accurately diagnosing whether or not a refrigerator malfunctions. The fault diagnosis method of a variable speed refrigeration system according to the present invention is a method for diagnosing whether or not a variable speed refrigeration system malfunctions. Whether or not the variable refrigeration system malfunction is diagnosed by analyzing residual information defined as a difference between a real-time electric current applied in real time to the variable speed refrigeration system and an estimated electric current applied to the variable speed refrigeration system when the variable speed refrigeration system functions normally.

Description

{Fault diagnosis method of variable speed refrigeration system}

The present invention relates to a fault diagnosis method of a variable speed refrigeration system.

A refrigerator is a refrigerant circulation type cooling device composed of components such as a compressor, a condenser, an expansion valve, and an evaporator, that is, a refrigerant circulation type cooling device composed mostly of components. Most of the failures are caused by deterioration of a refrigerator, This partial failure progressively progresses and accumulates in a particular area, so that the performance and cooling capacity of the freezer are deteriorated over time, which can cause serious damage to the stored articles and requires a very careful approach Field.

In rural areas, the production period of agricultural products is divided into high season and low season. In peak season, prices are plummeted due to mass production. In low season, prices are skyrocketing. Therefore, the construction of a low-temperature storage facility for controlling the supply-demand imbalance of agricultural products and storing the harvested crops at a low temperature is promoted, and the refrigerator is installed in the low-temperature storage facility. However, Therefore, the economic loss caused by this is huge.

Accordingly, Korean Patent Laid-Open Publication No. 10-2005-0090597 (hereinafter referred to as "prior patent") proposes a method for diagnosing whether or not a freezer breaks down. In the prior patent, the temperature change of the inlet and outlet of the compressor, the condenser and the evaporator is detected, and the fault condition of the refrigerator is diagnosed. However, there is a limit in that the diagnosis speed is somewhat slow in the method of detecting the temperature change and diagnosing the failure. Therefore, it is necessary to develop a method that can diagnose the failure more quickly.

Korean Patent Laid-Open No. 10-2005-0090597 (Title: Fault Diagnosis System for Refrigerators)

It is an object of the present invention to provide a fault diagnosis method of a variable speed refrigeration system capable of diagnosing whether or not a refrigerator is broken down more quickly and accurately.

A method for diagnosing a fault of a variable speed refrigeration system according to the present invention is a method for diagnosing a fault of a variable speed refrigeration system,

The present invention analyzes the residual information defined as the difference between the real time current applied to the variable speed refrigeration system in real time and the estimated current applied to the variable speed refrigeration system when the variable speed refrigeration system operates normally to diagnose whether the variable speed refrigeration system is malfunctioning .

According to the present invention, when the residual information has a positive value, it is diagnosed that the condenser of the variable-speed refrigeration system is defective. If the residual information has a negative value, the evaporator of the variable- It is preferable to diagnose that a defect has occurred.

According to the present invention, it is possible to diagnose the failure of the variable speed refrigeration system in real time from the change of the current applied to the variable speed refrigeration system. Accordingly, it is possible to diagnose the failure of the variable speed refrigeration system quickly and simply.

1 is a schematic block diagram of a fault diagnosis system (FDDS) according to an embodiment of the present invention.
2 is a schematic configuration diagram of the failure diagnosis module.
3 is a configuration diagram showing a method of acquiring residual information in the failure detection unit.
4 is a graph showing a comparison result between the measured current and the current estimated by the NFM (1).
5 is a device diagram for diagnosing faults in a variable speed refrigeration system using current, temperature, and pressure information.
6 is an algorithm relating to failure detection.
7 is a fault identification algorithm according to the reduction of heat transfer area of the condenser.
8 is a graph showing the average residual information according to the reduction of heat transfer area of condensation.
FIG. 9 is a graph for acquiring data for the NFM 2. FIG.
10 is a device diagram for fault diagnosis of a variable speed refrigeration system using current information.
11 to 13 are graphs showing the results of fault diagnosis according to the reduction of heat transfer area of the condenser.
14 is a graph showing a comparison of residual information, pressure, and temperature due to the current immediately after the heat transfer area of condensation is reduced by 50%.
15 is a graph showing the inverter frequency and the opening degree of the electronic expansion valve according to the reduction of the heat transfer area 30% in which condenser contamination is simulated.
16 is a graph showing the condensation temperature due to condenser contamination.
17 is a graph showing compressor discharge pressure due to condenser contamination.
18 is a graph showing residual information due to condenser contamination.
19 and 20 are graphs showing the results of fault diagnosis for condenser contamination of VSRS using only current information.
Figs. 21 and 22 are graphs showing the diagnostic results of fault diagnosis of the evaporator fan failure of the VSRS using the FDDS (2).
23 is a graph showing an ideal Morel's line diagram of the basic refrigeration cycle.
24 is a graph of the change in the ph diagram according to the reduction of the heat transfer area of the condenser.
25 is a graph of the change in ph diagram according to an evaporator fan failure.

Hereinafter, a fault diagnosis method of a variable speed refrigeration system according to a preferred embodiment of the present invention will be described with reference to the accompanying drawings.

First, the refrigeration system is basically composed of an evaporator, a condenser, a compressor, and an expansion valve. In recent years, a variable speed refrigeration system (hereinafter, referred to as VSRS) having a compressor motor driven by an inverter has been increasing to save energy and improve efficiency. In the present invention, it is tried to diagnose the failure of such a variable speed refrigeration system.

Hereinafter, the contents and the results of the present invention will be summarized briefly. Then, the configuration of a fault diagnosis and diagnosis system (hereinafter referred to as " FDDS ") for implementing a fault diagnosis method of a variable speed refrigeration system according to the present invention, A method of diagnosing a failure of a variable speed refrigeration system using a diagnostic system will be described in detail.

The FDDS according to the present invention includes a fault detection part for detecting a fault and a fault diagnosis part for identifying the location of the fault and the extent of the fault using the information obtained from the fault detection part.

The failure detection in the failure detection unit uses the residual information of the current. The residual information is obtained from the difference between the real-time detection current ( I _d ) reflecting the state of the actual system and the estimated current ( I _e ) generated from the current model obtained under various operating conditions in a normal state without a fault. The estimated current is generated in the No Fault Model (NFM) obtained under the normal state. The NFM is obtained by linera regression analysis based on the data obtained from experiments under various operating conditions for the basic refrigeration cycle unit designed for the experiment.

In the present invention, two NFMs are examined. The first is the NFM (1) when the temperature and pressure information are included with the other control variables as independent variables, and the second is the NFM (2) when the temperature and pressure information are excluded from the independent variables. At this time, the dependent variable is set to current in both NFM (1) and NFM (2). Independent variable for NFM (1), it is composed of temperature and seven pressure information including the operation amount of the drive frequency (f), when the NFM (2), consists of an inverter frequency (f) and the electronic expansion valve opening degree (VO) do.

As described later, among the two proposed NFMs, the NFM (2) can eliminate the effect of the time constant of the temperature and pressure information as well as the compensation effect of the feedback controller on temperature and pressure. In addition, by reducing the number of information to be detected, it is possible to minimize the price required for the sensor, the trouble of the sensor, and the possibility that noise accompanies detection information. Therefore, it can be said that NFM (2) is more preferable, and this can be confirmed through the experimental results to be described later.

The fault diagnosis part identifies the location of the fault and identifies the level of the fault. Since the fault diagnosis unit plays a role of fault identification and fault identification using the residual information generated by the fault detection unit, it is necessary to analyze the characteristics of the residual information according to the fault of the heat exchanger. As a result of analysis, I have found. One is the pattern of the residual information, and the other is the size of the residual information. The tendency of the residual information shows a tendency to increase in the direction of positive (+) direction in case of condenser contamination, and to increase in the direction of negative (-) in case of evaporator fan failure.

Also, the size of the residual information increased in proportion to the degree of contamination in case of condenser pollution, and in the case of evaporator fan failure, it decreased rapidly after pan failure simulation. Using the two characteristics of the residual information on the failure of the heat exchanger, an algorithm of the fault diagnosis section including fault identification and fault identification is generated.

In order to verify the effectiveness of the fault detection and fault diagnosis algorithms proposed in the present invention, a fault diagnosis practical test system of VSRS is constructed. Experimental system consists of three parts: VSRS, VSRS compressor, and control unit for controlling the electronic expansion valve and FDDS. The VSRS uses a single refrigeration system and the controller sets the room temperature and the degree of superheat of the system as a control variable so that a feedback controller composed of fuzzy logic is implemented on a programmable logic controller (PLC). Further, in order to drive the stepping motor of the squirrel cage induction motor and the electronic expansion valve of the compressor, a frequency variable device, an inverter, and a step motor drive were used, respectively. FDDS is implemented using PC (personal computer) and DAQ (Data Acquisition, NI) equipment.

In the experiment, the heat exchanger failure simulation of VSRS reduced the heat transfer area artificially in the case of the condenser, and forcedly stopped the evaporator fan at the desired time in the case of the evaporator.

The FDDS guarantees a certain precision, and the smaller the number of detected information, the better the quick responsibility and reliability, the better the performance. Therefore, in the present invention, the NFM (2) excluding the existing temperature and pressure information is finally proposed. In order to compare the performance of the proposed method, we first experimented with NFM (1) with temperature and pressure information, and then experimented with NFM (2) sequentially and then compared the results. In addition, the cause of the current fluctuation at each fault was inferred through conceptual analysis using the p-h diagram for each failure of the heat exchanger.

The terms used in the present invention will be briefly described as follows.

Fault: If the system characteristic or parameter is outside the allowable deviation from the specified condition

 Residual: a failure indication variable based on the deviation of the estimate computed by the measurement and the process model or its deviation

 Fault detection: Determining whether a fault occurs in the system

 Fault isolation: Determination of fault type, location and detection time

 Fault identification: Determination of fault size and time-varying characteristics

 Fault diagnosis: includes fault identification and identification, and determines the type, location, size and detection time of the fault

Hereinafter, the present invention will be described in detail.

1 is a schematic block diagram of a fault diagnosis system (FDDS) according to an embodiment of the present invention. Referring to FIG. 1, the FDDS 100 mainly includes a fault diagnosis module 10 and an evaluation module 20. The fault diagnosis module 10 diagnoses a fault in the event of a fault, takes action according to the result of the fault diagnosis or informs the user.

2 is a schematic configuration diagram of the failure diagnosis module. Referring to FIG. 2, the fault diagnosis module 10 includes a fault detection unit 11 and a fault diagnosis unit 12. The failure diagnosis unit 11 generates residual information and confirms whether or not a failure has occurred. The failure diagnosis unit 12 performs a failure identification and identification of the failure (failure) using the information generated by the failure detection unit 11 . The location of the fault can be detected according to the path of the residual information when various faults occur and the degree of the fault can be grasped by the size of the residual information. The aspect of the residual information is independent of each failure, and the reliability of the failure diagnosis can be secured if the reliability of the residual information is high.

3 is a configuration diagram showing a method of acquiring residual information in the failure detection unit. The purpose of the failure detection unit 11 is to check whether a failure has occurred and to generate residual information for failure diagnosis. As the residual information generating mechanism, the fault detecting unit is very important, so it is sensitive to faults and should be configured to be insensitive to other faults, noise, disturbance, and the like. To do this, the variables must be selected appropriately for the purpose of fault diagnosis and the noise should be minimized when the selected information is measured. Although various methods are used to generate the residual information, the present invention uses a model-based method.

In order to generate a steady state model of the VSRS, the steady state values according to the inverter frequency of the VSRS and the opening degree of the electronic expansion valve are collected, and the values are subjected to a linear regression analysis to generate the NFM. Then, as shown in FIG. 3, the estimated current value by the NFM is compared with the real time current value to generate residual information.

Hereinafter, a system for real-time fault diagnosis will be described. As mentioned above, two systems are introduced according to the fault-free model of the fault detector. First, the fault diagnosis system (NFM1) using current, temperature and pressure will be explained and the fault diagnosis apparatus and algorithm will be introduced. Secondly, the fault diagnosis system (NFM2) using only current information will be described, and the apparatus and algorithm will be explained in the same manner.

1. Explanation of NFM (1)

The most important factor in diagnosing faults is the reliability and responsiveness of fault diagnosis results. In order to improve the reliability, the viscosity may be deteriorated and the reliability may be deteriorated in order to improve the viscosity. However, reliability and flexibility are all important factors in fault diagnosis, so we can not give up one factor. In order to obtain both of these factors simultaneously, it is most important to select the NFM variable to detect the failure. The choice of the NFM's parameters depends not only on the reliability and fault tolerance of fault detection but also on the performance of the overall fault diagnosis.

First, the current, inverter frequency, temperature, and pressure are selected as the variables of the NFM (1). The reason is because we wanted to reflect all the circumstances of the refrigeration system. NFM (1) due to current, inverter frequency, temperature and pressure is shown in Table 1 below.

[Table 1]

Table 1 shows regression analysis data for NFM (1). The inverter frequency and the opening of the electronic expansion valve were kept constant, and representative values of the respective values were measured when the steady state was reached. The measured variable is the I _r, f, VO, T _a, T _o, T _c, T _SH, P _out, P _in the high, of which the dependent variable I _r, independent variables f, T _a, T _o, T _c , T _SH , P _out , and P _in . VO is less influential than other independent variables.

As a result of linear regression analysis, NFM (1) is shown in Equation (1).

...........식 (1)

... (1)

The dependent variable Y represents I _r , the independent variable X ₁ is f , X ₂ is T _a , X ₃ is T _o , X ₄ is T _c , X ₅ is T _SH , X ₆ is P _out , X ₇ is P _in . The coefficient of determination R square of equation (1) is 94.9%, which shows very high fitness. Thus, it can be seen that NFM (1) reflects the steady state of the system.

4 is a graph showing a comparison result between the measured current and the current estimated by the NFM (1). Referring to FIG. 4, the phenomenon that the measured current is divided into a high current and a low current is related to the inverter frequency. High current is measured at the inverter frequency of 60 [Hz], and low frequency is measured at the other frequencies.

5 is a device diagram for diagnosing faults in a variable speed refrigeration system using current, temperature, and pressure information. The refrigeration system consists of a single refrigeration system consisting of a compressor, a condenser, a receiver, an electronic expansion valve, and an evaporator. A feedback controller using fuzzy logic is implemented as a PLC (Programmable Logic Controller) for control, and information necessary for fault diagnosis is acquired by NI-DAQ device for fault diagnosis and programmed as Labview.

The feedback controller changes the compressor rotation speed and the opening degree of the electronic expansion valve so as to keep the room temperature and the superheat, which is the evaporator inlet / outlet difference, constant. By changing the number of revolutions of the compressor, the room temperature is kept constant and energy is saved. And, by changing the opening degree of the electronic expansion valve, the superheat degree is kept constant, so that the efficiency is maximized. A V / f = const control type is used for the inverter for changing the compressor rotational speed, and a stepping motor driver is used for changing the opening degree of the electronic expansion valve. The set values of room temperature and superheat degree are 3 and 6, respectively, and realize capacity control and superheat control independently. The parameters measured for control and fault diagnosis are current, inverter frequency, electronic expansion valve opening, room temperature, superheat, outdoor temperature, condensation temperature, compressor discharge pressure, compressor suction pressure. The current is measured by connecting a CT (Current Transformer) to one of the three phases between the inverter and the compressor.

On the other hand, when measuring current information for fault diagnosis, it is necessary to correct the current value because ripple occurs due to the switching method of the inverter. To stabilize the current value measured by CT, the exponential moving average was averaged using a smoothing filter, and the maximum value per 30 seconds was designated as the representative value ( I _r ). Since the representative value of the current varies depending on the frequency, it is appropriate to determine whether the refrigeration system is malfunctioning or not. Since the inverter frequency and the opening of the electronic expansion valve are determined by the feedback controller, the command value is transferred from the D / A Module of the PLC to the NI-DAQ device. Temperature information is measured using a T-type thermocouple and is collected by a PLC thermal module. The pressure is collected by the PLC's A / D Module using high and low pressure gauges. In order to eliminate the noise of temperature and pressure information, the optimal position of each sensor was selected, and the moving average was transferred to the NI-DAQ every 10 seconds by the PLC, and a smoothing filter was designed and used as Labview. At this time, the exponent of the smoothing filter is set to 50 or 100.

6 is an algorithm relating to failure detection. The failure detection unit generates residual information, which is a measure of the failure diagnosis, and uses the residual information to determine whether a failure has occurred. I _d and I _e are the same in the steady state because the difference between the current I _d measured in real time and the current I _e estimated by NFM (1) is the residual information. However, as to reduce the heat transfer area of the condenser is large difference between the I _d and I _e. Therefore, the residual information becomes large, and when the residual information is the minimum size that can be relied on for the failure, it becomes a criterion for determining whether or not a failure occurs. When the occurrence of the fault is confirmed, the residual information is transmitted to the fault diagnosis section for fault diagnosis.

7 is a fault identification algorithm according to the reduction of heat transfer area of the condenser. Fault diagnosis can be divided into fault identification and fault identification, but fault identification is not necessary because it deals only with faults due to the reduced heat transfer area of the condenser. Therefore, only algorithms for fault identification are described below.

A cycle is required for fault identification. This is because of the reliability of fault identification. The measured current I _d is due to noise from sensors and measuring instruments, and the estimated current I _e is also due to noise from sensors such as temperature and pressure. The judgment period for the fault diagnosis is defined as the fault diagnosis period. The failure diagnosis period in Fig. 7 should not exceed one half of the time until the system becomes dangerous due to a failure. The fault diagnosis cycle of this system is set to 300 seconds as a result of considering the dependability and reliability. It is checked whether the residual information exceeds 0.001A, which is the range of the non-fault state, within 300 seconds. If the time exceeding 0.001A exceeds 150 seconds, the fault is diagnosed. Otherwise, the fault information is output normally. Define the time that is counted for failure identification as t _over . When fault identification is initiated by t _over , the collected residual information is divided by 300 seconds and the average value per second is calculated. The value is defined as _mean residual ( I _mean ). Then, the failure is identified according to the size of the average residual information, and the degree of failure is determined.

FIG. 8 is a graph showing average residual information according to reduction in heat transfer area of condensation, which is data for determining the degree of failure, which is a result of fault identification. After reducing the heat transfer area of condensation to the system to which NFM (1) is applied, the average residual information at that time is collected. When the heat transfer area of condensation is reduced to 10 ~ 70%, the average residual information tends to increase as the heat transfer area decrease increases. However, when the experimental result is 10%, the reliability of the average residual information can not be secured and it is considered that failure detection is difficult. Therefore, the condenser failure corresponding to 20 ~ 30% is defined as Low, the condenser failure corresponding to 40 ~ 50% is defined as middle, and the condenser failure corresponding to 60% or more is defined as High, and applied to the fault identification algorithm.

2. Description of NFM (2)

Previously, we introduced FDDS (NFM (1)) using current, temperature, and pressure information. However, since the temperature and pressure information have a compensation effect by the feedback controller and a large time constant, the fault diagnosis period for ensuring reliability in the FDDS also becomes long. In addition, due to a large number of detection information, it is accompanied by a large number of sensors, which not only raises the price, but also has a high possibility that sensor failure and noise accompany the reliability. Therefore, it seems to be the best way to exclude temperature and pressure information as a parameter for fault diagnosis. In the following, FDDS using only current information is introduced, excluding temperature and pressure information.

In selecting variables for NFM (2), the temperature and pressure information are excluded and current, inverter frequency, and electronic expansion valve opening are selected. This is because the inverter frequency, which is the command value of the controller, and the opening of the electronic expansion valve can provide the fastest information to the change by the feedback controller. In the case of temperature and pressure information, since the time constant is large and the response is slow, the diagnosis is delayed in case of an emergency failure, which can cause the system to be overloaded or extremely inoperable.

Table 2 below shows some of the I _rms , f , and VO for NFM (2) generation. The current information used the rms value. The inverter frequency f and the electronic expansion valve opening VO were fixed, respectively, and the values were recorded when the steady state was sufficiently attained.

[Table 2]

FIG. 9 is a graph for acquiring data for the NFM 2. FIG. 9 (a) shows changes of f and VO . changes in f and VO is a command value for obtaining an effective value of the current information f have caused in 5 [Hz] unit in the range of 40 ~ 60 [Hz], VO is 60 to 100 [%] for each f In the range of 10%.

FIG. 9 (b) shows the change of I _rms according to the change of f and VO . The trend of I _rms was similar to VO rather than f . Therefore, it can be predicted that I _rms will be more affected by VO than f .

When the linear regression analysis is performed to generate NFM (2) using the data in Table 2, the following equation (2) is derived.

................................식 (2)

................................ Equation (2)

The independent variables X ₁ and X ₂ in equation (2) are f and VO , respectively, and the dependent variable Y is I _rms . R Square of Equation (2) is 93.3%, indicating high fitness. Therefore, equation (2) is suitable as NFM (2).

10 shows a device diagram for fault diagnosis of a variable speed refrigeration system using current information. The diagram of the fault diagnosis apparatus of the variable speed refrigeration system is similar to that of FIG. 5 described above, so only the difference will be described. The set room temperature and superheat for the feedback controller were set to 11 and 6 respectively. The parameters measured for control and fault diagnosis are current, inverter frequency, and electronic expansion valve opening. When measuring the current information for fault diagnosis, it is necessary to correct the current value because ripple occurs due to the switching method of the inverter. The ripple information is an abnormal current that occurs during the switching process. Therefore, the inverter frequency reference value is one, but the output value includes various frequencies. There are hardware filtering methods and software filtering methods. In the present invention, a high pass filter and a low pass filter are programmed using Labview to minimize software. In case of high pass filter, it is set to 38Hz. In case of low pass filter, it is set to 62Hz so that the current information of 40 ~ 60Hz can be stably received. Since the inverter frequency and the electronic expansion valve opening are determined by the fuzzy logic in the PLC, the information was transferred from the D / A Module of the PLC to the NI-DAQ device. However, as noise is generated in the transmission process, the smoothing filter is designed as Labview to minimize the noise. Since the information of the inverter frequency and the degree of opening of the electronic expansion valve is not a sensor value such as temperature and pressure information, the number of the smoothing filter is set to 50 because there is not much noise.

The algorithm will be described as follows.

The failure detection unit generates residual information and determines whether or not a failure has occurred. Residual information is generated by the difference of the current I _e and the inverter current I _d which is measured in real-time estimation by NFM (2), determines the failure has occurred is no fault condition, the range of the residual information -0.5 ~ 0.2 [A of ], It is set to send a fault occurrence signal.

The fault diagnosis unit operates by the fault occurrence signal of the fault detection unit. Since the residual information contains the ripple of the inverter, the noise of the measuring instrument, the noise of the NI-DAQ device, and the noise information of the PLC module, the error information can be identified and the fault can be identified by moving averages every 60 seconds . By excluding the temperature and pressure information with the compensation effect by the large time constant and the feedback controller, it is possible to set the period shorter than 300 seconds in the fault diagnosis system by NFM (1). If a fault occurs, the fault is first identified. In the fault identification, the residual information is calculated as the average residual information, which is the residual information per second. When the average residual information rises, the condenser is broken down. If the average residual information rises, the fault location is identified by the evaporator failure. After fault identification is completed, the fault is identified. The algorithm is designed according to the location of the fault. In the case of condenser contamination, the degree of failure is set to 30% if the average residual information is in the range of 0.2 to 0.5 [A], and the system is designed to have the degree of failure at 50% when the average residual information is more than 0.5 [A]. In the case of a failure of the evaporator fan, it is necessary to quickly set the temperature of the system because the system may be overloaded quickly and the indoor environment may be deteriorated due to a rise in the room temperature. Therefore, if the average residual information is less than -0.5 [A], the evaporator fan failure is designed to be output.

Hereinafter, experimental results will be described. Table 3 shows specifications of the variable freezer used in the experiment.

[Table 3]

<NFM (1) Experimental Results>

The fault diagnosis was performed according to the decrease of the heat transfer area of condensation by using FDDS of VSRS using NFM (1).

Table 4 shows the time and amount of reduction of the heat transfer area of the condenser. The heat transfer area of the condenser was variously reduced to 10 ~ 70%. After each failure, the system was allowed to recover to the normal state so that it would not affect the next failure.

[Table 4]

11 to 13 are the results of diagnosis of the fault according to the reduction of the heat transfer area of the condenser. 11 (a) shows I _d and I _e . Depending on the degree of failure, I _d increased and I _e decreased significantly. It can be seen that the influence of the compressor discharge pressure P _out among the independent variables of the NFM (1) is large. If the heat transfer area of the condenser is reduced, the evaporator inlet temperature rises and the superheat degree decreases. Accordingly, the opening degree of the expansion valve is reduced by the feedback controller, the amount of refrigerant is reduced, and the circulation of the refrigerant is not smooth. Also, the room temperature tends to rise because the refrigerating capacity of the refrigerant is reduced due to the generation of the flash gas of the refrigerant, and the compressor rotation speed tends to increase accordingly. The refrigerant expands in the non-condensed state and enters the evaporator. Since a small amount of refrigerant receives a large amount of compressed air, the compressor discharge pressure rises. In the NFM (1), the independent variable X ₆ of the compressor discharge pressure has a negative sign, so that I _e is greatly reduced as shown in (a).

11 (b) shows the residual information and the average residual information according to the reduction of the heat transfer area of condensation. The residual information is generated by the result of (a), and it can be seen that the tendency at 10 to 20% decrease in the heat transfer area of condensation is not significantly different from the steady state. However, it can be seen that the tendency of residual information rises differently from the steady state from 30% reduction of heat transfer area of condensation. As the degree of failure increases, the size of the residual information increases. As a result, average residual information is also increased in the fault diagnosis period.

12 (a) shows the change in pressure due to the decrease in the heat transfer area of condensation. Although the compressor suction pressure did not show a large change, it can be confirmed that the compressor discharge pressure reflects the reduction of the heat transfer area of the condenser.

FIG. 12 (b) shows the change in temperature due to the decrease in the heat transfer area of condensation. It can be seen that the control variables of the feedback controller, the room temperature and the superheat degree, tend to keep constant after the failure. Therefore, it is not appropriate as a parameter of fault diagnosis. The outdoor temperature is also kept constant, which is not suitable as a parameter for fault diagnosis. However, in case of condensation temperature, condenser outlet temperature reflects the decrease in condensation heat transfer area.

Fig. 13 shows the result of fault identification according to the reduction of heat transfer area of condensation. In the fault diagnosis system using NFM (1), the decrease of 10 ~ 20% of the heat transfer area of condensation did not detect the fault, so the fault was not identified. However, it was precisely responded to a decrease of more than 30%, and as a result, it can be confirmed that the fault is accurately identified.

The results of FDDS (1) by the NFM (1) showed that the condensation was more accurate in the case of condensation heat transfer area than 30%, and the fault diagnosis period was 300 seconds. The FDDS (1) It will be sufficient.

14 is a graph showing a comparison of residual information, pressure, and temperature due to the current immediately after the heat transfer area of condensation is reduced by 50%.

14 (a) shows the residual information immediately after the failure, the average residual information, and the fault identification result. It can be seen that the residual information immediately after the failure has risen sharply. The size of the change is very small, about 5mA, but it is stable due to the filter and the size of the current information is very reliable. It can be seen that after the rise of the residual information, the average residual information rises after 300 seconds of the fault diagnosis cycle, and the fault is identified.

14 (b) shows the compressor outlet pressure and the condensation temperature immediately after the failure. Only two of the various pressures and temperatures are presented because the remaining pressure and temperature are difficult to characterize. Compressor outlet pressure and condensation temperature also have information to diagnose faults. The combination of the information makes the residual information, thereby diagnosing the fault. However, the individual information showed a large vibration amplitude due to the compensation effect of the feedback controller. This is a vibration that can be sufficiently exhibited by sudden changes in the load under no fault conditions. Therefore, it is considered that a minimum of 1000 seconds after vibration stabilization is required for diagnosis after fault occurrence. However, it can be seen that the result of the current information is oscillated with a magnitude comparable to the steady state after the occurrence of the fault, and it is easy to use the current information to have more reliability in the fault diagnosis.

On the other hand, VSRS uses a feedback controller. The feedback controller is designed to keep the room temperature and superheat constant. As the heat transfer area of the condenser decreases, the temperature and pressure change. At this time, the room temperature and the superheat also change. The feedback controller changes the inverter frequency and the electronic expansion valve opening to keep these two values constant. This not only keeps the room temperature and superheat constant but also compensates for other temperature and pressure information. To verify the compensation effect of the temperature and pressure information by the feedback controller and to compare the residual information by NFM (2) with the temperature and pressure information, the advantage of the residual information in the fault diagnosis was ascertained. The heat transfer area of the condenser was 30% Lt; / RTI &gt;

15 shows the inverter frequency and the opening degree of the electronic expansion valve according to the reduction of the heat transfer area of 30% in which condenser contamination is simulated. After the failure simulation, the evaporator inlet temperature rose sharply, and the superheat degree dropped sharply. As a result, the opening degree of the electronic expansion valve is reduced by the feedback controller. The inverter frequency also changes to keep the changing room temperature at the set temperature.

Figure 16 shows the condensation temperature due to condenser contamination. The reason why only the condensation temperature is suggested is because the room temperature and the superheat degree are kept constant as a control target of the feedback controller, so that the information for the fault diagnosis is meaningless. After the failure simulation, the condenser did not sufficiently perform heat exchange with the outside, and the condensation temperature rose. The room temperature and superheat of the VSRS changed, and the feedback controller compensated for the temperature and pressure, and the condensation temperature oscillated to a large extent. From 8000 seconds after 500 seconds after the failure simulation, there was no significant difference compared to the condensation temperature before the failure. Also, since the temperature information is sensitive to the surrounding environment, the temperature information for the fault diagnosis of the VSRS can not be reliably obtained as a result of 16.

17 shows the compressor discharge pressure due to condenser contamination. In the case of the suction pressure of the compressor, the change before and after the failure was insignificant. However, compressor discharge pressure was sensitive to condenser contamination. After the failure simulation, the discharge pressure of the compressor increased, but it remained at a constant level after 8000 seconds, just like the condensation temperature. The pressure information is slightly different before and after the failure for the condenser pollution compared to the condensation temperature, but it is not a reliable difference.

18 shows residual information due to condenser contamination. The residual information changes according to the load of the compressor. Since the load of the compressor is directly connected to the state of the refrigerant, the residual information rises after the failure simulation and converges to a constant value by the compensation effect like the temperature and the pressure. However, unlike temperature and pressure information, the residual information showed a difference from immediately after failure simulation to a comparable level before failure simulation. The average value before and after the failure was about 0.2 ~ 0.3 [A]. This value is due to f , VO and I _d of NFM (2), and it is only due to the current information. Therefore, considering the characteristics of the current information, the difference before and after the failure can be sufficiently reliable.

<NFM (2) Experimental Results>

19 and 20 are diagnostic results of the VSRS contamination of the condenser using only current information. The heat transfer area of the condenser is reduced by 30% and 50% in the range of 1000 to 2050 seconds and 4400 to 4900 seconds, respectively.

19 (a) shows changes in inverter frequency and opening degree of the electronic expansion valve. It can be seen that as the degree of failure increases, the opening degree of the electronic expansion valve decreases sharply and the reduction amount increases.

FIG. 19 (b) shows residual information and average residual information due to condenser contamination. It can be seen that as the degree of failure increases, the size of the residual information becomes larger, and the current average residual information to be used for the diagnosis of the fault also becomes larger. It can be seen that the average residual information has sufficient reliability, and this information is used for fault diagnosis.

20 (a) and 20 (b) show fault identification and fault identification results, respectively. Fault identification result Identify faults correctly after fault simulation. Also, the fault identification result accurately identifies the degree of fault after the fault simulation. However, it can be seen that the fault is diagnosed once every 1740 seconds, but it can be solved easily by the filter design to stabilize the fault diagnosis.

As a result of Figs. 19 and 20, it can be seen that the FDDS (2) has sufficient reliability and fast response to VSRS condenser contamination.

Figs. 21 and 22 are diagnostic results of fault diagnosis of the evaporator fan failure of the VSRS using the FDDS (2). I simulated an evaporator fan failure in 1150 seconds.

21 (a) shows changes in inverter frequency and opening degree of the electronic expansion valve. It can be seen that the opening of the electronic expansion valve is reduced due to the decrease of the superheat due to the decrease of the evaporator outlet temperature after the evaporator fan failure, and the inverter frequency is increased with the increase of the room temperature.

FIG. 21 (b) shows the residual information and the average residual information. In the case of an evaporator fan failure, the current residuals show a sharp drop. This is because the refrigerant does not heat-exchange in the evaporator, which reduces the load on the compressor. In the event of an evaporator fan failure, problems may arise in compressor compression, so quick action is required. Residual information diagnoses failures every 60 seconds, providing faster information than temperature and pressure information, minimizing system damage.

22 shows fault identification and fault identification results. If the evaporator fan fails, the refrigeration system will be overloaded and the experiment can not be continued. However, it can be seen that the current information can accurately detect the location and occurrence of the fault even in a short time.

As a result of FIGS. 21 and 22, it can be seen that the fault diagnosis can be performed using the current information for the evaporator fan failure of the VSRS.

As discussed above, the fault diagnosis results using the current information are shown for the evaporator fan failure and the heat transfer area reduction of the condenser with respect to the failure of the heat exchanger. Thus, the principle of changing the current information of a compressor can be understood by considering a toy car with a DC motor that plays with it as a child. For example, if you turn on the power while holding the wheels of the car, the motor will be driven. This is a situation where the load of the motor is so great that the motor does not turn. At this time, the counter-electromotive force of the motor becomes close to 0, and the current becomes close to that, and the motor is driven. Conversely, if there is no load on the motor, the counter-electromotive force increases and the current decreases. This principle also applies to the compressor of the refrigeration system. That is, the increase or decrease of the current effective value is determined by the amount of the compressor.

이며 COP는 압축기 일량을 냉동능력으로 나누어준

이다.Fig. 23 shows an ideal morel curve of the basic refrigeration cycle. Here,

And the COP divides the compressor capacity by the refrigeration capacity

to be.

가 되게 되어 증가하게 된다. 그 결과 압축기 모터의 역기전력이 줄어들게 되고, 전류는 증가하게 된다. 그리고 COP는 압축기 일량은 늘어나고, 냉동능력은 줄어들게 되어 줄어들게 된다.24 is a change in the ph diagram according to the decrease in heat transfer area of the condenser. When the heat transfer area of the condenser decreases, the condensation failure occurs and the point 2 moves to the point 2. At this time, the evaporator inlet temperature rises and the superheat degree is reduced, and the electronic expansion valve opening degree is reduced by the feedback controller. Therefore, the amount of refrigerant flowing through the refrigeration system is reduced, thereby reducing the refrigeration capacity. When the freezing capacity is reduced, the room temperature rises, so that the number of revolutions of the compressor is increased by the feedback controller. As a result, the ph diagram is drawn as 1234. Compressor workload

. As a result, the counter electromotive force of the compressor motor is reduced, and the current is increased. And the COP decreases as the compressor workload increases and the refrigeration capacity decreases.

25 is a change of the p-h line according to the failure of the evaporator fan. If the evaporator fan fails and stops, the refrigerant will not absorb the load in the room. Therefore, point 4 of the evaporator outlet moves to point 4. At this time, the evaporator outlet temperature decreases and the degree of superheat decreases. Therefore, the opening degree of the electronic expansion valve is reduced by the feedback controller, and the amount of refrigerant is reduced. And since the evaporator does not exchange heat, the room temperature rises. As a result, the compressor rotational speed is increased by the feedback controller. Thus, if an evaporator fan failure occurs, the refrigeration cycle is drawn as 1234. Since the amount of compressor work is isentropic, the amount of compressor work is significantly reduced in the region of liquid and gas mixture of refrigerant. Therefore, since the compressor load is reduced, the counter electromotive force of the compressor motor is increased, and the current is reduced.

As described above, in the present invention, a real-time fault detection and fault diagnosis algorithm using a current information for a heat exchanger failure of a variable speed refrigeration system (VSRS) is constructed, a fault diagnosis system (FDDS) is constructed, The validity of FDDS was verified.

In this paper, we propose two models of NFM (1) and NFM (2) for fault detection. The FDDS (1) %, And FDDS (2) showed accurate and quick FDD results for over 30% reduction of the heat transfer area of condenser and evaporator fan failure.

FDDS is very important for reliability and fast response to faults. The performance and reliability of FDDS are determined by the choice of model parameters for FDDS, so these two performances are closely related to each other. NFM (1) used inverter frequency, temperature and pressure information, and NFM (2) used inverter frequency and electronic expansion valve opening information excluding temperature and pressure information. As a result of FDDS test by each NFM, it is confirmed that the reliability of FDDS (1) and FDDS (2) is secured as a result of simulation of heat exchanger failure. However, FDDS (2) Was 5 times shorter than that of FDDS (1). As a result, it was found that the fastness was better. In the NFM (2), temperature and pressure information with compensation effect of large time constant and feedback controller are excluded from the model, and the number of measurement sensors for extracting fault information is reduced from 8 to 1, It is interpreted that the filtering processing time can be minimized. Thus, the FDDS (2) using the NFM (2) can more accurately and quickly perform the FDDS by using only the current information according to the load of the compressor in the heat exchanger failure state of the VSRS.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is clearly understood that the same is by way of illustration and example only and is not to be taken by way of limitation in the embodiment in which said invention is directed. It will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the scope of the appended claims.

100 ... Fault diagnosis system
10 ... fault diagnosis module 11 ... fault detection section
12 ... fault diagnosis unit 20 ... evaluation module

Claims (2)

A method for diagnosing the failure of a variable speed refrigeration system,
The present invention analyzes the residual information defined as the difference between the real time current applied to the variable speed refrigeration system in real time and the estimated current applied to the variable speed refrigeration system when the variable speed refrigeration system operates normally to diagnose whether the variable speed refrigeration system is malfunctioning And the failure diagnosis method of the variable speed refrigeration system. The method according to claim 1,
If the residual information has a positive value, diagnoses that a defect has occurred in the condenser of the variable speed refrigeration system,
And diagnosing that a fault has occurred in the evaporator of the variable speed refrigeration system if the residual information has a negative value.

Images