JP7094177B2 - Life estimation device for power semiconductor devices - Google Patents

Description

The present application relates to a life estimation device for a power semiconductor element that estimates the life of a power semiconductor element such as a thyristor, and is particularly provided for a thyristor exciter that supplies a DC current to a field winding of a generator. It relates to a life estimation device for a power semiconductor device that estimates the life of a thyristor.

The thyristor excitation method, which is one of the methods for exciting the field winding of a generator, uses a thyristor to supply a desired direct current to the field winding, and has a simple structure and quick response. There is an advantage that there is. On the other hand, in the thyristor, like other power semiconductor devices, the wire bonding portion on the chip surface or the solder bonding portion inside the element deteriorates due to continuous use, and there is a possibility that an abnormality or failure due to deterioration may occur. In order to prevent such abnormalities, it is necessary to monitor the condition of the thyristor during operation by regular inspections and checks. However, since the life of power semiconductor devices depends on the ambient temperature and operating conditions, there is a possibility that rapid deterioration cannot be dealt with only by regular inspections and checks.

Therefore, the temperature of the power semiconductor element is estimated, the number of heat stresses based on the amplitude of the estimated temperature change and the number of heat stresses based on the estimated temperature change rate are calculated and integrated, and these heat stress times and the power semiconductor element are calculated and integrated. It has been proposed to estimate the operable life of a power semiconductor element using an inherent permissible number of heat stresses (see, for example, Patent Document 1). Such life estimation is based on the power cycle model, and it is possible to estimate the life of a failure according to a thermal fatigue accumulation model such as a wear failure.

Japanese Unexamined Patent Publication No. 8-126337

However, deterioration of power semiconductor devices and associated failures and abnormalities do not always follow the thermal fatigue accumulation model. For example, mixing of impurities in the manufacturing process or ensuring insufficient insulation withstand voltage causes an accidental increase in leakage current over time, but such a model of accidental failure is a failure model different from the model of accumulation of thermal fatigue. , It is difficult to estimate the life with the technique of Patent Document 1. Therefore, it is necessary to monitor the deterioration status of the power semiconductor element by periodic inspection / check, and there is a problem that the monitoring cost increases.

The present application discloses a technique for solving the above-mentioned problems, and the technique disclosed in the present application is a power semiconductor device capable of estimating the life of a power semiconductor device regardless of a failure model. This is to obtain the life estimation device of.

The life estimation device for a power semiconductor element disclosed in the present application includes a temperature measuring means for continuously measuring the temperature of the power semiconductor element during operation and a temperature of the power semiconductor element measured by the temperature measuring means. An average processing calculation unit that sequentially calculates the section average value of each of a plurality of sections having a predetermined length from the series data and the change amount of the section average value, and the change amount of the section average value and the section average value. It is equipped with a control temperature arrival time estimation unit that estimates the control temperature arrival time, which is the time when the temperature of the power semiconductor element reaches a predetermined control temperature, and the control temperature arrival time estimation unit is an interval average . When calculating the amount of change in the average value in the current section when calculating the amount of change in the value sequentially , the current section is used as the reference section, and one of the sections before the reference section is the section to be compared. If the difference between the section average value of the reference section and the section average value of the section to be compared is greater than or equal to a predetermined value, the difference is taken as the amount of change in the section average value in the current section , and the difference is If it is smaller than the predetermined value, the comparison target section is updated by setting the comparison target section to the previous section, and the difference is obtained and the difference is compared with the predetermined value again. The control temperature arrival time estimation unit repeats the update of the section to be compared until the difference becomes equal to or more than a predetermined value .

According to the technique disclosed in the present application, the life of a power semiconductor device can be estimated regardless of the failure model.

It is a block diagram which shows the thyristor excitator which concerns on Embodiment 1. FIG. It is a block diagram which shows the life estimation apparatus of the power semiconductor element in Embodiment 1. FIG. It is a figure which shows the hardware composition of the life estimation apparatus of the power semiconductor element in Embodiment 1. FIG. It is a figure which shows the example of the format of the temperature value table which concerns on Embodiment 1. FIG. It is a flow diagram which shows the operation of the life estimation apparatus of the power semiconductor element in Embodiment 1. FIG. It is a flow chart which shows the section average value comparison processing which concerns on Embodiment 1. It is a characteristic diagram which shows the estimated straight line of the control temperature arrival time.

Embodiment 1.
Hereinafter, the first embodiment will be described with reference to FIGS. 1 to 7. FIG. 1 is a block diagram showing a thyristor exciter according to the first embodiment, and FIG. 2 is a block diagram showing a life estimation device for a power semiconductor element according to the first embodiment. The thyristor exciter 50 supplies an exciting current to the field winding 92 of the generator 91, and has a plurality of thyristors, that is, a thyristor unit 51 having a semiconductor element for power, and an AVR (Automatic Voltage) that controls each thyristor. It is equipped with a Regulator) 52. When the generator 91 is driven by a turbine (not shown) and an exciting current is supplied to the field winding 92 from the thyristor section 51, the output voltage of the generator 91 is stepped down by the exciting transformer 93 and the cyclist section 51. Is applied to. Further, the output voltage of the generator 91 and the current from the generator 91 are input to the AVR 52 as feedback via the transformer 94 and the current transformer 95, respectively. The AVR 52 internally includes a gate control signal generation unit 521, and outputs a gate control signal g that controls the thyristor of the thyristor unit 51 according to the feedback current IR.

The life estimation device 10 of the power semiconductor element estimates the life of the thyristors 511 to 513 of the thyristor unit 51 shown in FIG. 2, and continuously measures and measures the temperatures of the thyristors 511 to 513 during operation. Calculation by the temperature sensor 11 that outputs the temperature value as the temperature value data D, that is, the temperature measuring means, the life estimation unit 12 that estimates the life of each of the thyristors 511 to 513 using the temperature value data D, and the life estimation unit 12. Based on the results, the treatment determination unit 13 that determines the necessity of treatment for each of the thyristors 511 to 513, and the treatment execution unit 14 that executes the treatment for each of the thyristors 511 to 513 based on the determination result of the treatment determination unit 13. I have.

The life estimation unit 12 acquires the temperature value data D from the temperature value data sampling unit 121 that samples the temperature value data D output by the temperature sensor 11 at a constant sampling cycle and the temperature value data sampling unit 121, and has a tabular format. A section is configured by acquiring temperature value data D from the temperature value table storage unit 122 and the temperature value data sampling unit 121, which are stored as a temperature value table, and dividing the time-series temperature value data D into predetermined numbers. Then, the average processing calculation unit 123 that calculates the section average value of each section and the amount of change in the section average value, and the control temperature arrival time estimation unit 124 that calculates the time for the temperatures of the cyclists 511 to 513 to reach the control temperature. It is equipped with. The interval average value calculated by the average processing calculation unit 123 and the amount of change thereof are sent to the temperature value table storage unit 122 and stored in the temperature value table. The "controlled temperature" is the temperature at which an offline response to the target thyristor, such as inspection or replacement, is required. Specific numerical values are preset according to the specifications of the target thyristor or the usage environment. Since the temperature value data sampling unit 121 performs sampling in real time from the temperature sensor 11 that continuously measures and outputs, the difference between the measurement time of the temperature sensor 11 and the sampling time of the temperature value data sampling unit 121. Is small enough.

The treatment determination unit 13 determines the content of the "treatment" for each of the thyristors 511 to 513 based on the calculation result of the control temperature arrival time estimation unit 124, and sends the treatment signal s according to the determination result to the treatment execution unit 14. Output. The treatment execution unit 14 treats the gate control signal g according to the value of the treatment signal s to obtain the post-treatment gate control signal x, and transmits the post-treatment gate control signal x to the respective thyristors 511 to 513. Here, the "treatment" is, for example, "mask processing" or "waiting". When the "treatment" is "mask processing", the value of the treatment signal s becomes "0" so that the target thyristor is turned off, and the value of the post-treatment gate control signal x is set regardless of the value of the gate control signal g. It becomes "0". This indicates that the target thyristor is forcibly stopped. When the "treatment" is "waiting", the value of the treatment signal s becomes "1", and the post-treatment gate control signal x becomes equal to the gate control signal g. In this case, the target thyristor continues to operate according to the gate control signal g.

Since the temperature measurement, control, and life estimation of the cyclists 511 to 513 are performed independently for each thylister, the temperature value data D, the gate control signal g, the treatment signal s, and the post-treatment gate control signal x shown in the figure are used. , The temperature value data D1 to D3 corresponding to each of the thyristors 511 to 513, the gate control signals g1 to g3, the treatment signals s1 to s3, and the post-treatment gate control signals x1 to x3 are included. The same applies to the following description.

Next, the hardware configuration that realizes each functional unit shown in FIG. 2 will be described. FIG. 3 is a diagram showing a hardware configuration of a life estimation device for a power semiconductor element according to the first embodiment. The life estimation device 10 of the power semiconductor element includes a processor 81 that executes various calculations, a storage device 82 that stores a program for operating the processor 81, the results of various calculations, and the like, and temperature value data D from the temperature sensor 11. It is composed of an input circuit 83 that receives an input signal such as a gate control signal g from the AVR 52 and transmits it to the processor 81 and the storage device 82, and an output circuit 84 that outputs the post-treatment gate control signal x to the thyristor unit 51.

Next, the temperature value table will be described. FIG. 4 is a diagram showing an example of the format of the temperature value table according to the first embodiment. As shown in FIG. 4, the temperature value table includes "data ID", "temperature value", "section ID", "section average value", and "difference from the section average value of the previous section". Hereinafter, it will be described in detail.

The "data ID" indicates the time series number of the temperature value data D sampled by the temperature value data sampling unit 121, and corresponds to the sampling time. The "temperature value" is a temperature value indicated by the temperature value data D. The "interval ID" identifies the interval defined by the average processing calculation unit 123, and indicates the number of the interval in the time series. In the first embodiment, the number of temperature value data D constituting one section is set to h. The specific numerical value of h may be determined from the time constant of the temperature sensor 11, the sampling period for acquiring the temperature value data D, and the like. Further, in determining h, detailed information provided by the vendor, such as specifications of thyristors 511 to 513, may be taken into consideration.

In FIG. 4, the temperature value is expressed in the form of D (k, j) so that the temperature value data D is the temperature value data in which section and which number. When the data ID of the temperature value data D is m, the section ID is k, and the time series number in the kth section is j, the following equation (1) holds.

なお、図４においては式（２）を簡略化して記載している。
また、実施の形態１では１区間中のすべての温度値から算出した単純平均を区間平均値をとしているが、それぞれの温度値が重みづけされた加重平均であってもよい。また、区間中の最大値および最小値を除いて区間平均値を算出してもよい The "interval average value" is calculated by the average processing calculation unit 123 as described above. The interval average value avr (k) of the kth interval is expressed by the following equation (2).

In FIG. 4, the equation (2) is simplified and described.
Further, in the first embodiment, the simple average calculated from all the temperature values in one section is used as the section average value, but each temperature value may be a weighted average. Further, the interval average value may be calculated by excluding the maximum value and the minimum value in the interval.

式（４）においてｉ＝ｋ－１の場合、式（５）で示すように第１区間の区間平均値との差となる。

なお、上述したように、サイリスタ５１１～５１３の温度測定はそれぞれについて独立して行われるので、温度値テーブルもそれぞれのサイリスタについて独立して作成される。このため、温度値テーブル記憶部１２２は３つの温度値テーブルを記憶することとなる。 As shown in the figure, the "difference from the section mean value of the previous section" is everything before the reference section, from the difference from the section mean value one section before to the difference from the section mean value of the first section. Includes the difference from the interval mean value for the interval of. That is, when the reference section is the k-th section, the difference from the section mean value of one section before is Δ (k), and the difference from the section mean value of any section (before i section) is Δ (k, If i), the following equations (3) and (4) are obtained.

When i = k-1 in the equation (4), it is the difference from the interval average value of the first interval as shown in the equation (5).

As described above, since the temperature measurement of the thyristors 511 to 513 is performed independently for each, the temperature value table is also created independently for each thyristor. Therefore, the temperature value table storage unit 122 stores three temperature value tables.

Next, the operation will be described. FIG. 5 is a flow chart showing the operation of the life estimation device for the power semiconductor element according to the first embodiment. Here, the m-th sampling will be described as an example. The current section is assumed to be the kth section.
First, the temperature value data D is sampled (step ST001). The information (data ID, temperature value) of the temperature value data D acquired by the sampling is sequentially written in the temperature value table.

Next, the difference between the temperature value D (m) of the temperature value data acquired in this sampling and the temperature value D (m-1) of the temperature value data acquired in the previous sampling (D (m) -D (m-). 1)) is calculated to obtain the temperature rise value, and it is determined whether or not this temperature rise value is equal to or less than the allowable temperature rise value ΔTm (step ST002). If the temperature rise value exceeds the allowable temperature rise value, the process proceeds to step ST003. In this case, since it is recognized that a sudden temperature rise has occurred, it is determined that an abnormality including a failure of the temperature sensor 11 has been detected, and processing such as stopping the thyristor exciter 50 and the generator 1 is performed (step ST003). ). Specifically, the treatment determination unit 13 transmits a treatment signal s of "0" (gate forced stop) to the treatment execution unit 14, and the treatment execution unit 14 performs mask processing on the gate control signal g. The thyristor exciter 50 and the generator 1 are stopped.
The allowable temperature rise value ΔTm, which is a threshold value, is preset in the design of the thyristor exciter 50.

When the temperature rise value is not more than the allowable temperature rise value ΔTm, it is determined whether sampling is performed for one section from the initial state or the previous estimation of the control temperature arrival time (step ST004). If sampling for one section has not been performed, the process returns to step ST001 and the (m + 1) th sampling is performed. After that, steps ST001 to ST004 are repeated until sampling for one section is performed.

When sampling for one section is performed, the section average value avr (k) is calculated and written in the temperature value table (step ST005). At this time, the difference Δ (k, i) from the section average value of the previous section is also calculated and written in the temperature value table.

Next, the interval mean value comparison process is performed (step ST006). FIG. 6 is a flow chart showing an interval mean value comparison process. In the section mean value comparison process, in order to grasp the degree of change in the section mean value, the current (kth section) section mean value avr (k) and the previous section section mean value are calculated until a significant amount of change is obtained. We will compare them one by one. First, the arithmetic mean value avr (ki) to be compared is set to the interval average value avr (k-1) one section before. That is, i = 1 is set (step ST101).

Next, the change amount ΔTa of the interval average value is obtained, and it is determined whether the change amount ΔTa is equal to or greater than the significant temperature increase value ΔTeff (step ST002). Since the amount of change ΔTa of the interval mean value is equal to Δ (k, i) of the temperature value table, the corresponding Δ (k, i) may be read from the temperature value table. When the amount of change ΔTa of the interval average value is equal to or greater than the significant temperature increase value ΔTeff, it is determined that a significant temperature increase is observed between the current interval average value and the interval average value of the (ki) section. The change amount ΔTa of the interval average value is saved, and the interval average value comparison process is terminated (step ST103).

If the amount of change ΔTa of the interval average value is smaller than the significant temperature increase value ΔTeff, it is assumed that no significant temperature increase is observed, and i is incremented by 1 to set the arithmetic mean value to the interval average value of the previous interval (). Step ST104). After that, the process returns to step ST102 and the change amount ΔTa of the interval average value and the significant temperature increase value ΔTeff are compared. Step ST102 and step ST104 are repeated until the change amount ΔTa of the interval average value becomes equal to or greater than the significant temperature increase value ΔTeff. The value of the significant temperature rise value ΔTeff is predetermined based on the temperature characteristics of the target thyristor, the value of the control temperature, and the like. Further, in the first embodiment, the change amount ΔTa of the section average value and the significant temperature increase value ΔTeff are compared, but the change rate obtained by dividing the change amount ΔTa of the section average value by the section average value avr (k) is obtained. It may be determined whether or not a significant temperature increase is observed based on this rate of change.

ここで、式（６）におけるａ、ｂは以下の式（７）のとおりである。

ａ、ｂを算出することで図７に示す推定直線Ｌ１を求めることができる。
なお、ΔＴａ＝Δ（ｋ、ｉ）と一般化した場合のａ、ｂは以下の式（８）のとおりである。

なお、式（６）により得られる管理温度到達区間ｋｅは、以下の式（９）により管理温度到達時間ｔｅに変換可能である。

ここで式（９）におけるＴｓは温度値データサンプリング部１２１のサンプリング周期である。ｈは、上述したとおり１つの区間を構成する温度値データＤの個数である。 After the section average value comparison process is completed, the control temperature arrival time te is estimated using the current section mean value avr (k) and the change amount ΔTa of the section mean value (step ST007). Here, it is assumed that ΔTa = Δ (k, 1). In the first embodiment, as shown in FIG. 7, the estimated straight line L1 is obtained by approximating the temperature T with a linear function of T = aX + b, and the section ID at which the temperature T reaches the control temperature Tc (T = Tc). The control temperature reaching interval ke is calculated by the following equation (6). Here, X is a continuation of the values of the section ID.

Here, a and b in the formula (6) are as shown in the following formula (7).

By calculating a and b, the estimated straight line L1 shown in FIG. 7 can be obtained.
In addition, a and b in the case of generalizing ΔTa = Δ (k, i) are as shown in the following equation (8).

The controlled temperature arrival section ke obtained by the equation (6) can be converted into the controlled temperature arrival time te by the following equation (9).

Here, Ts in the equation (9) is the sampling period of the temperature value data sampling unit 121. As described above, h is the number of temperature value data D constituting one section.

Next, it is determined whether or not the estimation result of the control temperature arrival time te has changed (step ST008). If there is no change in the control temperature arrival time te, the process ends. If there is a change, the treatment is changed according to the change (step ST009). As can be seen from the equation (9), the control temperature arrival time te is uniquely determined from the value of the control temperature arrival section ke, so whether or not the control temperature arrival time te has changed depends on whether the control temperature arrival section ke has changed. You can see if it is. Hereinafter, the section average value avr (k + 1) in the section ID (k + 1) will be described separately for each case.

(1) When avr (k + 1) = avr (k + 1) _1 As shown in FIG. 7, the point (k + 1, avr (k + 1) _1) is located on the estimated straight line L1. Therefore, when avr (k + 1) = avr (k + 1) _1, the estimated straight line in the section ID (k + 1) coincides with the estimated straight line L1, and the controlled temperature reaching section ke1 becomes equal to the controlled temperature reaching section ke1. Therefore, there is no change in the estimated value of the control temperature arrival time and no treatment change is made.

(2) When avr (k + 1) = avr (k + 1) _2 As shown in FIG. 7, the point (k + 1, avr (k + 1) _2) is located above the estimated straight line L1. Therefore, when avr (k + 1) = avr (k + 1) _2, the estimated straight line L2 in the section ID (k + 1) has a larger slope than the estimated straight line L1, and the controlled temperature reaching section ke2 is smaller than the controlled temperature reaching section ke. Become. Since this indicates that the estimated life of the target thyristor has been shortened, treatment changes such as accelerating the time for performing regular maintenance and the time for replacing the target thyristor are implemented.

(3) In the case of avr (k + 1) = avr (k + 1) _3 As shown in FIG. 7, the points (k + 1, avr (k + 1) _3) are located below the estimated straight line L1. Therefore, when avr (k + 1) = avr (k + 1) _3, the slope of the estimated straight line L3 in the section ID (k + 1) is smaller than that of the estimated straight line L1, and the controlled temperature reaching section ke3 is larger than the controlled temperature reaching section ke. Become. Since this indicates that the estimated life of the target thyristor has been extended, treatment changes such as delaying the time for performing regular maintenance and the time for replacing the target thyristor are implemented.

When the time reaches the control temperature arrival time te, measures are taken according to the temperature of the target thyristor. If the temperature is below the control temperature Tc, the target thyristor and the entire device are continued to operate. When the temperature is equal to or higher than the control temperature Tc, the target thyristor is stopped, and the entire system including the thyristor exciter 50 or the generator 91 is stopped depending on the configuration of the device. If the system is redundant, start the standby system and continue operation.

The life estimation device 10 of the power semiconductor element repeatedly and continuously executes the operation described above with respect to the thyristors 511 to 513 of the thyristor exciter device 50 in operation. That is, every time the temperature value data D for one section is acquired from the temperature sensor 11, the average processing calculation unit 123 sequentially calculates the section average value and the amount of change thereof. The control temperature arrival time estimation unit 124 sequentially estimates the control temperature arrival time using the newly calculated interval average value and the amount of change thereof, and updates the estimated value of the control temperature arrival time.

According to the first embodiment, the life of the power semiconductor element can be estimated regardless of the failure model. More specifically, with an average processing calculation unit that sequentially calculates the section mean value and the amount of change in the section mean value in a section with a predetermined length from the time series data of the thyristor temperature measured by the temperature sensor. , It is equipped with a control temperature arrival time estimation unit that estimates the control temperature arrival time from the section average value and the amount of change thereof. Failures and abnormalities due to deterioration of semiconductor elements are accompanied by temperature rise, so if the control temperature arrival time is estimated using temperature information such as the interval average value calculated sequentially and the amount of change thereof, a detailed failure model will not be clear. However, the control temperature arrival time can be estimated, and the failure model that can be monitored is not limited. In addition, it is possible to perform estimation with a certain degree of accuracy by sequentially performing estimation, and to make appropriate treatment changes according to changes in the estimation result. Therefore, by optimizing the maintenance implementation time and the semiconductor element replacement time, it is possible to suppress an increase in monitoring cost and improve the operation rate of the entire device.

In addition, since it is equipped with a treatment execution unit that executes mask processing on the gate control signal that controls the target thyristor according to the temperature value of the thyristor measured by the temperature sensor, the processing according to the temperature of the thyristor is performed in real time. This can be done, and accidental abnormalities and failures can be quickly detected and troubles can be prevented.

In addition, if the difference between the previous measurement value and the current measurement value by the temperature sensor is an abnormal temperature rise value, the gate control signal is masked and forcibly stopped, resulting in a sudden abnormality. You can respond to the situation.

In addition, since the amount of change in the interval average value used to estimate the control temperature arrival time is greater than or equal to the significant temperature rise value, it prevents estimation based on temporary temperature fluctuations or insignificant minute changes, and is more accurate. Can be highly estimated.

In the first embodiment, after the temperature value table shown in FIG. 4 is created, the section average values are compared using Δ (k, i) written in the temperature value table in the section average value comparison process. However, the difference between the current section average value and the comparison average value may be calculated as necessary in the section average value comparison process. In this case, it is not necessary to calculate and store Δ (k, i) after the difference between the current interval average value and the arithmetic mean value becomes the significant temperature increase value ΔTeff or more. For example, if Δ (k, 1) is equal to or greater than the significant temperature increase value ΔTeff, the calculation of Δ (k, 2), Δ (k, 3) and the like can be omitted. Further, since Δ (k, i), which is the difference from the section average value of the previous section, can be omitted from the temperature value table, the data size of the temperature value table can be reduced.

Further, although the thyristor has been described above as an example, the technique of the first embodiment can be applied to other power semiconductor devices such as MOSFET (Metal Oxide Semiconductor Field Effect Transistor) and IGBT (Insulated Gate Bipolar Transistor). be.

Although the present application describes exemplary embodiments, the various features, embodiments, and functions described in the embodiments are not limited to the application of a particular embodiment, either alone or. Various combinations are applicable to the embodiments.
Therefore, innumerable variations not illustrated are envisioned within the scope of the art disclosed in the present application. For example, it is assumed that at least one component is modified, added or omitted.

10 Life estimation device for power semiconductor elements, 11 Temperature sensor, 12 Life estimation unit, 122 Temperature value table storage unit, 123 Average processing calculation unit, 124 Control temperature arrival time estimation unit, 13 Treatment judgment unit, 14 Treatment execution unit, 50 thyristor exciter, 51 thyristor section, 511, 512, 513 thyristor, D temperature value data, g gate control signal, s treatment signal, x post-treatment gate control signal, L1, L2, L3 estimated linear

Claims (4)

A temperature measuring means that continuously measures the temperature of power semiconductor devices during operation,
From the time-series data of the temperature of the power semiconductor element measured by the temperature measuring means, the section average value of each of a plurality of sections having a predetermined length and the amount of change in the section average value are sequentially calculated. The average processing calculation unit to calculate and
A control temperature arrival time estimation unit that estimates the control temperature arrival time, which is the time when the temperature of the power semiconductor element reaches a predetermined control temperature, using the section average value and the amount of change in the section average value. Equipped with
When calculating the amount of change in the section mean value in the current section when the control temperature arrival time estimation unit sequentially calculates the amount of change in the section mean value , the control temperature arrival time estimation unit uses the current section as a reference section and the reference. When one of the sections before the section to be compared is set as the section to be compared, and the difference between the section mean value of the reference section and the section mean value of the comparison target section is equal to or more than a predetermined value. The difference is taken as the amount of change in the section mean value in the current section .
When the difference is smaller than the predetermined value, the section to be compared is updated by setting the section to be compared to the previous section to acquire the difference and to obtain the difference and the predetermined difference. The comparison with the given value is performed again.
The control temperature arrival time estimation unit is a life estimation device for a power semiconductor element, which repeats updating of a section to be compared until the difference becomes equal to or more than a predetermined value . The first aspect of the present invention further comprises a treatment execution unit that performs a masking process for turning off a gate control signal for controlling the power semiconductor element according to the temperature value of the power semiconductor element measured by the temperature measuring means. The life estimation device for the power semiconductor element described. The power semiconductor according to claim 2 , wherein the treatment executing unit performs the mask processing when the difference between the previous measured value and the current measured value by the temperature measuring means is equal to or larger than a predetermined threshold value. Device life estimation device. Further provided with a storage unit for storing the difference between the interval average values of the two sections different from each other in correspondence with the combination of the IDs of the two sections.
The control temperature arrival time estimation unit reads the difference between the section average values corresponding to the combination of the ID of the reference section and the ID of the section to be compared from the storage unit, thereby performing the section average of the reference section. The life estimation device for a power semiconductor element according to any one of claims 1 to 3 , wherein the difference between the value and the interval average value of the section to be compared is acquired.

Images