Thermal monitoring of a converter

Abstract

The thermal monitoring of a converter for an electric motor of a vehicle should be improved. Therefore, the invention relates to a method in which the temperature of the converter is measured. In addition, the output power of the converter is determined. The temperature of the converter is estimated using a thermal model of the converter having the output power as an input variable. The estimated temperature is compared with the measured temperature, and a corresponding monitoring signal is provided. Thus, a deviation of the measured temperature for the calculated temperature and thus a corresponding error in the cooling system can be detected even in the partial-load range.

Claims

1. A method for thermal monitoring of a converter for an electric motor of a vehicle comprising: measuring a temperature of the converter; determining a power outputted by the converter; providing a thermal model of the converter, said thermal model having the power outputted by the converter as an input variable, and being a linear function over time; estimating the temperature of the converter by means of the thermal model using the determined outputted power as the input variable; comparing the estimated temperature with the measured temperature; generating a monitoring signal as a function of the comparison; and generating a warning signal in response to the monitoring signal, when a difference between estimated and measured temperature exceeds a preset level, thereby informing a driver that cooling of the converter is not taking place.

2. The method of claim 1, further comprising calculating with the thermal model a power loss as a function of the outputted power.

3. The method of claim 2, wherein in the thermal model the power loss is averaged and normalized over a preset calculation time period, and the normalized power loss occurs in the thermal model as a factor for the estimation of the temperature.

4. The method of claim 1, further comprising experimentally determining a time constant for the thermal model.

5. The method of claim 1, wherein a cooling capacity with which the converter is cooled is used in the thermal model as a variable.

6. The method of claim 1, wherein a start temperature of the converter which is measured when the vehicle is started up or as a starting point for an estimate forms the basis in the thermal model.

7. The method of claim 1, wherein the measured start temperature is used as a variable in the thermal model.

8. The method of claim 1, wherein the thermal model for estimating the converter temperature is as follows:
(t):=[T.sub.0(1+t/T)]*{tilde over (P)}.sub.v where T.sub.0 is the start temperature for the estimation, {tilde over (P)}.sub.v, is a normalized, mean power loss in the most recent calculation time period and T is the experimentally determined time constant.

Description

BRIEF DESCRIPTION OF THE DRAWING

(1) The present invention will now be described in detail with reference to the attached drawings. in the drawings:

(2) FIG. 1 shows a temperature model having a first time constant and

(3) FIG. 2 shows a temperature model having a second time constant which is halved compared with the first time constant.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

(4) The thermal time constant, in other words that time in which the converter heats up by a predetermined temperature value, differs significantly with and without a functioning water cooling system. In the case of specified normal use, the time constant can be easily determined and stored in the converter or the monitoring system thereof. With the time constant, a thermal model of the converter temperature can be created.

(5) An important factor of the thermal model is the power loss of the converter. The power loss is that power which is ultimately converted into heat in the converter. It can be determined according to the following formula:
P.sub.v=(1)P.sub.ges

(6) where corresponds to the efficiency of the converter, in other words the quotient of output power and input power. P.sub.ges corresponds to the input power, in other words the total power flowing into the converter. Said power P.sub.ges can be calculated from the variables normally present for regulating the drive, such as intermediate circuit voltage and output current.

(7) The efficiency of the converter can be determined for example experimentally during a type approval test. In this situation, a dependency of the efficiency on other variables, such as ambient temperature, can under certain circumstances also be taken into consideration.

(8) From type approval test results regarding heating tests it is moreover possible to determine the thermal time constant of the system, in other words that time in which the temperature of the system increases by a preset temperature value. Said thermal time constant is, as has already been indicated above, stored in the converter and is then available for calculation or estimation of the converter temperature to be expected in a particular operating state. The corresponding thermal model for estimating the converter temperature can be as follows:

(9) $\left(t\right):=\left[T_0\left(1+\frac{t}{}\right)\right]*{\tilde{P}}_v$

(10) where T.sub.0 is the start temperature for the estimation, {tilde over (P)}.sub.v is a normalized, mean power loss in the most recent calculation time period and T is the experimentally determined time constant. The start temperature T.sub.0 can correspond to the temperature of the converter when the vehicle starts up, but as a rule will not do so. The start temperature T.sub.0 is generally speaking rather that temperature which serves as the starting point for a calculation time period. The mean power loss {tilde over (P)}.sub.v is normalized in respect of a preset power or an experimentally determined power.

(11) The thermal model presented above has deliberately been kept very simple. This means that a relatively low amount of processing power is required for estimating the temperature. In principle however, other temperature models can also be used for estimating the converter temperature. In particular, they can also be significantly more complex with regard to the calculation. For example, the thermal model can be more complex if an additional air cooling of the converter is taken into consideration.

(12) The temperature calculated or estimated with the aid of the thermal model is compared with a measured temperature of the converter. In the event of an error the measured temperature already deviates significantly from the calculated temperature even in the partial-load range. This means that for example in the case of a vehicle which is moved with low power on the flat a considerably higher temperature is measured than is expected. Even if the measured temperature has still not reached a system-critical point the opportunity is already available to report the failure of the cooling system to the driver. The converter is then still fully functional and the vehicle is thereby also still fully maneuverable. As a result of the reconciliation or comparison between estimate and measurement a deviation of the thermal time constant is determined which is the first consequence of the failure of the liquid cooling system as a result of a leak/failure of the coolant pump. The rapid temperature rise is in turn a consequence of the change in said time constant. In other words, a cause of the rapid rise in temperature is detected which was not previously accessible. It thereby becomes possible to bring the advance warning time for the vehicle driver back to values such as are familiar to the driver from conventional internal combustion engines.

(13) The estimation and monitoring of the temperature of the converter can be implemented with the aid of software in the converter. Since this is a matter of thermal states, the calculation can take place in low-priority time slices. It will nevertheless even then have a sufficiently accurate outcome.

(14) FIG. 1 and FIG. 2 in each case show the temperatures plotted over time for different thermal models. The linear model described above is presupposed. The value 1 is assumed as the normalized mean power loss {tilde over (P)}.sub.v. The progression of the estimated temperature (t) illustrated in FIG. 1 then results for a time constant T=50 s and a start temperature T.sub.0=25 C. The estimated temperature rises from the initial value 25 C. to the final value of 50 C. after 100 s.

(15) If on the other hand a thermal time constant T=50 s is determined, then the linear progression shown in FIG. 2 results. According to that the estimated temperature (t) rises from the initial value 25 C. to 75 C. In the case of FIG. 2 the converter therefore has a smaller thermal time constant than in FIG. 1. The reason for this may consist in the fact that the converter from FIG. 2 is cooled less effectively than the converter from FIG. 1.

(16) If the measured temperature in each case lies significantly above that which has been estimated with the aid of the thermal model, then this can be regarded as an indication of the fact that the cooling is not operating in the desired manner. As a result of the plausibilization according to the invention of the measured converter temperature it is not necessary to monitor the cooling system directly. In particular, it is not necessary to measure the flow rate and the pressure in the cooling system, from which a failure of the coolant pump or a leak with loss of pressure could be deduced. It is frequently possible to dispense with additional sensors which would be accompanied by additional costs and an increase in complexity and susceptibility to faults of the system.