Zero Output Research Articles

Quantitative assessment methods and qualitative evaluation system for the complementary characteristics of multi-energy

Assessing and analyzing the complementary characteristics of renewable energy (RE) is crucial for designing, operating, and optimizing multi-energy complementary systems (MECSs). However, unified and precise quantitative descriptions of the complementary and stability characteristics among various energy outputs in MECSs have lacked attention and research. Here, this study innovatively proposed a mathematical model for the multi-energy complementarity index (MECI), which considers the complementarity rates of multiple energy outputs during zero and non-zero output periods, and a mathematical model for the multi-energy volatility index (MEVI), which accounts for fluctuation thresholds and the overall volatility of output processes. An evaluation system for multi-energy complementarity characteristics qualitative analysis has been established. The natural output processes of RE at three MECSs in China were applied in the case calculations and verification. Results show that the hydropower rated discharge (Qrating) has a significant negative correlation with MECI, with the MECI decreasing by an average of 0.0046 for every 5 m³/s increase in Qrating. The relationship between the Qrating and MEVI shows an overall negative correlation with local fluctuations. Notably, The MECI of the BeiPan River MECSs exhibits significant seasonal characteristics, with the MEVI in summer (0.378) and autumn (0.395) higher than those in spring (0.132) and winter (0.160), closely related to the natural seasonal variations of the three energy sources: water, wind, and solar. We believe that the study can assist in evaluating and making decisions on the multi-energy complementarity characteristics of RE bases in the future, making a significant contribution to achieving dual carbon goals.

Mixture-of-Rookies: Saving DNN computations by predicting ReLU outputs

Deep Neural Networks (DNNs) are widely used in many application domains. However, they require a vast amount of computations and memory accesses to deliver outstanding accuracy. In this paper, we propose a scheme to predict whether the output of each ReLu activated neuron will be a zero or a positive number in order to skip the computation of those neurons that will likely output a zero. Our predictor, named Mixture-of-Rookies, combines two inexpensive components. The first one exploits the high linear correlation between binarized (1-bit) and full-precision (8-bit) dot products, whereas the second component clusters together neurons that tend to output zero at the same time. We propose a novel clustering scheme based on analysis of angles, as the sign of the dot product of two vectors depends on the cosine of the angle between them. We implement our hybrid zero output predictor on top of a state-of-the-art DNN accelerator. Experimental results show that our scheme introduces a small area overhead of 5.3% while achieving a speedup of 1.2x and reducing energy consumption by 16.5% on average for a set of diverse DNNs.

Optimal control strategy by tracking and compensation for ICPS under the DoS attack

AbstractTo solve the security problem of industrial cyber‐physical systems (ICPS) suffering from the Denial of Service (DoS) attack, this paper proposes an optimal control strategy with functions of tracking and compensation control. The system model, being depicted as a discrete linear time‐invariant system combined with an improved Bernoulli model, is established to describe the system under DoS attacks by using the feedback control theory. The optimal control strategy is completed by designing linear quadratic Gaussian (LQG) controller consisting of a tracking regulator and a Kalman filter to achieve tracking control of ICPS whose equilibrium state is neither zero state nor zero output, where a periodic event‐triggered mechanism is adopted for ensuring the system's stability. The compensation control strategy is completed by designing a compensator to compensate the system to a certain extent through the multistep state prediction compensation strategy after a DoS attack being detected, where the detector is adopted to detect the DoS attack based on the measurement residual obtained by the Kalman filter. The simulation test is conducted by using Simulink/True Time tool with a ball‐beam system as the physical object, whose results show that the tracking regulator can achieve the tracking control of ICPS; compensation strategy can reduce the impact of the attack on the system to a certain extent.

SNG production with net zero outflow of CO2 in an integrated energy system: An energy and economic aspects

Integrated energy systems are getting high attention due to their synergistic effects on both the energy efficiency and process economy. This study demonstrates the similar effect of coupling four processes such as solid oxide electrolyzer cell (SOEC)-based hydrogen production from renewable energy, synthetic natural gas (SNG) production, steam Rankine cycle (SRC) for energy recovery, and amine-based SNG upgrading. Integration of two SRCs with SNG produces 1.8 MW of power that is assumed to reduce the load on SOEC. In addition, monoethanol amine (MEA)-based chemical absorption provides two benefits comprising SNG upgrading and capture of unreacted CO2. The proposed process integration claims the net zero output of CO2 in purified SNG stream as the captured CO2 is recycled back to the SNG reaction system for process enhancement. Conclusively, proposed integrated process synergistically enhances the SNG production efficiency by up to 86 % while the whole integrated process efficiency is calculated at 77.3 %. Corresponding to process economy, unit production price of SNG is observed to be 0.223 $/kWh, which is comparable to other existing processes. Moreover, methane parity analysis and future predictions are made with expected natural gas price to estimate the possible reliance on SNG in coming decades.

Design of Measuring and Controlling Circuit for Digital Output MEMS Gyroscope with Zero Output Correction

Design of Measuring and Controlling Circuit for Digital Output MEMS Gyroscope with Zero Output Correction

An integrated slacks-based super-efficiency measure in the presence of nonpositive data

The slacks-based measure (SBM) and super-efficiency SBM (S-SBM) models are inoperable when inputs or outputs have zero or negative data. In this paper, a mixed-binary linear program integrating SBM and S-SBM is developed to handle zero and negative input and output values. The model is units invariant, unlike Tone and Chang et al. (2020)’s BP-SBM model. The model also addresses the limitations of modified SBM (MSBM) (Sharp et al., 2007) by dropping the requirement for minimum inputs and maximum outputs and avoids the sensitivity of the models using a range adjusted measure (RAM) to the range of the maximum and minimum values. Three sets of data are used to test the model and compare it with BP-SBM and the model proposed by Lin et al. (2019). The first data set is a simple illustration to simplify understanding of the data translation and application of the three models. The second data set comprises 30 Taiwanese electrical machinery listed firms and contains negative outputs. The third data set is the energy activities of 30 Chinese provinces with two bad outputs that are treated as negative data. The kernel density estimations show the distributions of efficiency scores as well as the Simar–Zelenyuk adapted Li tests for the model comparisons of the bootstrapped distributions for the latter two data sets. The results of these data sets validate the usefulness and practicability of our proposed approach.

Zero-Aware Low-Precision RNS Scaling Scheme

Scaling is one of the complex operations in the Residue Number System (RNS). This operation is necessary for RNS-based implementations of deep neural networks (DNNs) to prevent overflow. However, the state-of-the-art RNS scalers for special moduli sets consider the 2k modulo as the scaling factor, which results in a high-precision output with a high area and delay. Therefore, low-precision scaling based on multi-moduli scaling factors should be used to improve performance. However, low-precision scaling for numbers less than the scale factor results in zero output, which makes the subsequent operation result faulty. This paper first presents the formulation and hardware architecture of low-precision RNS scaling for four-moduli sets using new Chinese remainder theorem 2 (New CRT-II) based on a two-moduli scaling factor. Next, the low-precision scaler circuits are reused to achieve a high-precision scaler with the minimum overhead. Therefore, the proposed scaler can detect the zero output after low-precision scaling and then transform low-precision scaled residues to high precision to prevent zero output when the input number is not zero.

Optimizing nonlinear activation function for convolutional neural networks

Activation functions play a critical role in the training and performance of the deep convolutional neural networks. Currently, the rectified linear unit (ReLU) is the most commonly used activation function for the deep CNNs. ReLU is a piecewise linear function that will output the input directly if it is positive, otherwise, it will output zero. In this work, we propose a novel approach to generalize the ReLU activation function using multiple learnable slope parameters. These learnable slope parameters are optimized for every channel, which leads to the learning of a more generalized activation function (a variant of ReLU) corresponding to each channel. This activation is named as fully parametric rectified linear unit (FReLU) and trained using an alternate optimization technique by learning one set of parameters, keeping another set of parameters frozen. Our experiments show that the method outperforms ReLU and its other variant activation functions and also generalizes over various tasks such as image classification, object detection and action recognition in videos. The Top-1 classification accuracy of FReLU on ImageNet improves by 3.75% for MobileNet and $$\sim $$ 2% for ResNet-50 over ReLU. We also provide various analyses for better interpretability of our proposed activation function.

Application Study on Zero Output Technology of Low Pressure Cylinder of 350MW Supercritical Unit

In response to the requirement of national thermal power flexibility reform policy and the need of regional central heating development, a 350MW supercritical unit in a power plant has been transformed with zero output technology of low pressure cylinder. In this paper, the scope and content of low pressure cylinder zero output transformation are discussed, peak shaving and heating capacity after transformation are analyzed, the difference between operation effect and calculation result after unit transformation is compared by test, and the matters needing attention in future operation are given, which provides reference for technical transformation of low pressure cylinder zero output of other power plant units.

Output Monomial Reachability and Zero Output Controllability of Positive Switched Systems

In this paper, we present a sufficient condition for the output reachability of discrete-time positive switched systems. Besides, necessary and sufficient conditions for output monomial reachability and zero output controllability are provided. Further, some examples are shown.

AC-Voltage Harmonic Control for Stand-Alone and Weak-Grid-Tied Converter

This article presents an ac voltage controller that can operate under a wide range of loads, thanks to its high robustness to plant-model variations. A robust voltage controller is required in a droop-controlled inverter that operates connected to a grid or in island mode with other generators. In such modes of operation, the voltage source converter can experience large load-impedance variations depending on the number of parallel generators or the grid impedance value. Concerning the controller performance, the transient response is improved by selectively minimizing the output impedance of the system at the frequencies that typically contain the largest components of the output current. Complete harmonic control is achieved due to its zero output impedance at a set of design-selected frequencies. The proposed controller only measures the output voltage, and it contains a single-loop structure that uses all available bandwidth. In addition, a fast converter-current estimator provides protection against faults and inrush output currents. The design process only requires, as input parameters, the LC filter values, the sampling frequency, and a set of harmonics where load-current circulation is expected.

Quadratic Buck–Boost Converter With Zero Output Voltage Ripple at a Selectable Operating Point

We introduce a new dc–dc converter topology with the following features: first, quadratic voltage gain, which allows the converter to work over a wide voltage range with a minimal variation of duty cycle and without the use of extreme duty cycles; second, output voltage ripple mitigation, which can be accomplished with respect to a fixed but otherwise arbitrary duty cycle by proper selection of components and switching signals. Since the proposed converter can be designed to have zero output voltage ripple for the nominal gain (duty cycle), this allows the use of small values of capacitances. The underlying principles and steady-state equations in continuous conduction mode are presented in detail. Experimental tests are presented to validate the analysis.

(Plenary) Challenges in Going from Laboratory to Megawatt Scale PEM Electrolysis

Renewable hydrogen is becoming an increasingly important component of the transition away from fossil fuel use and towards reduction in carbon dioxide production. Hydrogen is the intermediary between primary energy sources and end products in many chemical processes such as ammonia generation, refining, and biogas processing, and is currently mainly produced by reforming of natural gas. Hydrogen from electrolysis can both make a strong environmental impact on these industries and also improve utilization of intermittent renewable energy sources such as wind and solar by leveraging otherwise stranded resources. Hydrogen can also enable improved integration of the electrical grid with the transportation sector and industrial processes, by serving as a storage medium until needed (Figure 1). Because of hydrogen's end use flexibility, electrolysis is a key component of Europe's strategy for energy storage and management of these renewable energy sources on the electrical grid. Proton exchange membrane (PEM) electrolysis is especially well suited to energy capture because of the dynamic range and ability to quickly ramp up and down from near zero output to full capacity. PEM electrolysis was largely developed in the 70’s by GE for life support applications where reliability and reproducibility were key. As the technology was translated to commercial markets it started to compete with the incumbent alkaline water electrolysis on cost, safety (no mixing of O2 and H2), ease of maintenance (no caustic) and operational flexibility (differential pressure operation, idling). While the resulting commercial success of PEM electrolysis suggests that the current state of the technology is sufficiently mature, the truth is that while PEM electrolyzers have demonstrated > 50,000 hrs lifetimes and > 99% reliability in the field, the technology and manufacturing techniques have not changed significantly from those early days. In traditional industrial applications the customer rarely cares about efficiency as much as reliability. At the same time, the robustness and manufacturability of the technology (cell, stack and system engineering), have allowed the systems to be scaled from laboratory scale (0.01kg/day) to megawatt scale (100kg/day) with no loss in performance or reliability. To date the research community has mostly focused on advancing PEM fuel cell technology, producing staggering PGM reduction, improvement in efficiency and translation to high throughput manufacturing. Thus electrolysis technology development has lagged behind PEM fuel cells, providing significant opportunities for continued improvements. Thinner membranes, reduced PGM usage, and improved manufacturing techniques have all demonstrated viability for electrolysis, and have the potential to make higher cost and performance impact. As the application areas for PEM electrolysis start to shift from traditional markets to fueling and renewable energy capture, efficiency, capital cost and ethical resource use become significant factors. Thus advancing the state of the art, translating laboratory advances to commercial PEM stacks and then manufacturing them at scale, while maintaining reliability becomes paramount. Questions that arise are: How do you test reliability, without testing for 50,000 hrs? How do you pick research “winners” at the rotating disk electrode (RDE) level and develop them to large scale PEM membrane electrode assemblies (MEAs). Where is the gap in research at the laboratory scale based on commercial failures? This talk will discuss the challenges in continued scale up, translating laboratory scale findings to commercial systems as well as some of recent advancements and impact on cost. Figure 1

(Plenary) Challenges in Going from Laboratory to Megawatt Scale PEM Electrolysis

Renewable hydrogen is becoming an increasingly important component of the transition away from fossil fuel use and towards reduction in carbon dioxide production. Hydrogen is the intermediary between primary energy sources and end products in many chemical processes such as ammonia generation, refining, and biogas processing, and is currently mainly produced by reforming of natural gas. Hydrogen from electrolysis can both make a strong environmental impact on these industries and also improve utilization of intermittent renewable energy sources such as wind and solar by leveraging otherwise stranded resources. Proton exchange membrane (PEM) electrolysis is especially well suited to energy capture because of the dynamic range and ability to quickly ramp up and down from near zero output to full capacity. This paper will discuss the challenges in continued scale up, translating laboratory scale findings to commercial PEM systems as well as some of recent advancements and impact on cost.

Renewable Hydrogen: Near Term Solutions and Long Term Potential

Renewable hydrogen is becoming an increasingly important component of the transition away from fossil fuel use and towads reduction in carbon dioxide production. Hydrogen is the intermediary between primary energy sources and end products in many chemical processes such as ammonia generation, refining, and biogas processing, and is currently mainly produced by reforming of natural gas. Hydrogen from electrolysis can both make a strong environmental impact on these industries and also improve utilization of intermittent renewable energy sources such as wind and solar by leveraging otherwise stranded resources. Because of hydrogen's end use flexibility, electrolysis is a key component of Europe's strategy for energy storage and management of these renewable energy sources on the electrical grid. Proton exchange membrane (PEM) electrolysis is especially well suited to energy capture because of the dynamic range and ability to quickly ramp up and down from near zero output to full capacity. PEM electrolysis technology development has lagged behind PEM fuel cells, providing significant opportunities for continued improvements. Thinner membranes, reduced PGM usage, and improved manufacturing techniques have all demonstrated viability for electrolysis, and have the potential to make higher cost and performance impact. At the same time, the basic technology is mature and has been scaled to multi-megawatt capacity. This talk will discuss the challenges in continued scale up, as well as some of recent advancements and impact on cost. A potential roadmap for next generation technology evolution will also be presented. There has also been significant interest in alkaline exchange membranes and ionomers (AEMs) for electrolysis applications. The AEMs are growing in stability, and electrolysis offers an alternative outlet for these materials which has operational advantages vs. the fuel cell. While electrolysis occurs at higher voltages and higher pressure, there are other operational characteristics which are less challenging than the AEM fuel cell application, such as water management within the membrane and through the cell. A comparison of traditional liquid alkaline, PEM electrolysis, and the potential for AEM electrolysis will be discussed.

Improving the Regulation Range of EV Battery Chargers With <italic>L3C2</italic> Resonant Converters

The demand for electric vehicles has expanded rapidly for both industrial and transportation applications. In parallel, new battery technologies capable of deeply discharging and powering electric vehicles over long periods of time have been introduced and made available in the market. Due to the increasing complexity of charging algorithms, battery chargers are exposed to demanding operating requirements. Chargers should not only work under different loading conditions (from absolutely zero to maximum output power) but also regulate a wide output voltage range (from near zero to 1.5 times nominal voltage). In this paper, a multiresonant $L3C2$ resonant converter is introduced that can cover nearly all of the regions in the battery V–I plane, from near zero output voltage, zero output current to maximum output power. By adding one capacitor, the topology is able to extend the operating frequency beyond the ${{{\bm LLC}}}$ topology and achieve significant regulation improvements. Near free-ripple charging current for batteries charging is achieved in different states of charge without using burst mode operation, effectively increasing the battery life cycle. In addition, soft transitions are obtained for all the switches (MOSFETs and diodes) resulting in high efficiency, reliability, power density, and low-noise operation of the charger. A complete set of simulation and experimental results, extracted from a 96-VDC 950-W $L3C2$ resonant converter, demonstrates the characteristic features of the proposed topology for battery charger applications, while providing a comparative example with the ${\bm LLC}$ topology.

(Invited) Key Barriers to Commercial Deployment of Next Generation Proton Exchange Membrane Electrolyzers

Hydrogen is gaining increased international attention based on the imminent roll out of fuel cell vehicles, as well as the growing need for energy storage to manage high levels of renewable and inherently intermittent energy sources on the electrical grid. Hydrogen from electrolysis based on proton exchange membrane (PEM) technology in particular is of interest due to the lack of carbon footprint, scalability, ability to ramp from near zero output to 100% rated capacity or greater, and the lack of corrosive electrolyte. While electrolysis and fuel cells are not typically as efficient on an electrical round trip basis vs. many battery systems, hydrogen is highly deployable, providing flexibility to optimize the grid system. Hydrogen can be used to crosslink different infrastructures such as the electrical grid, the transportation grid, and chemical processing. For example, hydrogen produced using excess renewable energy can either be fed back to the electrical grid via fuel cells, used to fuel hydrogen-powered vehicles, or used to upgrade conversion of biogas to methane. PEM electrolysis has the technical maturity and reliability to serve these applications, but is still handicapped by legacy materials and manufacturing processes vs. the next generation technology advancements that have been made in fuel cells. Feasibility of translating many of the elements of current fuel cell technology to electrolyzer systems has been demonstrated, but the specific requirements of electrolysis cells provide additional technical gaps which prevent materials developed for fuel cells to drop seamlessly into electrolysis processes. Some customization and optimization of materials and processes are needed to enable implementation in electrolysis cells. There are also areas of development needed including understanding the impact of process scale up on resulting material properties, translation from machined proof of concept prototypes to parts fabricated by high speed manufacturing methods, and integration of research efforts on different components into a cohesive cell design. In addition, scale to larger cell sizes brings new fundamental questions and materials and design challenges. These issues are often masked by the typical metrics and models currently in use for electrolysis, such as the DOE H2A model. The H2A model is a very detailed tool which provides relative comparisons between technologies based on constant assumptions, but does not reflect present commercial production, partially because this information is typically considered proprietary by companies. The H2A model is instead based on technology which has been demonstrated at a high enough technical fidelity to prove feasibility of commercialization within 5 years. It does not take into account the often challenging scale up and manufacturing development, or the investment required for the remaining hurdles to commercialization. It also assumes production of 500 units/year at 1500 kg/day hydrogen output per unit, or almost 2000 MW of annual capacity. No commercial company has reached even 5% of this total annual hydrogen output in sales, and thus large extrapolations are being made at both a manufacturing readiness level and volume production level in the model. Electrolysis also does not currently have the same well-defined material targets and metrics as fuel cells, such as hydrogen and oxygen crossover, platinum group metal loading, electrocatalyst support stability, and bipolar plate resistivity and corrosion resistance. These quantitative metrics would help to highlight specific progress being made in these areas vs. the overall system metrics which dampen the impact of a single component. Balance of plant is also an important contributor to cost, efficiency, and reliability which is not typically addressed in detail in government technology roadmaps. This talk will present the limitations in commercial PEM electrolysis which drive cost and efficiency, the large potential for improvements in these areas for electrolyzers, and remaining technical challenges that need to be addressed to provide a full solution. Specific barriers which remain in translating the output of successful research programs to commercialization will also be addressed, including manufacturing and integration challenges. Finally, comparison between approaches in the United States and Europe in supporting technology development will be presented.

Orthogonal Golay excitation for range side lobe elimination in dual-frequency harmonic imaging

Dual-frequency (DF) tissue harmonic imaging has been developed to take advantage of not only harmonic signal at second harmonic (2f0) frequency but also the inter-modulation harmonic signal at fundamental (f0) frequency for simultaneous nonlinear detection. Though phase-encoded Golay pair can improve the signal-to-noise ratio of DF harmonic signal at both f0 and 2f0 frequencies, conventional matched filtering cannot correctly decode the crosstalk from harmonic components at DC and third harmonic (3f0) frequency and will lead to range side lobe artifacts in DF harmonic imaging. For orthogonal Golay pair, however, exchanging the decoding filter will output zero for the signal and keep the crosstalk the same. Therefore, the output of exchanged filtering can be subtracted from that of the original matched filtering to completely remove the spectral crosstalk. Compared to phase inversion method, the proposed orthogonal Golay decoding does not require additional transmits to cancel the unwanted DC and 3f0 harmonic interferences and thus the achievable frame rate remains the same. Various experiments have been performed to verify the efficacy of the proposed orthogonal Golay decoding. Results from hydrophone measurements indicate that the proposed method effectively suppresses the spectral overlap between the harmonic signal and the interference. Corresponding range side lobe level (RSLL) can be suppressed by 10–20dB when the signal bandwidth is 60%. B-mode harmonic imaging also demonstrates a reduction of side lobe magnitude (SLM) by 8dB at 2f0 frequencies.

Humidity-Induced Charge Leakage and Field Attenuation in Electric Field Microsensors

The steady-state zero output of static electric field measuring systems often fluctuates, which is caused mainly by the finite leakage resistance of the water film on the surface of the electric field microsensor package. The water adsorption has been calculated using the Boltzmann distribution equation at various relative humidities for borosilicate glass and polytetrafluoroethylene surfaces. At various humidities, water film thickness has been calculated, and the induced charge leakage and field attenuation have been theoretically investigated. Experiments have been performed with microsensors to verify the theoretical predictions and the results are in good agreement.

Design and evaluation of a vibration sensor for measurement-while-drilling

For the specific functional requirements, the design based on micro-quartz gyro was carried out to keep the sensor's detection side in isolation from the exciting vibrations in Measurement-While-Drilling (MWD). The supporting node position was calculated to achieve good resistance to the shock and vibration. A new mechanical structure prototyping based on Finite Element Method (FEM) was presented to achieve good zero output stability. The structure designation and performance optimisation of a solid-vibration-beam based gyro were also introduced, and the validity of the designed prototyping was proved through the simulation experiments.