JP2010124606A - Simulation method for estimating loss/temperature of semiconductor device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To calculate the maximum value Tjmax of junction temperature of a semiconductor device in a short time when the inverter output frequency changes in the one cycle of operation pattern of a variable speed PWM inverter apparatus from start to stop. <P>SOLUTION: Configuration or operation pattern of an inverter is input (201). According to the operational conditions of an inverter thus input, average generation loss of a semiconductor element is calculated in each section (202). T(c-f) and T(f-a) are calculated from the waveform of average generation loss thus calculated (204). Moveover, the maximum value T(j-c)max of temperature between junction and case is calculate while taking account of the output frequency (205-207). From the value of the Tc and the value of the T(j-c)max, time-series data of the maximum value Tjmax of junction temperature is obtained (208). A transient response correction is made with a lapse of time t immediately after switching the section, and the ultimate maximum value Tjmax of junction temperature is obtained by estimation (209). <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a simulation method related to the design of a variable speed PWM inverter device, and more particularly to a simulation method for estimating a generated loss and temperature of a semiconductor element.

  FIG. 7 is a diagram showing a typical circuit configuration of an inverter device, and the illustrated example relates to a three-phase bridge inverter device that outputs a three-phase alternating current. A plurality of semiconductor elements 101 are installed on the heat sink 103. The semiconductor element 101 is configured such that two elements are connected in series between DC buses, and each phase of the three-phase alternating currents u, v, and w is output from the connection point of the semiconductor elements 101. A DC link (coupling) voltage 102 (not shown) is applied to a smoothing electrolytic capacitor disposed between the DC buses.

  FIG. 8 is a diagram illustrating an operation example of a general inverter device. Section 1 is an acceleration period, and the output frequency f changes from 0 Hz to the rated frequency at a load current of 150%, for example. Section 2 is a steady operation period and operates at the rated frequency with 100% load. Section 3 is a deceleration section, and the output frequency f changes from the rated frequency to 0 Hz with a load current of 130%.

  In designing the inverter device, it is necessary to calculate the junction temperature Tj of the semiconductor element in such an operation pattern so that the junction temperature Tj falls within the maximum rated value of the semiconductor element.

  FIG. 9 is a diagram showing waveforms of output current, generated loss, and junction temperature Tj during the acceleration operation shown in FIG. When the effective value I of the output current is constant, the average value P of the generated loss is constant. However, the junction temperature Tj varies depending on the output frequency f, and increases as the frequency decreases. At this time, the maximum value T (j−c) max of the difference between the junction temperature Tj and the case temperature Tc is in an inversely proportional relationship as shown in FIG.

  As a conventional loss / temperature calculation method, the calculation method shown in Patent Document 1 is known. However, since the temperature ripple due to the output frequency of the inverter device is not taken into consideration, a large error occurs in the estimated value of the element temperature rise.

  Further, in Patent Document 2, the inverter current and output current frequency based on the frequency command and the torque command, and the current output time are used as variables to refer to the function stored in the temperature abnormality determination function storage unit. There is disclosed an elevator control microcomputer that includes a temperature abnormality determination unit that determines whether the temperature of a switching element constituting a main circuit is normal or abnormal and has a function of protecting the switching element from thermal destruction. The temperature abnormality determination function storage unit includes a case-junction junction temperature conversion calculation unit, and is input by the output current frequency data, the element generation loss calculated by the switching element generation loss calculation unit, and the applied element data input unit. Calculate the junction temperature change based on the applicable element characteristic data, and memorize a functional equation with the characteristic that the temperature rise peak value between the case of the switching element and the chip junction increases as the output frequency decreases. Is disclosed.

  In addition, as a method of calculating the temperature from the generated loss, a method of calculating the temperature rise model between the junction and case, between the case and the heat sink, and between the heat sink and the surrounding as shown in Fig. 3 as an equivalent first-order lag circuit. is there. In this method, the temperature change ΔT after t seconds when the generated loss is P can be obtained by the following equation (1), where Rth is the steady thermal resistance value and Tth is the first-order lag time constant.

ΔT = P × Rth {1-exp (t / Tth)} (1)
JP 2007-043783 A JP 2002-302359 A (paragraphs 0016, 0023)

  A method of calculating the temperature rise model as an equivalent first-order lag circuit is limited to a case where the generated loss P is constant. When P changes continuously, it is necessary to divide t into very short periods where P can be considered constant and repeatedly calculate ΔT. Therefore, when calculating the operation pattern from when the inverter device starts operation until acceleration / deceleration is repeated until it stops, the calculation amount becomes enormous if the temperature ripple due to the inverter output frequency is calculated accurately. There is a drawback.

  There is also a method of calculating the junction temperature Tj of the semiconductor element by a periodic function without using the above formula (1). For example, Patent Document 1 describes a method of decomposing a loss generated in an operation pattern into a frequency domain by Fourier transform, calculating a temperature rise of each frequency component, and then obtaining a temperature rise in the time domain by inverse Fourier transform. Has been. By using the periodic function, it is possible to calculate the temperature rise in the operation pattern with a small amount of calculation. However, since this method does not take the output frequency into consideration, there is a disadvantage that the element temperature rise cannot be calculated accurately.

  In the present invention, in order to solve the above-described problem, the maximum junction temperature of the semiconductor element when the inverter output frequency changes in the one-cycle operation pattern from the start to the stop of the operation of the variable speed PWM inverter device. An object of the present invention is to provide a simulation method for accurately calculating the value Tjmax with a small calculation time.

  In the present invention, a process for selecting a semiconductor element constituting an inverter device, an acceleration, a deceleration operation, and a stop from the start of operation of the inverter device to one cycle are regarded as one cycle, and the ambient temperature Ta, heat sink in this cycle. A step of inputting time series data of transient thermal resistance value Zth (fa), output load current value I, output frequency f, carrier frequency fc, control factor λ, power factor φ, and gate resistance Rg of the semiconductor device, Loss characteristics and junction-case transient thermal resistance value Zth (jc) and database storing case-heatsink transient thermal resistance value Zth (cf), loss characteristics of the selected semiconductor element and the input value Calculating the time series data of the generated loss of the semiconductor element based on the above, the calculated generated loss, the transient thermal resistance value Zth (cf) between the case and the heat sink, and the junction-case Based on the transient thermal resistance value Zth (fa), the process of calculating the temperature rise T (cf) between the case and the heat sink, the temperature rise T (fa) between the heat sink and the ambient temperature, and the period during which the inverter output frequency continuously changes In order to obtain the maximum value T (jc) max of the junction-case temperature rise T (jc) of the semiconductor element, T (jc) max in one cycle at two output frequencies within the period is calculated. The step of calculating the output frequency and the characteristic function of T (jc) max from the value of two T (jc) max, calculating T (jc) max at all frequencies, and the calculated junction-case temperature of the semiconductor element Step of calculating the maximum value Tjmax of junction temperature from the maximum value T (jc) max of the rise T (jc), the temperature rise T (cf) between the case and the heat sink, the temperature rise T (fa) between the heat sink and the ambient temperature, and the ambient temperature Ta A simulation method for accurately estimating the temperature of a semiconductor device comprising Is the law.

  According to the present invention, the maximum value of the junction temperature of the semiconductor element in consideration of the operation pattern of the inverter device can be accurately calculated in consideration of the inverter output frequency. In addition, since the number of calculations is smaller than in the prior art, the amount of calculation of loss and temperature and the burden of inverter device design can be reduced.

The best mode for carrying out the invention will be described below with reference to the drawings.
FIG. 1 is a diagram for explaining a flow of processing for calculating a junction temperature by a simulation method according to an embodiment of the present invention. 2 and 3 are diagrams showing the configuration of the semiconductor chip, case, and heat sink of the inverter device used in the junction temperature calculation of the present invention, and the thermal resistance model thereof.

  FIG. 2 is a diagram showing an outline of the configuration of the semiconductor chip, case, and heat sink of the inverter device. One case 104 is installed on one heat sink 103, and one or more semiconductor chips are provided on one case 104. 105 is built in, and there is a surrounding 106 for these. In the thermal resistance model shown in FIG. 3, the junction temperature Tj, case temperature Tc, and heat sink temperature Tf when the ambient temperature is Ta are calculated. Hereinafter, the flow of processing for calculating the junction temperature by the simulation method shown in FIG. 1 will be described based on the thermal resistance model shown in FIG.

  201 shown in FIG. 1 is an input unit for the configuration of an inverter device and an operation operation pattern. In the inverter operation condition and operation pattern input unit 201, a semiconductor element to be applied to the inverter device is selected, the ambient temperature Ta, the output load current value I of the inverter device, the output frequency f, the DC link voltage Vdc, the PWM carrier frequency fc, PWM The control factor λ, the power factor φ, the on-gate resistance Rg (on) and the off-gate resistance Rg (off) of the semiconductor element are input as time series data.

  First, as shown in FIG. 8, the acceleration period (section 1) in which the output frequency f changes from 0 Hz to the rated frequency, the steady operation period (section 2), and the deceleration section in which the output frequency f changes from the rated frequency to 0 Hz. (Section 3) is one cycle of PWM inverter operation.

  Then, the average generated loss calculation unit 202 calculates the average generated loss in the semiconductor elements in each section based on the inverter operating conditions in each section input. The generated loss is expressed for each semiconductor element by a function of the output load current value I, the output frequency f, the PWM carrier frequency fc, the PWM control factor λ, the power factor φ, and the gate resistance Rg of the semiconductor element. Or the like and registered in the database 203. The “IGBT” shown in the database 203 stores experimental result data of an IGBT (Insulated Gate Bipolar Transistor), which is a semiconductor element often used in an inverter device as shown in FIG. 7, as an example. It has been shown.

  In the T (cf) and T (fa) calculation unit 204 shown in FIG. 1, the case-heat sink temperature T (cf), the heat sink-ambient temperature are calculated from the average generated loss waveform calculated by the average generation loss calculation unit 202. Calculate the inter-temperature T (fa). The generated loss waveform is frequency-decomposed by Fourier transform, and each frequency component is multiplied by the case-heatsink transient thermal resistance value Zth (cf) and the heatsink-ambient transient thermal resistance value Zth (fa). A value is obtained for each frequency component of the temperature T (cf) and the temperature T (fa) between the heat sink and the ambient temperature. The time domain data of the case-heat sink temperature T (c-f) and the heat sink-ambient temperature T (f-a) are obtained by inverse Fourier transforming these values. Furthermore, time series data on the case temperature Tc and the heat sink temperature Tf is obtained by the following equations (2) and (3).

Tf = T (fa) + Ta (2)
Tc = T (cf) + Tf (3)
When the inverter output frequency changes within a certain section of the inverter operation pattern, the maximum value T (jc) max of the junction-case temperature considering the output frequency is calculated in 205 to 207 shown in FIG. . When the output frequency does not change, only the calculation of the T (jc) max calculation unit 205 at the maximum frequency in the interval is performed, and the T (jc) max calculation unit 206, T (jc) at the minimum frequency in the interval The calculation by the max frequency dependence calculation unit 207 is not performed.

  In general, when the generated loss is constant, the junction temperature Tj rises with a time constant unique to the semiconductor element. Therefore, the higher the output frequency, the smaller the junction temperature rise value in one cycle, and the relationship between the maximum value T (jc) max of the junction-case temperature and the output frequency f shown in FIG. ).

T (jc) max = a / f + b (4)
By obtaining the constants a and b in the above equation (4), the maximum value T (jc) max of the junction-case temperature at all frequencies can be calculated.

  First, in the T (jc) max calculation unit 205 at the maximum frequency in the section shown in FIG. 1, the maximum value T (jc) of the junction-case temperature rise T (jc) at the maximum output frequency in the section. Calculate max. The output current waveform of one cycle is regarded as a sine wave, and the inverter output current effective value I, output frequency f, PWM carrier frequency fc, DC link voltage Vdc, gate resistance value Rg, PWM control factor λ, power factor φ, and database 203 4 is used to calculate the instantaneous value of the generated loss due to switching in one output cycle as shown in the waveform of FIG. The calculated loss waveform is Fourier-transformed and decomposed into frequency components. Each frequency component is multiplied by the junction-to-case transient thermal resistance value Zth (j-c) of the semiconductor element to obtain frequency domain data of the junction-to-case temperature rise T (j-c). Calculate the time-series data of the junction-case temperature rise T (jc) in the time domain by performing inverse Fourier transform on the calculated junction-case temperature rise frequency domain data, and calculate the maximum junction-case temperature T (jc) Get max.

  Next, the T (jc) max calculation unit 206 at the minimum frequency in the section shown in FIG. 1 calculates the maximum value T (jc) max of the junction-case temperature at the minimum output frequency in the section. . Next, in the T (jc) max frequency dependence calculation unit 207 shown in FIG. 1, the two sets of output frequencies f obtained from 205 and 206 and the maximum value T (jc) max of the junction-case temperature are obtained. The constants a and b are obtained from the simultaneous equations of the above equation (4). The obtained constants a and b are substituted again into the equation (4) to obtain time series data of the maximum value T (j-c) max of the junction-case temperature in the section where the frequency changes.

For each section, calculations are repeated by repeating the operations in 205 to 207, and time series data of the maximum junction-case temperature T (jc) max in all sections is obtained.
In the Tj calculation unit 208 shown in FIG. 1, the case temperature Tc obtained from the T (cf) and T (fa) calculation unit 204 and the maximum junction-case temperature value T (jc) max are calculated as follows. From equation (5),
Tjmax = T (jc) max + Tc (5)
Thus, time series data of the maximum junction temperature Tjmax is obtained.

  The Tjmax transient response correction calculation unit 209 shown in FIG. 1 is a calculation unit that adds transient response correction to the maximum junction temperature Tjmax obtained by the above procedure. The maximum value Tjmax of the junction temperature obtained by the previous calculation procedure is a value in a steady state in the repetitive operation. Therefore, immediately after the section is switched and the operation mode is changed, it is in a transient state, and thus the calculated temperature rise value is not reached. Therefore, transient response correction is performed after the elapse of time t immediately after the section switching to obtain a final maximum junction temperature Tjmax, which is output from the calculation result output unit 210.

The temperature rise waveform immediately after the section switching is obtained from equation (1).
The value at the start of the interval is Tjmax (0), the value obtained in the Tj calculation unit 208 is Tjmax (sat), the temperature rise time constant specific to the semiconductor element stored in the database 203 is Tth, By substituting into (1), the following equation (6) for calculating Tjmax after transient response correction is obtained. And the calculation result by Formula (6) is shown in FIG.

Tjmax = Tjmax (sat) {1-(1-(Tjmax (0) / Tjmax (sat)) exp (-t / Tth)} (6)
As can be seen from the waveform shown in FIG. 6, the final maximum junction temperature Tjmax can be accurately obtained by simulation by adding the transient response correction after the time t has passed immediately after the section switching to the Tjmax before the transient response correction. It can be estimated.

  In the above example, as shown in FIG. 8, the acceleration period (section 1) in which the output frequency f changes from 0 Hz to the rated frequency, the steady operation period (section 2), and the output frequency f is 0 Hz from the rated frequency. The deceleration section (section 3) that changes up to is explained as one cycle of PWM inverter operation. This is for explaining an example in which a section such as “stop to rating, rating, rating to stop” is the simplest.

  Not only in this example, even when a complicated pattern with three or more sections such as “stop to rating, deceleration, reacceleration, rating, rating to stop” is defined as one cycle of PWM inverter operation The average generated loss calculation unit 202 can calculate the average generated loss in the semiconductor elements in each section based on the inverter operating conditions in each input section. The subsequent calculation method is the same.

  In this way, since the average generated loss in the semiconductor element can be calculated for each section, even in the case of a complex pattern cycle with three or more sections, the semiconductor element considering the operation pattern of the inverter device The maximum junction temperature can be accurately calculated in consideration of the inverter output frequency. Further, the amount of calculation of loss and temperature and the burden of designing the inverter device can be reduced with a smaller number of calculations than in the past.

It is a figure explaining the flow of the process for junction temperature calculation by the simulation method which concerns on embodiment of this invention. It is a figure which shows the structure outline | summary of the semiconductor chip of an inverter apparatus, a case, and a heat sink. It is a figure which shows the thermal resistance model employ | adopted by this embodiment. It is a figure which shows the current waveform of 1 period, a loss waveform, and a junction-case temperature waveform. It is a figure which shows the frequency dependence of T (j-c) max. It is a figure which shows the transient response correction | amendment of Tjmax performed in this embodiment. It is a figure which shows the typical circuit structure which comprises an inverter apparatus. It is a figure which shows the example of a general driving | running operation | movement of an inverter apparatus. It is a figure which shows each waveform of the inverter output current at the time of acceleration operation shown in FIG. 8, generation | occurrence | production loss, and junction temperature Tj.

Explanation of symbols

101 Semiconductor elements constituting the inverter device
102 DC link voltage
103 heat sink
201 Inverter operation condition and operation pattern input section
202 Semiconductor element average generated loss calculation section in each section
203 Semiconductor Device Characteristics Database
204 T (cf), T (fa) calculation unit
205 T (jc) max calculation section at maximum frequency in the interval
206 T (jc) max calculation section at minimum frequency in the interval
207 T (jc) max frequency dependence calculator
208 Tj operation part
209 Tjmax transient response compensation calculator
210 Calculation result output section

Claims (3)

In the simulation method for calculating the temperature rise value of the components of the inverter device in the operation pattern of one cycle from the start to the stop of the operation of the variable speed PWM inverter device composed of a plurality of semiconductor elements and a heat sink,
Selecting a semiconductor element constituting the inverter device;
When the ambient temperature Ta, the heat sink transient thermal resistance value Zth (fa), the output load current value I, the output frequency f, the carrier frequency fc, the control factor λ, the power factor φ, and the gate resistance Rg of the semiconductor element during one cycle. A process of inputting series data;
Semiconductor element loss characteristics, junction-case transient thermal resistance value Zth (jc) and case-heatsink transient thermal resistance value Zth (cf), database of the selected semiconductor element loss characteristics and the input Calculating time series data of the generated loss of the semiconductor element based on the operating pattern,
Based on the calculated generated loss, the case-heatsink transient thermal resistance value Zth (cf), and the junction-case transient thermal resistance value Zth (fa), the case-heatsink temperature rise T (cf), the heatsink- Calculating the ambient temperature rise T (fa);
In order to obtain the maximum value T (jc) max of the junction-to-case temperature rise T (jc) of the semiconductor element during the period in which the inverter output frequency changes continuously, in one cycle at two output frequencies within that period Calculating T (jc) max, obtaining a characteristic function of the output frequency and T (jc) max from these two T (jc) max values, and calculating T (jc) max at all frequencies;
Maximum value T (jc) max of junction-case temperature rise T (jc) of the calculated semiconductor element, case-heat sink temperature rise T (cf), heat sink-ambient temperature rise T (fa), ambient temperature Ta Calculating the maximum junction temperature Tjmax from
A simulation method comprising: In calculating the maximum value T (jc) max of the junction-case temperature rise, the characteristic function of the output frequency of the inverter device and T (jc) max, where a and b are constants,
T (jc) max = a / f + b
The simulation method according to claim 1, wherein the calculation method is as follows.   In the calculation of the maximum value Tjmax of the junction temperature, the temperature change immediately after the operation condition of the operation pattern changes is calculated by regarding the thermal resistance model of the inverter component as an equivalent first-order lag circuit. Item 2. The simulation method according to Item 1.