Oil-immersed transformers are to operate more long life than designed lifetime. Therefore it is necessary to maintain and replace aged transformers appropriately by applied deterioration diagnosis, to keep. In this paper, the way of deterioration for oil-immersed transformers is introduced. Examples of deterioration diagnosis technique research such as diagnosis focused on decomposition product from insulating paper, optical diagnosis and the temperature stress analysis are introduced.

View full abstract

This paper proposes a new voltage control scheme for distribution networks based on a multi-agent system. Multiple On-Load Tap Changing transformers including SVRs, each of which regulates multiple points of voltages, are optimally controlled. Large voltage fluctuations caused by PV generations are effectively regulated. We achieved an optimal control performance by developing an optimal control law that is applicable to decentralized autonomous control. The autonomy is accomplished by using a multi-agent system instead of the centralized control scheme. The effectiveness of the proposed scheme is demonstrated through numerical simulations with successful results.

View full abstract

Large amount penetration of renewable energies as distributed generators enlarges voltage fluctuation in distribution system. This fluctuation might be suppressed within allowable range by centralized voltage control. However, centralized control needs high-cost broadband communication network. This paper proposes a decentralized control method approximately minimizing voltage deviation with local measurements. In the proposed method, regulators control voltage independently with decentralized control models determined by linear regression of minimum voltage deviation control. The simulation results showed that the proposed method regulated voltage within allowable range in severe conditions with a large amount of renewable energies.

View full abstract

The concept of microgrid (MG), which can integrate a number of distributed generators and realize high efficient energy management, is attracting great attentions. The authors are investigating the feasibility and effectiveness of DC supplying MG (DCMG) for an office building, from the viewpoint of environmental-load reduction. In this paper, an optimal scheduling for DCMG, which can consider the PV forecast errors and their impacts on operation performance, is introduced. Additionally, second time interval operation algorithm is also introduced. Effectiveness of the proposed algorithm is evaluated through computational simulations using the actual data.

View full abstract

It is important to maintain the voltage profile in distribution systems with a large amount of distributed generators (DGs). Though it is effective to control reactive power to maintain voltage profile, there is a possibility that active power curtailment is needed. To solve this problem, the authors have developed so far the pricing method of reactive power controlled by DGs to give economic incentives for those DG owners. In order to realize the method in real-time, we suppose that measurement data by sensor switches are available to determine the pricing and the price signal can be sent to DG owners through smart meters. Based on the concept of “shallow”, the required amount of resource in order to procure the reactive power is imposed to all the DG owners equally. The effectiveness of the proposed method was tested by numerical simulation by using a distribution system model consisting of one feeder line.

View full abstract

It is of prime importance to maintain voltage profile within the proper range in distribution systems with a large amount of photovoltaics (PVs) and electric vehicles (EVs). In particular, there is a possibility that LDC logic which is utilized for the control of load ratio tap transformer (LRT) does not work properly when reverse power flow is included partially. Hence, in this paper, we have developed a new LRT control method based on the sensor information supposing that some section switchgears with sensors are introduced in the future distribution systems. Specifically, the extreme value of voltage profile is estimated by convergence calculation in the section between the section switchgear with sensor and LRT. Moreover, the voltage at the end node can be estimated by LDC method using sensor information of section switchgear. The proposed method was tested using a distribution system model and its effectiveness was shown.

View full abstract

Demand response can achieve peak-cut and peak-shift of the electric demand by delivering the load restraint of customers, based on Demand Response (DR) signal sent by an electric power company. However, the demand response restrains the load amount of many consumers simultaneously, so the voltage of the whole distribution system is greatly fluctuated depending on the distribution form such as line length, amount of load restraint, etc. Therefore, we propose the voltage control methods which can avoid voltage deviation in consideration of a time constant of demand response. Furthermore, we propose the BESS cooperating voltage control methods with LRT, SVR and verify the effectiveness using experiment of distribution system simulator.

View full abstract

It becomes an important issue to maintain voltage profile within the allowable range in distribution systems with a large number of distributed generators (DGs), such as Photovoltaic (PV). Although reactive power control by power conditioning system (PCS) is effective to mitigate the voltage increase, there is a possibility that active power curtailment is needed to provide enough amount of reactive power due to current limitation. Here, it is expected that the active power curtailment can be avoided if a cooperative voltage control between load ratio tap controller (LRT) and PCS is realized. To this end, system operators have to know the amount of reactive power controlled by PCS. Hence, in this paper, we developed a new state estimation method of the reactive power using the measurement data obtained by the switchgears with sensors. In addition, a probabilistic approach was applied because an accurate estimation is theoretically impossible only by using the limited number of measurement data. The proposed method was tested in a hierarchic distribution system model and its effectiveness was shown through simulations.

View full abstract

The introduction of photovoltaic generation (PV) is expanding and, in accordance with that, it is concerned that power system stability deteriorates by PV without a system stabilizing function. Therefore, PV is required to have system stabilizing functions such as fault ride through (FRT) and dynamic voltage support (DVS) function. It is important to evaluate an influence of PV with these functions on power system stability. In this paper, influence of applying these functions to PV on transient stability and voltage stability is evaluated. As evaluation methods, not only simulation method but also power-angle curve (P-δ curve) of synchronous generator and torque speed curve of induction motor are used. As stated above, improvement effect on power system stability of FRT and DVS function by multiple evaluation is verified.

View full abstract

In this paper, we have constructed an Auto Regressive(AR) and Neural Network (NN) combined type prediction model for responsive change in the room temperature trend due to the Fast Automated Demand Response (FastADR) power limitation of office building air-conditioning facilities. We defined the average of differences between room temperature and setpoint of each indoor unit for the entire building as a FastADR side-effect index for the building. Prediction experiments using an actual office building showed that the root mean square prediction error of our model was 0.23 degrees Celsius for 5 minutes after the FastADR. This prediction ability is considered sufficient to estimate the side-effect of FastADR power limitation.

View full abstract

There are two causes, inductive lightning surge and direct lightning stroke, as the aerial insulated wire accidents accompanying lightning in a power distribution system. When the overvoltage due to the inductive lightning surge invades to the wire core, the creeping discharge can develop along the wire surface from the free end of the binding wire just after a flashover of the post insulator at the wire supporting point. This creeping discharge may give rise to the accidents such as a melting or snapping of the wire. The creeping discharge along the wire surface has the positive or negative polarity. Positive creeping discharge is restricted to the area where a positive lightning generates. Only a few examples have been reported on the positive creeping discharge and its characteristic has many unsolved points. In the previous studies, we have observed the positive creeping discharges along the wire surface under the negative inductive lightning surge with the peak values in the range |Vm| ≦ 80kV. In this study, the positive creeping discharges are examined newly in the range |Vm| > 80kV. It is reported that the positive creeping discharges are greatly affected by the negative corona discharges generating from the wire surface in |Vm| ≧ 95kV.

View full abstract

In general, the degradation span of 6.6kV XLPE cable with plural connections is determined by DC leakage current measurement. This method needs to disconnect and reconnect the shielding layer and the outer semi-conducting layer at the cable connecting part, and hence it takes a lot of time and cost.  We investigated a new method to locate the water tree degradation by DC leakage current. In 6.6kV XLPE cables the pulse-like leakage current called “kick” due to water tree degradation is often observed. Consequently, we detected the currents at the DC applying terminal and another terminal, and transformed them to the signals like a sine-wave having the phase difference exactly proportional to the time difference of arrival (TDOA) by using ceramic filters at 455kHz. The position of the water tree is calculated from TDOA obtained by the phase difference. In the experiments to locate the water tree using 6.6kV XLPE cable samples up to 85m long, the mean squared error was 2.2m. From the results, it is possible to locate the water tree in 6.6kV XLPE cable 2km long with an error of about 20m by optimizing the center frequency of the ceramic filter.

View full abstract