suffering a higher risk of voltage violations, as shown in Figs. 4.5 and 4.6 below. As sketched above we run the real-timecontrol heuristic H rt for many cases, each with a different δ and different device sizes depending on that δ. For two of the cases with δ = 0.1 and δ = 1, the real-time traces of voltage magnitude and realpower loss are shown in Fig. 4.5. The traces of voltage and power loss are also plotted for a benchmark case with only a fixed capacitor (whose size equals C 0 ∗ + C ∗ s + q ∗ f when δ = 0.1) and no control. We see that, in the benchmark case with only a fixed capacitor, the larger time-independent reactive power injection from the fixed capacitor results in higher voltages and losses at nearly all times compared to the two cases with controls. The case δ = 0.1 generally biases the voltage above the case δ = 1 (no chance constraints). This bias protects the system against experiencing an undervoltage when the load suddenly increases, as revealed near the end of the day when the voltage in the case δ = 1 dips below 0.99 pu. This extra voltage safety provided by the chance constraints incurs increased power loss during periods when the case δ = 0.1 biases the voltage up with additional reactive power injections. For each case with different δ, we record the proportion of 5-second samples in one day at which the voltage drops below 0.99 pu (indeed the voltage never swings above 1.01 pu so those recorded are all the samples with voltage violations). We also sum up the real-timepower loss over one day, and add up the cost of that power loss and the average capital cost in one day. Fig. 4.6 shows the proportion of samples with voltage violations and the one-day total (capital plus power-loss- induced) cost for different δ. As δ is decreased, the voltage control becomes more reliable as demonstrated by the significantly decreasing proportion of samples with voltage violations in the upper subfigure. The increased reliability only brings a modest increase in cost, as shown in the lower subfigure. Not shown in Fig. 4.6 is a benchmark case with only a fixed capacitor and no control. In that case the fixed capacitor is set high enough so that the voltage never drops below 0.99 pu, but the cost is as high as $215/day due to the high power loss. We also consider another benchmark case in which there is only a D-STATCOM and there are no fixed and switchable capacitors. In this case the deterministic-constrained OPF problem C f (p, C 0 , c ∗ s , q f ) is solved every 5 seconds in realtime with C 0 = c ∗ s = 0 and q f
This chapter demonstrates how text-based specifications can be converted into a temporal logic spec- ification language using a representative single-line diagram as an example. Given a set topology for an electricpower system, as seen in the single-line diagram from Figure 2.1 and a set of system requirements formalized in linear temporal logic, we automatically synthesize a control protocol for an electricpower system on a more-electric aircraft. The resulting controller allows generators and APUs to connect and disconnect to buses through the closing and opening of contactors. The health status of each generator/APU is uncontrollable, and thus considered an environment action. The controller reacts to changes in the environment and is guaranteed, by construction, to satisfy the desired properties even in the presence generator failures. We synthesized a centralized controller where statuses of all components (generators, contactors, and buses) are known. We also created dis- tributed and decentralized controllers by refining the overall system specifications. This refinement involves additional assumptions and guarantees between subsystem interfaces (i.e., specifications on the components that interact with other subsystems). For a distributed controller, we implemented a master/slave architecture where one subsystem has full authority for routing power to the other subsystem. In the decentralized controller design, we allow power exchange between two subsystems to flow in both directions, again refining the interface specifications.
Summary. — The article describes electricpowersystems from a physicist’s point of view. In contrast to common introductory textbooks on powersystems, the em- phasis is on the physical design, that is the material selection and the choice of the geometrical shape, of the fundamental components as it follows from the function and serves the main purpose. Why do power system components look the way they look? This is the question addressed in an accessible way. Four fundamental com- ponents are needed to make the most elementary power system: overhead transmis- sion lines, transformers, synchronous generators, and circuit breakers. High-voltage overhead lines make eﬃcient long-distance transmission of electricpower possible. Transformers step up the power from the generating plant and cascade it down to the ﬁnal consumption. For their ability to control, independently, real and reactive power, synchronous generators are the most common type of generators. And it is only through the immediate extinction of plasma arcs in circuit breakers that short- circuit currents can be interrupted and faulty segments of the grid disconnected.
Pre-Scheduled LS techniques to meet demand and supply imbalance is used in many developing countries. Expected error in load forecasting, manual control, excessive LS and complete feeder disconnection make this method nonviable for powersystems. Automated under frequency LS methods are based on the frequency drop or rate of frequency deviation , . This technique is relatively simple to execute but time delays make its response considerably slow. Time delays are kept in eﬀective range to avoid nuisance trips causing unnecessary LS. Additionally it may result in tripping of distributed generation source .
The central module in our implementation provides support for pe- riodic real-time tasks in Linux. This is done by attaching call-back functions to hooks inside the Linux scheduler and timer tick han- dlers. This mechanism allows our modules to provide tight timing control as well as override the default Unix scheduling policy for our real-time tasks. Note that this module does not actually define the real-time scheduling policy or the DVS algorithm. Instead, we use separate modules that provide the real-time scheduling policy and the RT-DVS algorithms. One such RT scheduler/DVS module can be loaded on the system at a time. By separating the underly- ing periodic RT support from the scheduling and DVS policies, this architecture allows dynamic switching of these latter policies with- out shutting down the system or the running RT tasks. (Of course, during the switch-over time between these policy modules, a real- time scheduler is not defined, and the timeliness constraints of any running RT tasks may not be met). The last kernel module in our implementation handles the access to the PowerNow! mechanism to adjust clock speed and voltage. This provides a clean, high-level interface for setting the appropriate bits of the processor’s special feature register for any desired frequency and voltage level. The modules provide an interface to user-level programs through the Linux /procfs filesystem. Tasks can use ordinary file read and write mechanisms to interact with our modules. In particular, a task can write its required period and maximum computing bound to our module, and it will be made into a periodic real-time task that will be released periodically, scheduled according to the currently- loaded policy module, and will receive static priority over non-RT tasks on the system. The task also uses writes to indicate the com- pletion of each invocation, at which time it will be blocked until the next release time. As long as the task keeps the file handle open, it will be registered as a real-time task with our kernel extensions. Al- though this high-level, filesystem interface is not as efficient as di- rect system calls, it is convenient in this prototype implementation, since we can simply use cat to read from our modules and ob- tain status information in a human readable form. The PowerNow! module also provides a /procfs interface. This will allow for a user-level, non-RT DVS demon, implementing algorithms found in other DVS literature, or to manually deal with operating frequency and voltage through simple Unix shell commands.
operation costs triple of low peaks operation). 2. CONCEPT OF POWERLOAD MANAGEMENT The load management is a new concept of distribution of electricity aiming at a more efficacious supply network system. Such a control system should satisfy the needs of consumers at the lowest possible peak loading. There is a strong upward tendency in using load management throughout the world. The direct load management systems have passed through the experimental stage and have been now adapted as an everyday practice is a great many of supply network systems. The economic grounds for introduction of these systems have been justified throughout the world and also certain. The load management is a process going along with electricity conservation which decreases total electricity consumption, while the load management is intended for consumption control over a certain period of time . Load management is defined as sets of objectives designed to control and modifies the patterns of demands of various consumers of a power utility. This control and modification enables the supply system to meet the demand at all times in most economic manner.
G. K. Morison, B. Gao, P. Kundar in  has showed how voltage instability can occur and the situations in which the modeling of loads, under load tap changers and generator maximum excitation limiters will impact the system voltage stability. Modeling of loads has significant effect on the accuracy of voltage stability analysis [23-24] and investigated the dynamic nature of voltage instability considering dynamic load. To study dynamic voltage stability of a system we need to consider dynamic model for all the elements in the power system  and capture all the dynamics of different elements in the system to find out the exact reason for voltage collapse. In  voltage instability is associated with tap-changing transformer dynamics by defining the voltage stability region in terms of allowable transformer settings. In  they have employed a nonlinear dynamic model of OLTC, impedance loads and decoupled reactive power voltage relations to reconstruct the voltage collapse phenomenon and developed a method to construct stability regions. In  analyzed dynamic phenomenon of voltage collapse by dynamic simulations using induction motor models has been analyzed and it explains how voltage collapse starts locally at weakest node and spreads out to the other weak nodes.
A concrete approach offering the least amount and finest location of load shedding was presented in . The proposed technique uses a multi-stage and non-linear approach to find the minimum load shed at each stage. Genetic Algorithms were executed in the Hydro-Quebec system to estimate the amount of load shed in[10, 11],but the approach is unable to grip a broader range of load behavior, different scenarios and short-term voltage instability problems. GA was utilized to investigatefor optimal supply restoration approach in the network of distribution system.Likeisealternativestudy showedan optimization tool builton GA to estimate and perform load shed. To solve steady State load shedding problem a novel application of the GA presented in . A new adaptive load shedding technique using GA is proposed in . The load buses are ranked from the strongest to the weakest. The weakest bus is considered the best option for load shedding. The voltage stability margin is highly influence by the weakest buses in an interconnected power system. Therefore, the identification of weak buses is necessary for planning and operation of powersystems in long-term studies.
Abstract— Electric Vehicles (EV) have limited air pollution and are many environment friendly, and due to their addition to carbon dioxide reduction, EVs are enchancing increasingly popular nowadays. The government also encourage and supporting the procedure of electric vehicles for the social. The electric vehicle - taxis have been discovered into the common transportation systems to increase EV market distribution. Various from regular taxis that can refuel in minutes, EV taxis’ recharging cycles can be as lengthy as one hour. Due to the lengthy cycle, the poor decision on the charging station, i.e., choosing one without hollow charging piles, may lead to a lengthy waiting time of more than an hour in the bad case. Therefore, choosing the right charging station is very necessary to reduce the overall waiting time. Considering that the waiting time can be a non negligible portion to the mistken work hours, the decision will naturally distrub the revenue of individual EV taxis. The current practice of a taxi driver is to choose a station heuristically without a global knowledge. However the heuristically choice can be a wrong one that leads to more waiting time. The proposed system provides a real-time charging station recommendation system for EV taxis via large-scale GPS text mining. By combining each EV taxi’s historical recharging events and real-time GPS trajectories, the data operational state of each taxi is predicted. Based on this data, for an EV taxi requesting a recommendation, recommend a charging station that
Plug-in hybrid electric vehicles (PHEVs) have become popular in recent years since they have advantages of low fuel cost and low carbon emissions  compared to conventional vehicles that use internal combustion engines. The performance of a PHEV depends highly on the on-board battery energy storage system (BESS). The choice of the BESS technology for an PHEV will depend on vehicle power and energy ratings, response time, weight, volume and operating temperature . Common BESS technologies for PHEV applications include lead-acid (Pb-Acid), lithium-ion (Li-ion), nickel-cadmium (Ni-Cd), nickel-metal hydride (Ni-MH)  and sodium-sulphur (NaS) . Among these battery technologies, Li-ion batteries have superior performance over other battery technologies and are preferably used for PHEVs applications. This is mainly attributed to several factors, like high energy density, high output current levels, flat discharge characteristics in the middle region, high energy-to-weight ratios, no memory effect, and a low self-discharge rate , . Fig. 5.1 shows the Li-ion battery packs for a 2013 Chevrolet Volt PHEV.
With the exponential growth of distributed energy resources, the MG is becoming a multidirectional network interconnecting a vast range of consuming devices. As an effective approach to increase reliability and decrease dependence on energy storage systems, larger scale MGs interconnecting more DGs are recommended, since by extension of power sharing between DGs, it would be possible alleviation of the problems created due to unreliable renewable energy resources. Accordingly, PQ issue of such a wide system needs to be attended more since the PQ situation of a certain point (or area) might significantly affect the other points (or areas) PQ level due to harmonic propagation , . Few literature works address the PQ of larger scale powersystems –. For example, in , a double-resistive APF is proposed to attenuate voltage harmonics and suppress harmonic propagation in radial powersystems respectively installed at the end feeder and at a specific place determined based on PQ concerns. The proposed method in  is not applicable in interconnected powersystems. In , a long-feeder simulator-based APF is proposed to reduce harmonic propagation in closed-loop distribution feeders. The proposed APF is recommended to be installed at the middle electrical point of the system. Likewise, the proposed method in  cannot be used in interconnected powersystems. In , an optimal utilization of DGs inverters (as compensators) is proposed. Taking into account the available capacity of each inverter, to utilize inverters toward harmonic suppression as
To address the design and control challenges of DP vessels, a hybrid propulsion system using batteries is known as a feasible solution due to its potential benefits. First, the utilisation of batteries allows an optimal control of energy flow between engine-generators and batteries. Depending on load demands and battery energy, one or more engines to be turned off and on to avoid operating in low efficiency zones. This results in significant improvements in fuel consumption, emissions and noise-vibration-hardness level of the engines. Second, similar to automotive applications (Son et al., 2018) the use of batteries offers the load levelling function in energy management. By using this function, engine-generators can work at their optimal constant working points with the highest energy efficiency disregarding the actual load fluctuations by dis-charging or re-charging batteries. Third, batteries can provide other advanced features, such as peak shaving (to avoid the need of large power sources to supply peak demands of highly variable pulsed loads, i.e. naval weapons and high-power radar) (Dedes et al., 2012), brake energy recuperation and back-up power (Geertsma et al., 2017). In order to enhance these functions, an agile control scheme is necessary to decide working profiles of engine-generators and batteries in such a way that the fuel consumption and emissions of the engines are minimised while the machine performance is guaranteed (Sun, 2015).
The operation objectives of the LFC are to maintain uniform frequency, to divide the load between generators and to control the tie line interchange schedules. The change in frequency and tie line power are sensed, which is the measure of the change in rotor angle . The error signal, and , are amplified, mixed and transformed into a realpower command signal , which is sent to the prime mover to call for an increment in torque. The prime mover, therefore brings change in the generator output by an amount which will change the values of and within the specified tolerance.
provided by the motor, then sends an appropriate signal to the motor controller. The EPS motor is thus controlled to provide more assist at low speeds and less assist at high speeds to provide better road feel to the driver. The EPS motor provides the necessary torque through the gearbox to the steering rack and pinion which steers the vehicle wheels accordingly. When a fault occurs with the torsion bar torque sensor, the EPS controller assumes no driver torque signal is sent and does not power the assist motor. This turns the EPS system to manual steering requiring the driver to provide the entire steering torque. This is an undesirable situation. To make the EPS system more reliable, a fault tolerant control using motor load torque observer is designed.
partitioning in the second step instead of exhaustive enumeration to make the algorithm takes time polynomial in the number of tasks, and polynomial in the number of processors. In  authors have compared 11 heuristics for mapping a set of independent tasks onto heterogeneous distributed computing systems. The best one that has minimum makespan, that is defined as the maximum completion time for the whole processors, was the Genetic Algorithm (GA) followed by Min-min algorithm. In , Chen and Cheng (2005) applied the Ant Colony Optimization (ACO) algorithm. They proved that ACO outperforms both GA and LP-based approaches in terms of obtaining feasible solutions as well as processing time. Abdelhalim  presented a modified algorithm based on the Particle Swarm Optimization (PSO) for solving this problem and showed that his approach outperforms the major existing methods such as GA and ACO methods. Then, his PSO approach is developed to can further optimize the solution to reduce the energy consumption by minimizing average utilization of processors (without using any energy or power model). Finally, a tradeoff between minimizing the design makespan as well as energy consumption is obtained. In  Visalakshi and Sivanandam, presented a hybrid PSO method for solving the task assignment problem. Their algorithm has been developed to dynamically schedule heterogeneous tasks onto heterogeneous processors in a distributed setup. It considers load balancing and handles independent non-preemptive tasks. The hybrid PSO yields a better result than the normal PSO when applied to the task assignment problem. The results are also compared with GA. The results infer that the PSO performs better than the GA. In , Omidi and Rahmani used PSO for task scheduling in multiprocessor systems as an important step for efficient utilization of resources. They considered independent tasks on homogeneous multiprocessor systems. Apart from all these efforts, this paper integrates the PSO approach with a polynomial-time partitioning techniques; Min-min and priority assignment. The proposed approach takes into account energy efficiency during task partitioning among heterogeneous cores in MPSoCs.
Data are used to setup a set of training data of 1400 patterns. Inputs and outputs are normalized to range from -1 to 1. After the ANN model is trained, the 10 hours forecast of electricload are calculated from the model and denomalized and then compared with the realtimeelectricload. Adaptive Learning Factor
As shown in Table 7, in stage I, the system load is at its valley, industrial loads which are price-sensitive are generally full load operation, it can increase electrical load of public facilities and resident by means of Auto DR, such as increasing the cooling rate of the refrigera- tion/freezer or the storage capacity of resident. In stage II, the wind power output decreases, before eight o’clock, when system load is low, it can be balanced by control- ling the load which have been adjusted in stage I and with interruptible load of 8 hours; after eight o’clock, when system load is at its morning peak with the load of resident and public facilities higher electricpower, it can be balanced by adding a interruptible load of 4 hours of industry; during this time it will be fluctuant, so it needs to supply some Auto DR. In stage III and stage IV, wind power output has a positive peaking characteristic, the load of resident and public facilities can modify load rate by means of Auto DR.
When the bus voltage is controlled concurrently by a conventional SVC control and ULTC, both having completely different time-response scales, the SVC reacts to the voltage deviations before ULTC. consequently, when the SVC output reaches the maximum capacity limit during the steady-state voltage variation, the SVC loses its functionality during the dynamic voltage variation. The newly proposed SVC control, against the conventional SVC control having a different time response than ULTC, would improve the coordination with ULTC by adapting the following settings. 1) The SVC reference voltage has to be equal to the ULTC reference voltage.
Approaches for compensation for data loss over network links have also been proposed, among others, by Nilsson  and by Ling and Lemmon , who posed the problem of optimal compensator design for the case when data loss was independent and identically distributed (iid) as a nonlinear optimization. A sub-optimal estimator and regulator to minimize a quadratic cost were proposed by Azimi Sadjadi  and this approach was extended by Imer et al. in  and Sinopoli et al. in [10-11]. However, most of the work aimed at designing a packet loss compensator. The compensator used the successfully transmitted packets to come up with an estimate of the plant state. This estimate is then used by the controller. Gupta et al. realized that sensors and actuators equipped with wireless or network communication capabilities would likely have some computational power. They took these computational capabilities into consideration and extended the above result further. They introduced an encoder at the sensor end. The compensator then became a decoder for the information being transmitted over the unreliable link. The control, the encoder and the decoder were jointly designed to solve the packet loss problem and to achieve an LQG optimal control.
The popularity of handheld devices, which are usually powered by batteries, has made power saving an important and practical issue in recent years. Techniques of power saving for user devices using mobile communication systems such as WiMAX and LTE (Long-Term Evolution) are parts of the major focuses in the literature. In this paper, two revised schemes of the authors ’ previously proposed power saving schemes for IEEE 802.16 are proposed to be applied in LTE. The proposed schemes, namely LTE-LBPS-Aggr and LTE-LBPS-Merge, estimate the input load by traffic measurement and the channel capacity by channel quality indicator (CQI) reports, calculate the length of the sleep cycle, and notify related user equipments (UEs) of the next radio-on time for receiving data. The difference between LTE-LBPS-Aggr and LTE-LBPS-Merge lies in the grouping of UEs for sleep scheduling. LTE-LBPS-Aggr treats all UEs in a group, while LTE-LBPS-Merge allows multiple groups of UEs in sleep scheduling. The simulation study shows that in comparison with standard-based mechanisms, the proposed schemes can achieve better power saving efficiency at the cost of moderate increase on delays and the signaling overhead.