Research Article  Open Access
Bahram Ashrafi, Mehdi Niroomand, "Novel Reduced Parts Online Uninterruptible Power Supply", Advances in Power Electronics, vol. 2012, Article ID 502763, 8 pages, 2012. https://doi.org/10.1155/2012/502763
Novel Reduced Parts Online Uninterruptible Power Supply
Abstract
This paper presents design consideration and performance analysis of novel reduced parts online threephase uninterruptible power supply (UPS) system. The proposed UPS system is based on reduced switch count dual bridge matrix converter. It employs only six power switches and results in reducing the cost of the system compared to conventional online UPS topologies, while achieving excellent performance. The performance of the proposed system is evaluated through simulation in terms of input/output waveforms quality and shows the viability of topology.
1. Introduction
Protecting sensitive loads like medical facilities, life supporting systems, data centers, telecommunication systems, and industrial processing systems are the main purpose of implementing uninterruptible power supplies (UPSs system) [1]. UPS system has to provide distortion free, high quality waveforms in order to ensure the reliable operation of the load, regardless of quality of the AC power source [2].
Different types of UPS systems are developed since the advent of UPS concept. Among these types, the online UPS is the most common type for low to medium power applications due to its superior performance such as high tolerance to the input voltage variation and precise regulation of the output voltage [3]. However, conventional online UPS topology employs big number of switches and as a consequence, suffers from high cost, which stops it from adopting in some cost sensitive applications like personal computers. So the idea of developing different topologies to reduce the cost of UPS systems attracts researcher’s attention [4].
In general, the cost reduction is achieved by developing topologies in order to [5](i)reduce the number of power switches,(ii)replace power switches with diodes, because diodes are cheaper than active power switches.
Employing the mentioned methods would further reduce the cost by eliminating the circuitry for driving power switches and simplifying the control circuit:(i)lowering the voltage stress across power switches,(ii)eliminating passive components like transformers, capacitors, and inductors.
Removing bulky passive elements and lowering the number of power switches would bring some advantages like greater compactness and higher reliability.
This paper present the design and performance analysis of a new online threephase UPS system based on dual bridge matrix converter topology. The concept of reducing the number of switches is applied by using 9 switch dual bridge matrix converter. Further switch reduction is obtained by using capacitor midpoint connection (two series capacitors in dclink) which leads to 6 switch topology.
The rest of the paper is organized as follows. In the first part, brief introduction to online UPS is presented and three different operation modes are discussed. In the next part the applied topology is introduced. It starts with dual bridge matrix converter and description of the switch reduction procedure from 24 to 9 and continues with capacitor midpoint connection concept, which further reduced the switches count to 6. Finally, simulation results are presented to demonstrate the feasibility of the proposed method.
2. Online Uninterruptible Power Supply (UPS)
Double conversion online UPS is depicted in Figure 1, based on IEC standard 620403.
As shown in Figure 1, this topology consists of rectifier, battery management set, battery pack, inverter, and static switches (bypass) [6]. As the term “Double Conversion” suggests, an online UPS converts power twice. At first, AC input voltage with all the distractions is converted into DC voltage. For stabilizing the DC voltage, a capacitor is used at the end of the rectification stage.
Second, DC voltage is converted back to regulated AC voltage to supply the load with high quality, uninterrupted power. The output AC voltage can even manage to have different frequency than input voltage, which is not possible with lineinteractive UPS. When the AC input is available (normal mode of operation), described double conversion process would occur. In online UPS, inverter is rated at 100% of the load in order to supply the load, and rectifier power rating is set to meet the 100% of the power demanded by the load and power needed to charge the battery bank [2].
When AC input is outside of the desired range, the UPS is disconnected from AC source and the power is drawn from the battery, so the output is not affected. The transition time for switching from power source to the battery set is several milliseconds, and the capacitor in the dclink supplies the load during transition time. So even there is a short interruption in power flow, the output voltage would remain continuous.
In brief, online UPS has 3 different operating modes: normal mode, stored energy mode, and bypass mode [1, 2].
2.1. Normal Mode of Operation
In this mode, input AC source supplies the load through double conversion process, which is ACDC via frontend rectifier and DCAC via backend inverter. The dclink voltage is also used to charge the batteries. In comparison with lineinteractive UPS, online topology has higher power rating and results in higher cost. The efficiency of online UPS is usually lower than lineinteractive one, so it consumes more power and produces more heat, which results in higher utility cost and additional cooling expense.
2.2. StoredEnergy Mode of Operation
In this mode the UPS is disconnected from AC input source, due to the distraction of the voltage source. The continuity of power flow is maintained by batteries and inverter. The duration of this mode depends on availability of AC line and batteries backup time.
2.3. Bypass Mode of Operation
In bypass mode, load is directly connected and fed by AC input source. It should be considered that the source frequency and load frequency should be equal in order to make the direct connection between source and load possible.
UPS operates in this mode when there is an internal malfunction or in case of maintenance.
The main advantages of online UPS are as follows [3, 4].(i)High tolerance against the input voltage variations.(ii)Regulation of the output voltage.(iii)Controllable output frequency. It can perform frequency conversion (e.g., 50 Hz to 60 Hz).
The main disadvantages of this topology are low power factor, low efficiency, and high THD.
The rectification stage causes distortion to the input current. So an extra power factor correction stage is needed, which leads to higher cost:(i)higher power losses compared to lineinteractive UPS,(ii)shorter life span compared to other topologies due to higher temperature.
Double conversion online UPS, with all mentioned drawbacks, is the most preferred topology in performance, power conditioning, and load protection. There is a large diversity in online UPS topologies to fulfill different tasks in different kind of applications, in range of few KVA to several MVA.
3. Matrix Converter
The matrix converter (Figure 2), first introduced in 1980, is an ac to ac force commutated converter structure. It is able to provide simultaneous voltage and frequency transformation [6].
Matrix converter employs fourquadrant switches, and by choosing proper modulation technique, it makes bidirectional power flow from source to load possible. So it inherently can regenerate energy back to source. Matrix converter does not employ any electrolytic capacitors or any other passive elements in dclink, so it shows high reliability. These features make matrix converter suitable topology in applications like AC drives or power supplies [7, 8].
Recent trends to make higher quality, more efficient power supplies, caused renowned attention and extensive research work on matrix converter due to its superior characteristics like [5]:(i)sinusoidal input and output waveforms,(ii)adjustable power factor,(iii)bidirectional power flow,(iv)compact design (no bulky energy storage is needed),(v)adjustable frequency.
In general, there are two major derivations in matrix converter topology. First one is conventional matrix converter in Figure 2, and the second one is dual bridge [9–11] approach which has “dc link” [5] and is depicted in Figure 3. Dual bridge matrix converter also called “indirect matrix converter” in the literature [11–13].
3.1. Conventional Matrix Converter (CMC)
CMCs have all the advantages mentioned above and its function, based on Figure 2, can be described as follows.
If the switching function of switches is defined as “0” when open and “1” when closed, the output voltage is defined by (1) as follows:
But there are some drawbacks that stop CMCs from being utilized in industry such as the following:(i)potential commutation problem,(ii)complicated control algorithm with the need of high performance microprocessor,(iii)realization of bidirectional switches, (iv)elaborate clamp circuit.
As a result, in the past few years the concept of dual bridge matrix converter has developed.
3.2. Dual Bridge Matrix Converter (DBMC)
Compared to CMCs, the recently introduced DBMC, Figure 3, has the following advantages [7].(i)The difficulty of commutation is reduced. All line side switches turn on and off at zero current and all load side switches commutate similar to a conventional DC/AC inverter.(ii)The number of switches can be reduced under certain constraints.(iii)The clamp circuit can be greatly simplified, only one capacitor and one diode are used.
Considering all the advantages that DBMCs offer and the fact that this topology makes the switch reduction possible [7], this topology is chosen and in the next two sections the switch reduction from 24 to 6 is described.
3.3. Steps to Reduce the Switch Number
Reducing the number of switches in DBMC is extensively discussed in the literature such as [7, 12, 14]. In this part, a brief review is presented.
In general 12 bidirectional switches in Figure 3 require 24 unidirectional switches, but by proper assumptions the number of switches can be reduced [7].
In the following part, switch reduction procedure is described in four steps.
Step 1. If, by employing appropriate control technique, the dclink voltage is set to be positive, then replacing bidirectional switches with unidirectional switches at the load side is possible. This leads to 18switch topology which is depicted in Figure 4 [7].
Step 2. According to Figure 5 [11, 14], and can share the same gate signal and can be replaced by one bidirectional switch and two diodes.
So 15switch [10, 11] topology as shown in Figure 6 can be achieved. This topology is also called “sparse matrix converter” in [11–13, 15].
Step 3. By assumption, further switches reduction from 15 to 12 is possible. This leads to 12switch topology in Figure 7. This topology is a standard 6switch rectifier and 6switch inverter connected back to back.
Step 4. As far as switches and () in Figure 6 are not conducting current and only , , and are conducting, the 9switch topology (also named “ultra sparse matrix converter [11–13, 15]”) is obtained and depicted in Figure 8.
By reducing the number of switches to 9, the presented topology has lost some of its desired features compared to standard and 24switch topology, like bidirectional power flow between source and load. In 9switch topology power only flows from source to load and power factor should always be higher than 0.866.
4. Proposed Topology
So far the 9switch topology is obtained. In this section, we intend to further reduce the switch count by applying the concept of sharing leg between rectification and inversion stages [13–16]. Rectifier with capacitor dclink is considered as one of the most promising structures for threephase to threephase power conversion [14].
In order to achieve this, a leg consists of two series capacitors is added and placed in parallel with rectification stage. The midpoint, tagged by “0” in Figure 9, is shared between an input leg and an output leg and results in elimination of one active switch and four diodes in rectification stage and two active switches in inversion stage.
By replacing the proposed 6switches converter with rectifier block and inverter block in Figure 1, the novel 6switches UPS is achieved which fulfill desired characteristics of double conversion UPS, such as compactness, output voltage regulation, and controllable frequency.
Three different operation mode of proposed UPS system are as follows. (i)In normal mode, double conversion process take place through proposed 6switches matrix converter.(ii)In storedenergy mode, the batteries are in parallel with inversion stage and feed the load.(iii)In bypass mode, the bypass switch in Figure 9 is closed and load is directly connected to the source.
Reliability is considered one of the most important characteristics of UPS system. It is indicated in the literature that 92% of all UPS failures are related to battery malfunction. This happens because of electrochemical nature of batteries. So providing the best operating condition for battery bank is of high importance when designing UPS system [1]. Since we can tightly regulate dclink voltage in presented topology, the battery bank, which is consist of two set of batteries, is directly connected to the split capacitor dc bus.
By applying appropriate control technique, achieving sinusoidal output waveforms and controlled dclink voltage is possible.
The control circuit in rectification stage consists of multiple control lops for each switches, one outer voltage loop and one inner current loop [1]. The first loop takes dclink capacitor voltage as a feedback signal and compares it with reference voltage, which calculates error signal and by processing that signal, reference signal for second loop can be created. In second loop input current is sampled and compared with generated reference in first loop and PWM waveform with 30 kHz switching frequency is created and applied to active switches.
The 4 switches output inverter is controlled in a high frequency PWM pattern in order to produce high quality output voltage. Reference voltages for each of two legs are 120 degree shifted from each other. So produced line voltages are 120 degree shifted. As a result, the line voltage shared leg is also 120 degree shifted which completes symmetrical system of line voltages.
The control technique for output converter uses two control loops for each leg, one outer voltage loop and one inner current loop. As mentioned above, two reference voltages for each leg are 120 degree shifted from each other. Output voltage is sampled to calculate the error signal. For achieving stable voltage during steady state operation, PID module is used.
Calculated voltage error is also used as a reference signal for current loop. This loop uses the output line current as feedback signal. This loop is faster and is used to improve the dynamic response of the output converter.
In the next section, simulation result of the proposed UPS system is provided. Using the aforementioned control strategee will guaranty the performance of the system. In the rated power, during the steady state operation, as simulation results demonstrate, THD is lower than 3.2%, efficiency is about 89%, and sinusoidal output voltage and current for different load conditions are achieved.
5. Simulation Results
Simulation is carried out by using PSIM software. In the simulation, all switches are considered ideal. The voltage and current waveforms of the input, dclink voltage as well as load current and voltage are depicted in this part to demonstrate the performance of proposed 6switch topology (Figure 9) and applied control technique. The system parameters of simulated circuit are given in Table 1.

Figure 10 shows the simulation result for input current. In Figure 11, voltage and current phase “a” are shown. According to simulation result, power factor in input stage is 0.94.
Simulation result for dclink voltage and current is shown in Figures 12 and 13.
DCLink voltage is controlled using hysteresis controller. As shown in Figure 12, the voltage is fairly fitted between two set values.
Dc current is always positive which satisfies Step 3 in switch reduction procedure.
Figure 14 shows the output current for threephase load connected to the backend inverter. The applied current controller shows acceptable performance to supply load with threephase sinusoidal current.
In Figure 15, the output voltage is shown. In order to demonstrate output voltage, a low pass filter is used and the cutoff frequency is set to 1000 Hz.
Demonstration of output voltage and current of phase “a” is presented in Figure 16. In this figure, the performance of the system for three different load conditions is shown, from resistive load to 120% of rated load. In this range, the proposed UPS system demonstrates a good performance in achieving sinusoidal waveforms and high power factor.
Figure 17 demonstrates output voltage in frequency domain. For calculating total harmonic distortion, the THD box in PSIM software is used. In steady state operation, THD is equal to 3.2%.
6. Conclusions
The performance of proposed 6switch online UPS is investigated in this paper. Applied topology is based on the reduced switch count dual bridge matrix converter. Reducing the number of power switches leads to low price, highly reliable, and compact design. The proposed topology shows excellent performance, produced high quality, sinusoidal output voltage, and current waveforms. Simulation results have been presented to confirm the advantages of the proposed novel UPS system.
References
 A. Emadi, A. Nasiri, and S. B. Bekiarov, Uninterruptible Power Supplies and Active Filters, CRC Press, 2005.
 H. R. Karshenas and M. Niroomand, “Design and implementation of a single phase inverter with sine wave tracking method for emergency power supply with high performance reference,” in Proceedings of the 8th International Conference on Electrical Machines and Systems (ICEMS '05), pp. 1232–1237, September 2005. View at: Google Scholar
 M. Niroomand and H. R. Karshenas, “Hybrid learning control strategy for threephase uninterruptible power supply,” IET Power Electronics, vol. 4, no. 7, pp. 799–807, 2011. View at: Publisher Site  Google Scholar
 M. Niroomand and H. R. Karshenas, “Design and implementation of a new hybrid learning control strategy for threephase UPS,” in Proceedings of the 11th International Conference on Electrical Machines and Systems (ICEMS ;08), pp. 1732–1737, October 2008. View at: Google Scholar
 A. Nasiri, S. B. Bekiarov, and A. Emadi, “Reduced parts threephase seriesparallel UPS system with active filter capabilities,” in Proceedings of the 38th IAS Annual Meeting, IEEE Industry Applications Conference, vol. 2, pp. 963–969, October 2003. View at: Google Scholar
 M. Niroomand and H. R. Karshenas, “Review and comparison of control methods for uninterruptible power supplies,” in Proceedings of the Power Electronic & Drive Systems & Technologies Conference (PEDSTC '10), pp. 18–23, IEEE, 2010. View at: Google Scholar
 A. Nasiri, “Digital control of threephase seriesparallel uninterruptible power supply systems,” IEEE Transactions on Power Electronics, vol. 22, no. 4, pp. 1116–1127, 2007. View at: Publisher Site  Google Scholar
 H. R. Karshenas and M. Niroomand, “Analysis and comparison of two control strategies for threephase seriesparallel UPS,” in Proceedings of the 11th International Conference on Electrical Machines and Systems (ICEMS '08), pp. 1726–1731, October 2008. View at: Google Scholar
 J. Holtz and U. Boelkens, “Direct frequency converter with sinusoidal line currents for speedvariable AC motors,” IEEE Transactions on Industrial Electronics, vol. 36, no. 4, pp. 475–479, 1989. View at: Publisher Site  Google Scholar
 L. Wei and T. A. Lipo, “A novel matrix converter topology with simple commutation,” in Proceedings of the 36th IAS Annual Meeting, pp. 1749–1754, October 2001. View at: Google Scholar
 J. W. Kolar, M. Baumann, F. Schafmeister, and H. Ertl, “Novel threephase ACDCAC sparse matrix converter,” in Proceedings of the 17th Annual IEEE Applied Power Electronics Conference and Expositions, pp. 777–791, Dallas, Tex, USA, March 2002. View at: Google Scholar
 P. W. Wheeler, J. Rodríguez, J. C. Clare, L. Empringham, and A. Weinstein, “Matrix converters: a technology review,” IEEE Transactions on Industrial Electronics, vol. 49, no. 2, pp. 276–288, 2002. View at: Publisher Site  Google Scholar
 J. W. Kolar, F. Schafmeister, S. D. Round, and H. Ertl, “Novel threePhase ACAC sparse matrix converters,” IEEE Transactions on Power Electronics, vol. 22, no. 5, pp. 1649–1661, 2007. View at: Publisher Site  Google Scholar
 L. Wei, T. A. Lipo, and H. Chan, “Matrix converter topologies with reduced number of switches,” in Proceedings of the IEEE 33rd Annual Power Electronics Specialists Conference (PESC '02), pp. 57–63, June 2002. View at: Google Scholar
 C. Klumpner, M. Lee, C. Pitic, P. Wheeler, and P. Zanchetta, “A new threelevel indirect matrix converter with reduced number of switches,” in Proceedings of the IEEE Industry Applications Conference 42nd Annual Meeting, pp. 186–193, September 2007. View at: Google Scholar
 L. Wei, T. A. Lipo, and H. Chan, “Robust voltage commutation of the conventional matrix converter,” in Proceedings of the IEEE 34th Annual Power Electronics Specialists Conference, pp. 717–722, June 2003. View at: Google Scholar
Copyright
Copyright © 2012 Bahram Ashrafi and Mehdi Niroomand. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.