KR101126706B1 - Method of complex context-aware for status monitoring of semiconductor equipment and system thereof - Google Patents

Abstract

The present invention includes the steps of receiving each sensing value from a sensor node having a plurality of sensors for detecting obstacles of the semiconductor equipment, monitoring whether the sensing value of the sensor is out of the corresponding threshold range, and the sensor If the sensing value of a particular single sensor is out of the corresponding threshold range, analyzing the failure of the specific single sensor and generating an alarm corresponding thereto, and in the state that the failure of the specific single sensor is detected, the sensors Monitoring the sensing value of the other single sensor during a set waiting time, and if the sensing value of the other single sensor is out of the corresponding threshold range, analyzing and responding to a complex failure caused by the specific single sensor and the other single sensor Complex situation recognition method for the status monitoring of the semiconductor device comprising the step of generating an alarm Provide system.
According to the complex situation recognition method and system for monitoring the status of the semiconductor device, it is possible to monitor the failure factors of the semiconductor device in real time to analyze the complex failure corresponding to the problem caused by the complex cause to inform the manager, etc. There is an advantage to prevent the risks in advance and to be able to respond quickly to failures.

Description

Method of complex context-aware for status monitoring of semiconductor equipment and system

The present invention relates to a complex situation recognition method and system for monitoring a state of a semiconductor device. More particularly, the state of a semiconductor device, which enables a real-time grasp of the current situation of an industrial semiconductor device using sensors having a multi-sensing function. The present invention relates to a complex context awareness method and system for monitoring.

The existing semiconductor equipment processing environment has a structure in which dozens of equipments are arranged at dense intervals inside a factory. Since the semiconductor process itself is aimed at a clean system, the plant's interior consists of a clean room, usually with one work manager assigned to each unit.

Most of the work performed on semiconductor equipment is automated, but some processes are handled manually by the task manager. When the manual work is necessary, an abnormal operation may occur in the equipment. The abnormal operation refers to a case in which the semiconductor equipment requires the inspection of the management worker during the normal operation of the automation work. Typically, there are micro vibrations and micro gas leaks.

Semiconductor equipment is initially installed in a vibration-free design when set up, but over time, due to the aging of the diaphragm, which cancels out the vibration of the glass, inadvertent vibrations often occur in the equipment itself or around the equipment. This is a factor that causes defective products in the semiconductor process.

In addition, in the burning test existing inside the semiconductor equipment, a high temperature and high pressure gas is used to clean the semiconductor parts or to test the normal operation of the parts. Most of these gases are toxic gases that are fatal to the human body. There is a danger that it may spread rapidly inside the factory and it may damage the safety of workers.

Therefore, real time check on the abnormal operation is important. Currently, in Korea, a condition monitoring system for dust and abnormal temperature detection is built in a semiconductor factory, but other matters are inadequate. In addition, in the conventional case, since a worker uses a passive state monitoring system that collects information by connecting a PDA-type terminal to each sensor in a 1: 1 manner, it is difficult to monitor in real time a potential failure or abnormality caused by a complex factor. There is a limit.

An object of the present invention is to provide a complex situation recognition method and system for monitoring the status of a semiconductor device that can be informed by real-time monitoring of various factors that are obstacles to the semiconductor device by analyzing a complex failure corresponding to a problem caused by a complex cause. There is this.

The present invention includes the steps of receiving each sensing value from a sensor node having a plurality of sensors for detecting obstacles of the semiconductor equipment, monitoring whether the sensing value of the sensor is out of the corresponding threshold range, and the sensor If the sensing value of a particular single sensor is out of the threshold range, analyzing the failure of the specific single sensor and generating an alarm corresponding thereto, and in the state that the failure of the specific single sensor is detected, the sensors Monitoring the sensing value of the other single sensor during a set waiting time, and if the sensing value of the other single sensor is out of the corresponding threshold range, analyzing and responding to a complex failure caused by the specific single sensor and the other single sensor Complex situation recognition method for the status monitoring of the semiconductor device comprising the step of generating an alarm Provided.

Here, the plurality of sensors may include a plurality of sensors selected from a temperature sensor, a vibration sensor, a pressure sensor, and a gas sensor.

The complex situation recognition method for monitoring the state of the semiconductor device may further include storing a sensing value of the specific single sensor or the other single sensor that is out of a corresponding threshold range.

The monitoring of the sensing value of the other single sensor may be controlled such that the waiting time is reduced as the number of sensors corresponding to the sensing value is in the normal range increases. The complex situation recognition method may further include displaying a result of comparing each sensing value sensed by the sensor with a corresponding threshold range in the form of at least one of an image, a graph, or text.

In addition, the present invention, the receiving unit for receiving each sensing value from a plurality of sensors for detecting the obstacles of the semiconductor device in real time, the first monitor unit for monitoring whether the sensing value of the sensor is out of the corresponding threshold range, As a result of the monitoring, when a sensing value of a particular single sensor among the sensors is out of a corresponding threshold range, the first alarm unit for analyzing the failure of the particular single sensor and generates an alarm corresponding thereto, and of the specific single sensor A second monitor unit configured to monitor a sensing value of another single sensor among the sensors for a predetermined waiting time in a state where a failure is detected, and when the sensing value of the other single sensor is out of a corresponding threshold range, the specific single sensor and the A semiconductor including a second alarm unit for generating a corresponding alarm by analyzing a complex failure by another single sensor That the complex situation for the health surveillance of non-providing system.

The complex situation recognition system for monitoring the state of the semiconductor device may further include a storage configured to store sensing values of the specific single sensor or the other single sensor, which are out of a corresponding threshold range.

In addition, the second monitor may adjust the waiting time to decrease as the number of sensors whose sensing value corresponds to a normal range increases.

The complex situation recognition system for monitoring the state of the semiconductor device may further include a display unit configured to display a result of comparing each sensing value detected by the sensor with a corresponding threshold range in the form of at least one of an image, a graph, or text. It may include.

According to the complex situation recognition method and system for monitoring the status of the semiconductor device according to the present invention, by monitoring the failure factors of the semiconductor device in real time can analyze the complex failure corresponding to the problem caused by the complex cause to inform the administrator Thus, there is an advantage in that anticipated risks are prevented in advance and rapid response to a failure can be made.

In addition, by performing the monitoring for a predetermined waiting time, it is possible to reduce the frequency of alarm generation, increase the power life of the sensor node, there is an effect that does not cause unnecessary warning to the operator. In addition, when the number of sensors whose sensing value is in the normal range increases, the waiting time for monitoring is reduced, thereby reducing the unnecessary monitoring process and saving power of the sensor.

1 is a block diagram of a complex situation recognition system for monitoring the state of a semiconductor device according to an embodiment of the present invention.
2 is a conceptual diagram of FIG. 1.
3 is a flowchart of a complex situation recognition method using FIG. 1.
4 is a more detailed flowchart of FIG. 3.
FIG. 5 is a graph comparing the frequency of alarm occurrences for the case where the method of FIG. 3 is applied and when the method of FIG. 3 is not applied.
6 and 7 are exemplary views of a monitoring screen for the method of FIG.

1 is a block diagram of a complex situation recognition system for monitoring the state of a semiconductor device according to an embodiment of the present invention. 2 is a conceptual diagram of FIG. 1.

Generally, several semiconductor equipments 10 are arranged at close intervals in a semiconductor factory. For the complex situational awareness system 200 of the present invention, one or more sensor nodes 100 for monitoring abnormal operation of the semiconductor device 10 are disposed outside or around the semiconductor device 10.

The sensor node 100 includes a plurality of sensors for detecting obstacles of the semiconductor device 10. Here, the sensors are used as the temperature sensor 110, vibration sensor 120, pressure sensor 130, gas sensor 140. Of course, the type and number of the sensors 110, 120, 130, and 140 are not necessarily limited thereto.

The sensor node 100 collects the sensing values of each of the sensors 110, 120, 130, and 140 and transmits them in a cycle set in the system 200. Hereinafter, the system 200 will be referred to as an analysis server 200.

The analysis server 200 may monitor and analyze sensing values received from the sensor node 100 in a wired or wireless manner, and generate an alarm regarding an analyzed risk factor, potential risk situation, and the like to warn an administrator or a worker in advance. The analysis server 200 includes a receiver 210, a first monitor 220, a first alarm 230, a second monitor 240, a second alarm 250, a storage 260, The display unit 270 and the input unit 280 are included.

3 is a flowchart of a complex situation recognition method using FIG. 1. Hereinafter, a method of recognizing a complex situation for monitoring the state of the semiconductor device will be described in detail.

First, each sensing value is received through the receiver 210 via wire / wireless from the sensor node 100 including the sensors 110, 120, 130, and 140 for detecting obstacles of the semiconductor device 10 (S210). In this case, each of the received sensing values is stored in the storage unit 260. The storage unit 260 stores the sensing values by dividing them by ID of each sensor node 100 so as to easily manage the received sensing values.

Thereafter, the first monitor unit 220 monitors whether the sensing values of the sensors deviate from a corresponding threshold range (S220). In the system 200, a corresponding threshold range is set for each sensor type. The setting of the threshold range may be automatically generated in the system 200 or set by an operator or an administrator through the input unit 280.

When the sensing value of a specific single sensor (Si) of the sensors is out of the corresponding threshold range, the first alarm unit 230 analyzes a failure of the specific single sensor (Si) and generates an alarm corresponding thereto. It generates (S230). The single failure alarm by the step S230 is immediately reported to the terminal (ex, mobile phone, PDA, PC, etc.) of the manager or worker side to enable a quick situation recognition and response.

For example, by detecting the vibration of the semiconductor device 10 itself and the periphery, and alerts the administrator when a vibration of a certain size or more occurs, it warns the expected risk and enables a quick response to the failure.

Of course, the sensing value of the specific single sensor Si outside of the corresponding threshold range may be stored and recorded in the storage unit 260. Of course, subsequent storage of additional sensing values is accumulated and stored, including previously stored information. The stored values can be usefully used for cause analysis.

Next, the second monitor 240 monitors the sensing value of the other single sensor Sj among the sensors for a predetermined waiting time (T seconds) in a state where the failure of the specific single sensor Si is detected ( S240).

At this time, in step S240, as the number of sensors corresponding to the sensing value is in the normal range increases, the waiting time (T seconds) is adjusted to decrease. This is because, if all the sensors remain in the normal range, the probability of occurrence of complex failures is reduced.

That is, as sensors with a normal sensing range increase, the monitoring time (the waiting time) is reduced, thereby reducing the unnecessary monitoring process and saving power of the sensor. On the contrary, as the sensors whose sensing values are not normal increase, the monitoring time (the waiting time) is increased to further increase the monitoring process, and to facilitate the recognition and coping of a complex disorder that may occur later.

After the step S240, when the sensing value of the other single sensor Sj is out of the corresponding threshold range, the second alarm unit 250 combines the specific single sensor Si and the other single sensor Sj. The failure is analyzed to generate an alarm corresponding thereto (S250). The complex fault alarm by the step S250 is also immediately reported to the terminal (ex, mobile phone, PDA, PC, etc.) of the manager or worker side to enable quick situation recognition and response.

Of course, the sensing value of the other single sensor Sj outside the corresponding threshold range may be stored in the storage unit 260. Of course, subsequent storage of additional sensing values is accumulated and stored, including previously stored information.

The display unit 270 displays the result of comparing each sensing value detected by the sensor with a corresponding threshold range as well as the above-described series of processes in the form of at least one of an image, a graph, or text, thereby providing visual monitoring. It can facilitate identification and lead to rapid awareness of the situation.

4 is a more detailed flowchart of FIG. 3. The following complex situation recognition algorithm is performed in the analysis server 200. Hereinafter, with reference to Figure 4, to look at a more specific embodiment of the present invention.

First, an initializing command of the sensors 110, 120, 130, and 140 is received (S410), and initial sensing information is received from the sensors 110, 120, 130, and 140 through the receiving unit 210 (S415). Accordingly, the receiver 210 receives the "received normal condition ACK" and sets an initial threshold range for each of the sensors 110, 120, 130, and 140 (S420).

Thereafter, the first monitor unit 220 monitors the sensing values of the sensors 110, 120, 130, and 140 in real time (S425). Next, it is determined whether a sensing value of the specific single sensor S (i) has occurred (S430). In this case, when there is no sensing value of the generated S (i) as a result of the determination in step S430, the step S425 is repeated.

If the sensing value of S (i) occurs as a result of the determination in step S430, the sensing value of S (i) is stored in the storage unit 260 (S435), and the first monitor unit 220 It is also determined whether the sensed value of the generated S (i) exceeds the corresponding threshold range (S440).

As a result of the determination in step S440, if the sensing value of S (i) is within the normal range without exceeding the threshold range, the step S425 is repeated. On the contrary, when it is determined that the sensing value of the S (i) exceeds the threshold range, the first alarm unit 230 analyzes the failure of the S (i) and generates an alarm corresponding thereto (S445). ).

In a state in which the failure for S (i) occurs, the second monitor unit 240 monitors the sensing value of the other single sensor S (j) for the waiting time (T seconds) (S450). The second monitor 240 determines whether a sensing value of S (j) has occurred during the waiting time (T seconds) (S455). At this time, if there is no sensing value of the generated S (j) as a result of the determination in step S455, step S425 is repeated.

If the sensing value of S (j) occurs as a result of the determination in step S455, the sensing value of S (j) is stored in the storage unit 260 (S460), and the second monitor unit 240 It is also determined whether the sensed value of the generated S (j) exceeds the corresponding threshold range (S465).

As a result of the determination in step S465, if the sensing value of S (j) is within the normal range without exceeding the threshold range, the step S425 is repeated. On the contrary, when it is determined that the sensing value of the S (j) exceeds the threshold range, the second alarm unit 250 analyzes the complex failure for the S (i, j) and generates an alarm corresponding thereto. (S470).

For the embodiment of the S (i, j) corresponding to an example of whether or not a complex situation for monitoring the state of the semiconductor equipment, see Table 1 and Table 2 below. The gas sensor can know whether the gas leaks and the amount of leaked gas, the temperature sensor can know the temperature around the semiconductor device 10, the vibration sensor can detect the vibration through the acceleration sensing, the pressure sensor is air The pressure information can be detected.

Complex Disability Judgment 1: S (i, j) division N2 gas (normal)
200 ppm or less N2 gas (leakage)
200 ppm or more Low temperature
-40 degrees-15 degrees Room temperature
15 degrees-25 degrees High temperature
25 degrees to 125 degrees N2 gas
(normal)
N / A
N / A Wait time (T)
decrease Wait time (T)
decrease Wait time (T)
Add,
Residual Gas Check N2 gas
(leakage)
N / A
N / A Add latency (T),
Gas leak Wait time (T)
Add,
Gas leak Wait time (T)
Add,
Gas leak Low temperature Wait time (T)
decrease Wait time (T)
Add,
Gas leak
N / A
N / A
N / A Room temperature Wait time (T)
decrease Wait time (T)
Add,
Gas leak
N / A
N / A
N / A High temperature Wait time (T)
Add,
Residual Gas Check Wait time (T)
Add,
Gas leak
N / A
N / A
N / A

Combined Disability Decision 2: S (i, j) Acceleration (stop)
(X, Y, Z axis)
0g Acceleration (change)
(X, Y axis)
+/- 2g {≠ 0g} Acceleration (change)
(Z axis)
+/- 2g {≠ 0g} Pressure (stop)
value : 0 Pressure (change)
value : 300 ~ 1100 {≠ 0} acceleration
(stop)
(X, Y, Z axis)
N / A Wait time (T)
Add Add wait time (T) Wait time (T) reduction Wait time (T)
Add,
Possible vibration Acceleration
Degree (change)
(X, Y axis) Add wait time (T)
N / A Add wait time (T) Add latency (T),
Left and right vibration possible Wait time (T)
Add,
Vibration of left and right acceleration
(change)
(Z axis) Add wait time (T) Wait time (T)
Add
N / A Add latency (T),
Vertical vibration is possible Wait time (T)
Add,
Vertical vibration pressure
(stop) Wait time (T) reduction Wait time (T)
Add,
Left and right vibration possible Add latency (T),
Vertical vibration is possible
N / A
N / A pressure
(change) Add wait time (T)
Possible vibration Add latency (T),
Vibration of left and right Add latency (T),
Vertical vibration
N / A
N / A

An example of the semiconductor equipment applied to Table 1 and Table 2 is as follows. Gases for semiconductors include CF4, C2F6, SF6, NF3, Freon, etc. as CFC and PFC-based gases, and SiH4, PH3, AsH3, B2H6, etc. as SiH4. Fluorine and chlorine series include F2, Cl2, HF, HCl, BF3, HBr, WF6, etc. have.

Chorus1110H-48, GS Tech's burn-in tester equipment, which is an example, has a test temperature range of -40 ° C to + 125 ° C for semiconductor parts and tests the semiconductor parts by changing the temperature within the range.

Table 3 shows the sensing range for each element by the sensors used.

division N2 gas Temperature pressure acceleration range Human body concentration (TLV) 200ppm or less -40-125 300-1100 mbar  16bit: +/- 2g

At this time, when the test environment is changed from low temperature to room temperature or high temperature using the Chorus1110H-48 equipment, condensation due to moisture is generated inside the tester equipment. Typically, condensation decreases with time at low temperatures at room temperature.

However, for the frequent test process, the experiment is performed by adjusting the temperature by force, but the test is performed while removing condensation by continuously spraying N2 gas to prevent internal condensation. If the test is performed in the presence of internal condensation, the pins of the semiconductor component may be exposed to moisture and corrode, adversely affecting the internal circuitry of the component, which causes an increase in defective products. When the burn-in tester is opened after the test and the part is taken out, traces of N2 gas remain inside the tester. Since the above-mentioned gases are mostly toxic gases, residual gas leakage can pose a serious danger to the operator.

According to this problem, the semiconductor device 10 should be monitored in real time, and should also be able to detect complex failures due to gas and temperature in real time. Referring to the examples of Tables 1 and 2, when the gas and temperature sensing values are in a steady state (within the initial threshold range), the waiting time T of the monitoring is reduced. This is because as described above, if all the sensors remain in a normal state, the probability of occurrence of a complex failure is reduced.

However, if any one of the sensors exceeds the threshold range, the likelihood of a subsequent compound failure is increased, so that the monitoring wait time T of the other sensor Sj is increased for this purpose. If the two sensors Si (Sj) exceed their respective threshold ranges within the waiting time T, the composite fault is considered to have occurred, and an alarm on S (i, j) is immediately notified to the operator. However, if the other sensor Sj does not exceed the threshold in overlap with a single sensor S (i) within the waiting time T, the waiting time T for monitoring is reduced.

Since the above process is repeatedly performed, when the waiting time T is reduced, power of the sensor can be saved. This has the effect of extending the life of the entire sensor node 100 soon, it can be broken down the notification according to the stage of danger can reduce unnecessary alarms and warnings to the operator.

In the conventional case, since there is no separate waiting time condition, a warning is continuously informed whenever an error occurs in the sensor value. However, in the case of the present invention, the sensing value of the other sensor (Sj) is monitored during the waiting time (T) which is set fluidly according to the steady state frequency of the sensors (110, 120, 130, 140) to perform the situation and generate an alarm The unnecessary alarm action can be reduced.

FIG. 5 is a graph comparing the frequency of alarm occurrences for the case where the method of FIG. 3 is applied and the case where the method of FIG. 3 is not applied. In fact, as a result of the experiment, the case of applying the situation according to the method of the present invention was able to reduce the alarm occurrence frequency by about 85% compared to the case of non-applied. That is, in the method of the present invention, since the other sensor is monitored only during the waiting time of the entire monitoring time, the alarm occurrence frequency can be reduced by that amount. This increases the power life of the entire sensor node 100 as a result, there is an effect that does not cause unnecessary warning to the operator.

6 and 7 are exemplary views of a monitoring screen for the method of FIG. This is implemented on the display unit 270 in the analysis server 200, and corresponds to the monitoring software for monitoring the state of the semiconductor equipment.

In FIG. 6, a list of sensor nodes and a test result for each sensor node are simply shown. FIG. 7 is a detailed analysis content screen of a sensor node, which displays sensing values for each sensing element (temperature, gas, pressure, acceleration (vibration)) in numerical values and graphs, and indicates a state of failure detection and the like over time. It also includes a function to accumulate text.

That is, each sensor node 100 can be displayed on the screen as necessary, and the state of the sensing value received from the displayed sensor node 100 is analyzed through graphs and numerical values. Here, when the specific single sensor (Si) exceeds the threshold value, whether the other sensor (Sj) is exceeded continuously during the waiting time (T), but if the other sensor (Sj) also exceeds the threshold value is a complex failure Recognize. The operator is informed of an alarm of S (i, j). However, if the threshold value of the other sensor Sj does not subsequently occur within the waiting time T, only the warning of the single sensor Si is accumulated, and the waiting time is reduced.

Although the present invention has been described with reference to the embodiments shown in the drawings, these are merely exemplary and those skilled in the art will understand that various modifications and equivalent other embodiments are possible. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

100: sensor node 110: temperature sensor
120: vibration sensor 130: pressure sensor
140: gas sensor
200: complex situational awareness system for monitoring the status of semiconductor equipment
210: receiving unit 220: first monitor unit
230: first alarm unit 240: second monitor unit
250: second alarm unit 260: storage unit
270: display unit 280: input unit

Claims (10)

Receiving respective sensing values from a sensor node having a plurality of sensors for detecting obstacles of the semiconductor equipment;
Monitoring whether the sensing values of the sensors deviate from a corresponding threshold range;
Analyzing a failure of the specific single sensor when the sensing value of the specific single sensor is out of a corresponding threshold range, and generating an alarm corresponding thereto;
Monitoring a sensing value of another one of the sensors for a set waiting time when a failure of the specific single sensor is detected; And
If the sensing value of the other single sensor is out of the corresponding threshold range, analyzing a complex failure caused by the specific single sensor and the other single sensor and generating an alarm corresponding thereto;
Monitoring the sensing value of the other single sensor,
And as the number of sensors whose sensing value is within a normal range increases, the waiting time is reduced. The method according to claim 1,
The plurality of sensors,
Complex situation recognition method for the state monitoring of semiconductor equipment including a plurality of sensors selected from the temperature sensor, vibration sensor, pressure sensor, gas sensor. The method according to claim 1 or 2,
And storing a sensing value of the specific single sensor or the other single sensor that is outside of the threshold range. delete The method according to claim 1 or 2,
And displaying a result of comparing each sensing value sensed by the sensor with a corresponding threshold range in the form of at least one of an image, a graph, and text. A receiver configured to receive respective sensing values from a plurality of sensors that detect obstacles of the semiconductor device in real time;
A first monitor unit configured to monitor whether the sensing values of the sensors deviate from a corresponding threshold range;
A first alarm unit configured to analyze a failure of the specific single sensor and generate an alarm corresponding thereto when a sensing value of a specific single sensor among the sensors is out of a corresponding threshold range as a result of the monitoring;
A second monitor unit configured to monitor sensing values of other single sensors among the sensors for a set waiting time when a failure of the specific single sensor is detected; And
If the sensing value of the other single sensor is out of the corresponding threshold range, and includes a second alarm unit for analyzing the combined failure by the specific single sensor and the other single sensor and generates an alarm corresponding thereto,
The second monitor unit,
The complex situation recognition system for monitoring the state of the semiconductor device to adjust so that the waiting time is reduced, as the number of sensors that the sensing value is within the normal range increases. The method of claim 6,
The plurality of sensors,
Complex situation recognition system for monitoring the status of semiconductor equipment including a plurality of sensors, such as temperature sensor, vibration sensor, pressure sensor, gas sensor. The method according to claim 6 or 7,
And a storage unit configured to store sensing values of the specific single sensor or the other single sensor outside the threshold range. delete The method according to claim 6 or 7,
And a display unit which displays a result of comparing each sensing value detected by the sensor with a corresponding threshold range in at least one of an image, a graph, and text.

Images