6.1 Comparison and connections with former research
Previous studies about the measurement of the frequency converter’s DC link electrolytic capacitor where operating principle is based on the constant DC magnetization feature of the DTC control, were not found in the literature.
However, in some previous studies (Kwang-Woon et al. 2008) researchers have used the similar measurement method, which was now further developed. The results obtained in this research were compared to the previous studies. The results based on simulation and practical measurements confirm that the measuring method principle works.
6.2 Objectivity
This research is possible to perform again by using the same methods, regardless of re-searcher. Practical measurements were based on independent and standardized measur-ing equipment.
Measurements were performed with three frequency converters; two of them were of dif-ferent types. The simulation and measurements for the ACS60403206 converter were carried out by changing the capacitance of the capacitor bank. Four different combinations were used. The results based on the simulation, practical measurement and literature research supported each other.
6.3 Reliability and validity
The reliability of the study was evaluated by comparing the results obtained from simula-tions, measurements, and literature. The calculations of the results are based on widely proven electrical engineering equations. In this measuring method, the measuring circuit includes the inverter switches (IGBT), the motor cable and the motor. So these factors may affect the measurement’s validity and reliability. Evaluations of these effects were excluded from the study.
The measurements were carried out using calibrated measuring instruments. The accura-cies of the measuring devices were evaluated. In addition, repetitive tests for the analyzer were performed to find out its measuring uncertainty. The uncertainty of reading from os-cilloscopes figure was also analyzed.
The reliability of the simulation results depends on how accurate the simulation models are compared to real test setup. Now the used model contained ideal components, which do not actually exist in real life. However, the simulation model described the functionality of the measurement method accurately.
The assessment of the measurement method was performed by measuring three different converters. The simulation and practical measurements for the ACS60403206 converter were carried out by changing the capacitance of the capacitor bank. Four different combi-nations were used. The simulation and measuring results are consistent with the previous-ly studied theory, simulation and measurements, which verify this study.
6.4 Assessment of the results and sensitivity analysis
Error and sensitivity analysis of the study was taken into account in the measuring equip-ment, reading uncertainty and practical measurements.
The error analysis of the measuring equipment took place in respect of the technical spec-ification provided by manufacturers. In addition, the analyzer measurement uncertainty was investigated with repetition tests. In the repetition test the same measurement was repeated four times with four different capacitors. Between measurements the capacitor was disconnected. Based on the technical specification the combined accuracy for the measuring equipment was about 3 % and for the analyzer below 1%.
The measuring reading uncertainty of the ESR and capacitance is caused by the DC link voltage jitter and current jitter. Thus, the determination of the precise reading point can be challenging, as it can be deduced in Figure 23. Measurement uncertainty in reading was evaluated by repetition tests with two triggered oscilloscope’s figures. Each triggered fig-ure contained three evaluable pulses which were read four times to estimate ESR and capacitance values. Based on the results the biggest deviation with capacitance values was 21 % and with ESR values it was 15 %. This gives an idea of the sensitiveness of the reading for current and voltage jitter. However, to this matter can be influenced by
reduc-ing the samplreduc-ing frequency of an oscilloscope whereby the jitter is reduced and the image becomes more readable.
For the simulation model the error and sensitivity analysis was not made, and it was left out from the subject. Practical measurements were performed with three frequency con-verters; two of them were different types. The simulation and measurements for the ACS60403206 converter were carried out by changing the capacitance of the capacitor bank. Four different combinations were used.
Based on the results of the simulation the biggest error for ESR was 1.2 % and for capaci-tance 3.7 %. Table 25 shows both measured converter types’ biggest individual meas-urement errors and the biggest deviations between measmeas-urements.
Table 25 Biggest individual measurement error and deviation rates.
Converter type ACS60403206 ACN63401003
ESR C ESR C
Biggest individual measurement error [%] 51 57 16 25
Biggest deviation between measurements [%] 19 19 1 6
The results revealed that there is a bigger measurement error and deviation for ACS60403206 than ACN63401003, which can be a result of the larger capacitor bank.
Because the temperature has a great impact on the capacitance and ESR values (Figure 7), it has to be taken into account. It is important to notice the temperature also when comparing the results of the new measurements to the old results. The measurements must be done preferably when the temperature of the capacitor bank has levelled near to the ambient temperature. However, the measurement itself is fast; triggering a single measurement is less than 300 µs, so it does not greatly affect the capacitors’ warming.
The measuring accuracy is affected by a capacitor bank design. A bigger capacitor bank contains more contact points which causes a bigger transfer resistance. For example, poor contacts can particularly affect the measured value of the ESR, since the nominal value of the capacitor's ESR is small, depending on the type of capacitor normally from a few mΩ to tens of mΩ at 300 Hz. The magnitude of the effect on the total ESR value de-pends on the capacitor bank structure, the number of capacitors connected in parallel and in series. This must be taken into account when measuring same time several inverters which are linked to the same DC link; a higher degree of accuracy is achieved by measur-ing the inverter at a time.
The magnitude of the change of the total ESR and capacitance values depends on the capacitor bank structure, the number of capacitors connected in parallel and in series. In a smaller capacitor bank the impact of a single capacitor increases on the measurement results when comparing to a larger capacitor bank. The magnitude of change in the total ESR and capacitance values is higher with ESR values, if for example one capacitor fails and becomes an open circuit. In conclusion it can be said that the smaller the capacitor bank is the easier the change in measurement results can be seen.
6.5 Key findings
Most important outcome of the study is that the practical measurement showed that the DC link capacitors measuring method is functioning by using the DTC control's a constant DC magnetization property.
A constant DC magnetization was used in this work to control the switches, however, it has its pros and cons. On the plus side can be considered the commissioning of simplici-ty, it requires only to check few parameters and change them, if not already chosen. On the minus side is the lack of adjustment of the pulse width, which is now the standard 48 µs. With a longer pulse width a higher ripple voltage of the DC link could be achieved which would improve the measuring accuracy. In section 5.6 the method's pros and cons are discussed in more detail as well as the method's suitability for the onsite measure-ments.
However, the simulation results and the experimental measurements indicate that the new measuring methods of the DC link capacitors operates with sufficient accuracy, at least if the aging phenomena are based on ESR values.
6.6 Novelty value of the results
In this research, a new measurement method has been applied. So, one of the most im-portant results was to identify the shortcomings and inaccuracies which are included in the now applied DC link capacitor's measuring method. Also this research provided infor-mation of the DC link capacitors’ aging.
6.7 Generalization and utilization of the results
The most important outcome of the study is that the DC link capacitors measuring method is functioning by using constant DC magnetization property of the DTC control. This was proved by the practical measurements and by the simulations. Due to operating principle, the measurement method can also be applied to other frequency converters, where the switches are controlled in a suitable way.
The measuring method can be taken into account when planning the future frequency converters. Then the method could be integrated into the frequency converter functions.
This new promising measurement method will be tested more on site. If the results are still good, this method will be used as an everyday tool.
6.8 Topics for future research
From the safety, measurement accuracy and reliability point of view, the measurement method could be developed further. Table 26 lists possible changes to the measuring method and evaluates their pros and cons.
Table 26. The modification for the measurement method, as well as their pros and cons.
Modification Pros Cons
A load coil instead of a motor
Better measurement accuracy:
o the impact of the motor and the cable could be left out
o accurately known load
Investment cost:
o parameterization takes more time, performing ID-run to the coil, and after the measurement returning the used values
o disconnecting and connecting the motor cables
o connecting and disconnecting the coil
Separate control unit for controlling switches or modi-fying existing pro-gram
Better measurement accuracy:
o pulse width can be adjusted, ena-bles sufficiently big ripple voltage
Development costs:
o requires software and possibly also hardware design
o no external sources of interference (motor, cable, switches and their
o measurement automation improves measurement accuracy
o fastest, automated measurement o just connecting new measuring
device to the DC link Better safety:
o safe, no effect to the outside