Abstract: A multi-objective optimized evaluation method of anti-seismic performance of a slope reinforced by a pile-anchor system is provided. The method includes: training an initial three-dimensional slope numerical calculation model to obtain an target three-dimensional slope numerical calculation model; determining numerical values to be imported into the target three-dimensional slope numerical calculation model according to deformation differences, to obtain a model analysis result; obtaining a simulation operation result according to a reinforcement scheme working condition table of the pile-anchor system; based on the model analysis result and the simulation operation result, evaluating anti-seismic reinforcing performance of the pile-anchor system to obtain comprehensive evaluation values, and then optimizing and evaluating the reinforcement schemes of an overall slope to be reinforced.

Abstract: A method of anomaly detection and energy-efficient inference determination includes receiving an input. A set of features of the input are extracted using an artificial neural network (ANN) to generate a latent representation of the input. A reconstruction of the input is generated using the ANN, based on the latent representation. A reconstruction error is computed based on the generated reconstruction and the input. The reconstruction error is compared to a predefined threshold to determine whether the in-distribution data or out-of-distribution data. An anomaly is detected in response to an out-of-distribution determination. A decision model is provided with the latent representation in response to the input being determined to be in-distribution data. In turn, the decision model computes an inference based on the latent representation.

Abstract: A system architecture of a fast transfer switch is provided, which includes a main power loop configured to be connected to at least two power supplies, transfer between the at least two power supplies, and output an AC signal; a first parallel bypass module configured to output an AC signal, the first parallel bypass module including an independent power supply; the main controller is configured to, in response to detecting the abnormality of the first power supply of the at least two power supplies, adjust the AC signal output by the first parallel bypass module during the transfer of the main power loop between the at least two power supplies, so as to transfer from the first power supply outputting the AC signal via the main power loop to the second power supply of the at least two power supplies outputting the AC signal via the main power loop.

Abstract: Embodiments of the present disclosure provide a voltage stabilizing circuit and a DC microgrid.

Abstract: The disclosure relates to a power supply device for a motor protector and a power supplying method thereof. The motor protector receives power. The power supply device comprises: a first power supplying unit, configured to receive and store the power and supply power to a tripping device of the motor protector; a second power supplying unit, configured to receive the power and supply power to a calculation and control device of the motor protector; a third power supplying unit, configured to receive and store the power and supply power to a reclosing device of the motor protector; and a control unit, configured to control an order in which the first power supplying unit, the second power supplying unit and the third power supplying unit receive the power.

Abstract: The present disclosure discloses a solid-state circuit breaker including an electronic switch and a mechanical switch connected in series, the mechanical switch includes: a handle micro-switch, configured to be actuated upon the handle rotating to the on position and released upon the handle leaving the on position; a contact micro-switch, installed to the housing and configured to be actuated upon the movable contact moving to a closed position and released upon the movable contact leaving the closed position. By detecting a state of the handle micro-switch and a state of the contact micro-switch, it is achieved that, upon the solid-state circuit breaker being switched from off to on, the mechanical switch is turned on first, and then the electronic switch is turned on, upon the solid-state circuit breaker being switched from on to off, the electronic switch is turned off first, and then the mechanical switch is turned off.

Abstract: A method and an apparatus for satellite laser broadband demodulation are provided. The method includes: setting a residual carrier to a carrier acquisition range of a receiver, pulling the residual carrier to an MHz level by adjusting a frequency of a local oscillator laser, and obtaining a precise carrier frequency according to an accurate frequency acquisition, such that the residual carrier enters a fast acquisition band of a carrier tracking phase-locked loop. After carrier acquisition is achieved, carrier tracking and data recovery processing are performed. According to the present disclosure, signal equalization of an ultra-high bandwidth/ultra-high bit rate can be implemented, and carrier acquisition, tracking, and demodulation are quickly achieved for a modulation signal in a high dynamic range.

Abstract: According to an aspect of the present disclosure, there is provided a switching circuit including a main power supply circuit connected between power sources and a load, wherein the power sources include a first power source and a second power source, and the main power supply circuit includes a transfer switch connected between the power sources and the load and configured to switch between the first power source and the second power source; and a power electronic switch for turning on or off the main power supply circuit, wherein the power electronic switch is connected between the transfer switch and the load.

Abstract: A method for compressing of a deep neural network model includes determining an architecture of a teacher model. An initial layer of the teacher model is preserved and included in a student model. Repeated layers of the teacher model having a same type are identified. The second of such layers is removed such that the student model includes fewer layers than the teacher model. Knowledge distillation is applied to train the student model. The identifying and removal of layers of the same type is repeated to compress the student model.

Abstract: A method for generating a personalized artificial neural network (ANN) model receives an input at a first artificial neural network. The input is processed to extract a set of intermediate features. The method determines if the input is out-of-distribution relative to a dataset used to train the first artificial neural network. The intermediate features corresponding to the input are provided to a second artificial neural network bases on the out-of-distribution determination. Additionally, the system resources for performing the training and inference tasks of the first artificial neural network and the second, personalized artificial neural network are allocated according to a computational complexity of the training and inference tasks and a power consumption of the resources.

Abstract: Certain aspects of the present disclosure are generally directed to apparatus and techniques for event state detection. One example method generally includes receiving a plurality of sensor signals at a computing device, determining, at the computing device, probabilities of sub-event states based on the plurality of sensor signals using an artificial neural network for each of a plurality of time intervals, and detecting, at the computing device, the event state based on the probabilities of the sub-event states via a state sequence model.

Abstract: A power distribution device is disclosed. The power distribution device includes: a first power source input end for connecting a first power source; a second power source input end for connecting a second power source; a first power source switch for connecting a first power source input end with a bus and switching on and off the power supply of the first power source; a second power source switch for connecting a second power source input end with a bus and switching on and off the power supply of the second power source; the bus for connecting with a load device and supplying power for the load device; and a load device switch for switching on and off the power supply to a relevant load device. Furthermore, A power distribution system is also disclosed.

Abstract: A method receives a first program code including one or more nested loops. A loop order is determined for the nested loop(s). The determined loop order aligns an input data layout and an output data layout. The nested loop(s) are transformed based on the loop order. A second program code is generated based on the transformed nested loop(s).

Abstract: The present invention discloses a power source transfer circuit for switching among a plurality of power sources to supply power to a load. The power source transfer circuit includes a first switch, a second switch and a bypass switch, wherein one of the plurality of power sources supplies power to the load via the first switch and the second switch, and wherein the bypass switch is connected in parallel to the second switch on a neutral line and is configured to control neutral overlapping duration during switching among the plurality of power sources. The power source transfer circuit including a bypass switch provided by the present disclosure can reduce the neutral overlapping duration to microseconds during power source transfer, so as to realize fast and reliable response.

Abstract: A modular parallel half-bridge integrated assembly with an annular layout is provided. The assembly includes a plurality of parallel sub-modules to improve a current capacity of the assembly. The sub-modules adopt an annular layout and are connected in parallel to balance currents of the sub-modules. The half-bridge integrated assembly includes a plurality of sub-modules, a plurality of heat sinks, a drive board, a direct current (DC) positive collecting busbar, a DC negative collecting busbar, and an alternating current (AC) collecting busbar. According to the assembly, since each insulated gate bipolar transistor (IGBT) is tightly bound to a capacitor, parasitic inductance of a commutation loop is small, realizing a small voltage overshoot and a fast switching speed for the IGBT module, so as to balance the currents of the sub-modules.

Abstract: A method for energy-efficient classification receiving, via a first circuit, an input data stream from one or more sensors. The first circuit detects, while a second circuit is in a dormant state, if a state change has occurred between a first input of the input data stream and a second input of the input data stream. The second input is a next succeeding input of the input data stream. The first circuit triggers the second circuit to perform a classification of the input data stream in response to detecting the state change.

Abstract: It is disclosed a motor thermal protection device and operation method thereof. The motor thermal protection device utilizes a power supply of a motor for power supplying and comprises: a capacitor timing circuit, configured to time after the motor is powered off; a power-off time determining unit, configured to perform a reading operation of reading an output voltage of the capacitor timing circuit after the motor is powered on, and determine a power-off time from when the motor is powered off to when the motor is powered on according to the output voltage; and a heat accumulation calculating unit, configured to calculate a remaining heat accumulation when the motor is powered on according to the power-off time and a heat accumulation when the motor is powered off.

Abstract: The present disclosure provides a multi-scale three-dimensional (3D) engineering geological model construction system and method. A regional geological model of a target region, a site geological model of each engineering site, and a drilling geological model of each drilling well are constructed. The geological model of each drilling well is superimposed to the site geological model of the corresponding engineering site in the way of step-by-step superimposition, and the site geological model of each engineering site fused with the drilling geological model is superimposed to the regional geological model of the target region. Thus, multi-scale geological model fusion of drilling well, engineering site, and regional mountain is realized.

Abstract: Various embodiments include methods and devices for weight layout transformation of a weight tensor. Embodiments may include, accessing a first memory to retrieve weights of the weight tensor in a transformed order that is different than an order for retrieving the weights for a calculation at a network layer of a trained machine learning model, and loading the weights to a second memory in the transformed order. Embodiments may further include retrieving the weights from the second memory in the transformed order, and reordering the weights to the order for implementing the calculation at the network layer of the trained machine learning model.