Synchronization in the AC-stacked PV-inverter topology has been shown utilizing powerline communication where a low-bandwidth synchronization signal is transmitted from the grid interface to a receiver in each inverter allowing the controls to obtain the necessary phase alignment to the grid voltage that allows for them to operate completely autonomously. The authors have been developing a channel model to describe the transmission and reception of this synchronization signal as it propagates down the powerline and the series connected power-electronic devices. This paper builds upon this model and examines the effects of the noise in the channel as it affects the reliability of the receiver to accurately determine if the synchronization signal is correct. Such reliability is key to the decentralized control of the AC-stacked PV-inverter topology. With the channel model that can determine the amount of transmitted current and the noise in the system, a power optimized transmission can be established that remains in an acceptable operating limit for reliability. Even though the results are for the AC-stacked PV-inverter architecture, the model is generalized enough for applicable for any series-connected power electronics.

This paper proposes a cost-effective approach to active power-factor correction that uses a smart meter installed at a service entry and a dedicated active power filter. The power circuit could be connected at a conventional 240V AC outlet in a home, or it could be installed at a transformer secondary and coordinate with several downstream meters. The smart meter is a non-intrusive load monitor (NILM) that measures the active, reactive, and harmonic content of the line current. This information is fed to the dedicated active power filter. The NILM determines the amount of reactive and harmonic current fed to the downstream loads, which may include the current supplied by the active power filter itself. Despite this, the reactive current can be reduced to nearly zero. The active power filter can also be provided with external commands from the grid operator. This scheme is particularly useful in virtual power plant (VPP) applications involving demand dispatch in support of distributed intermittent renewables. The paper describes the control scheme needed by the power circuit and presents experimental results.

This paper describes how smart meters can be used to for automatic diagnostics and control in HVAC systems. The specific smart meter used in this paper is the non-intrusive load monitor (NILM). The NILM can disaggregate the operating schedule of individual loads from aggregate current measurements. Combining load operating schedules with measurements from several environmental sensors, the NILM builds a representative thermal model for the building. Two different model formulations are proposed. The first is designed to detect refrigerant leaks in existing thermostatically controlled HVAC systems; the other is intended for use with next-generation variable-capacity systems. Experimental and simulation results are presented.

This paper describes the design and deployment of a large scale wireless sensor network (WSN) for monitoring the health of power equipment in a substation. The sensor network consists of 122 low power nodes that that are spread over an area approximately 1000 × 400 feet in size and perform monitoring of equipment such as transformers, circuit breakers, and compressors. All nodes communicate over a multihop wireless mesh network that uses a dynamic link-quality based routing protocol. A primary objective of this project is to develop effective monitoring applications for the substation using low-cost wireless sensor nodes that can sustain long periods of battery life. We study the battery consumption in the network and present a transmission scheme that conserves communication cost by enabling the sensor nodes to transmit observation samples only when their values are significantly different from those transmitted previously. Experimental results that demonstrate the performance of the sensor network for several monitoring applications are presented.

We present the design and performance of a wireless sensor network that is deployed in a substation for monitoring the health of power subsystems such as circuit breakers, transformers and transformer bushings. The sensor network consists of 45 low power nodes for temperature sensing and a high power node that is capable of performing more specialized computations. All nodes communicate over a multihop mesh network that uses a dynamic link-quality based routing protocol. The nodes are solar powered and are capable of operating autonomously for multiple years.

On modern computers, the performance of programs is often limited by memory latency rather than by processor cycle time. To reduce the impact of memory latency, the restructuring compiler community has developed localityenhancing program transformations, the most well-known of whichisloop tiling. Tiling is restricted to perfectly nested loops, but many imperfectly nested loops can be transformed into perfectly nested loops that can then be tiled. Recently, we proposed an alternative approach to locality enhancement called data shackling. Data shackling reasons about data traversals rather than iteration space traversals, and can be applied directly to imperfectly nested loops. We have implemented shackling in the SGI MIPSPro compiler which already has a sophisticated implementation of tiling. Our experiments on the SGI Octane workstation with dense numerical linear algebra programs show that shackled code obtains double the performance of tiled code for most of these programs, and obtains five times the performance of tiled code for some versions of Cholesky factorization. Data shackling has been integrated into the SGI MIPSPro compiler product-line.

Cache-coherent multiprocessors with distributed shared memory are becoming increasingly popular for parallel computing. However, obtaining high performance on these machines mquires that an application execute with good data locality. In addition to making efiective use of caches, it is often necessary to distribute data structures across the local memories of the processing nodes, thereby reducing the latency of cache misses.We have designed a set of abstractions for performing data distribution in the context of explicitly parallel programs and implemented them within the SGI MIPSpro compiler system. Our system incorporates many unique features to enhance both programmability and performance. We address the former by providing a very simple programmming model with extensive support for error detection. Regarding performance, we carefully design the user abstractions with the underlying compiler optimizations in mind, we incorporate several optimization techniques to generate efficient code for accessing distributed data, and we provide a tight integration of these techniques with other optimizations within the compiler Our initial experience suggests that the directives are easy to use and can yield substantial performance gains, in some cases by as much as a factor of 3 over the same codes without distribution.