3results about How to "Reduce electricity usage" patented technology

The present disclosure designs natural assistant interactions. The invention provides systems and processes for operating a virtual assistant to provide natural assistant interactions. According to one or more examples, a method includes, at an electronic device with one or more processors and memory: receiving a first audio stream, the first audio stream including one or more utterances; determining whether the first audio stream includes a lexical trigger; generating one or more candidate textual representations of the one or more utterances; determining whether the virtual assistant is to ignore at least one of the one or more candidate textual representations. If at least one candidate textual representation is to be ignored, generating one or more candidate intents based on candidate textual representations of the one or more candidate textual representations other than the at least one candidate textual representation to be ignored.

The utility model discloses a photovoltaic heat pump energy storage and supply system which comprises a power grid, a distributed photovoltaic power station, an energy storage power station, a water source heat pump unit, a heat storage container and a cold storage container. A second switch is connected between the power output end of the distributed photovoltaic power station and the power grid, the power output end of the distributed photovoltaic power station and the power input end of the energy storage power station are connected with a third switch, and the power output end of the energy storage power station and the power input end of the water source heat pump unit are connected with a fourth switch. A fifth switch is connected between the power grid and the power input end of the water source heat pump unit, and a sixth switch is connected between the power output end of the energy storage power station and the power grid. Through the arrangement of the first switch to the sixth switch, the electric energy generated by the distributed photovoltaic power station can be switched among the energy storage power station, the water source heat pump unit and the power grid in multiple operation modes, the diversification of the operation modes is increased, and the electric power is reasonably utilized.

The invention relates to the technical field of cloud computing, and discloses a data center resource intelligent scheduling method and system based on a digital twinning and space-time diagram neural network, and the method comprises the steps: constructing a digital twinning graph model, collecting the real-time data of various resources of a data center, and converting the real-time data into a graph structure; constructing a space-time load prediction model, extracting multi-dimensional features of various resources of the data center, and performing load prediction on the resources; by setting multiple optimization targets, a scheduling optimization model is constructed, and a resource scheduling strategy is generated; and performing simulation verification on the scheduling scheme of the resource scheduling strategy, evaluating a strategy execution result, and performing feedback optimization on the scheduling model according to an evaluation result. According to the method, scheduling hysteresis is remarkably reduced, multi-dimensional resource global optimization is realized, prediction real-time performance and accuracy are considered, efficient simulation verification capability is provided, continuous self-adaptive optimization is supported, and operation cost is reduced.