A building management system (BMS) including a controller having an adaptive interaction manager and an agent manager. The system further includes one or more input-output (I/O) devices, the I/O devices in communication with the adaptive interaction manager. The controller further including a number of BMS field devices. The I/O devices are configured to receive an input from a user, and further configured to communicate the input to the adaptive interaction manager. The agent manager is configured to determine if one or more existing software agents are capable of performing the desired action, and to automatically transmit the existing software agents to one or more of the BMS field devices based on the agent manager determining the existing software agents are capable of performing the desired action. The software agents are configured to automatically be installed in a processing circuit of the BMS field device to perform the required action.

The present invention relates to systems and methods for improved building systems management and maintenance. The present invention provides a system for providing an augmented reality-like interface for the management and maintenance of building systems, specifically the mechanical, electrical, and plumbing (MEP) systems within a building, including the heating, ventilation, and air-conditioning (HVAC) systems.

A method for interacting with a building management system (BMS) using intelligent software agents. The method includes receiving a user request from a multi-input device configured to accept vocal and textual inputs, and contextualizing the user request for a space and/or place and a corresponding user. The method further includes constructing a user skill level from the user request, and activating a customized BMS optimization process, the customized BMS optimization process determined by the intelligent software agents from the user skill level, the user request, and the space and/or place.

A computer-implemented method for training a deep neural network includes defining a loss function corresponding to the deep neural network, receiving a training dataset comprising training samples, and setting current parameter values to initial parameter values. An optimization method is performed which iteratively minimizes the loss function. During each iteration, a steepest direction of the loss function is calculated by determining the gradient of the loss function at the current parameter values. A batch of samples included in training samples is selected. A matrix-free CG solver is applied to obtain an inexact solution to a linear system defined by the steepest direction of the loss function and a stochastic Hessian matrix with respect to the batch of samples. A descent direction is determined, and the parameter values are updated based on the descent direction. Following the optimization method, the parameter values are stored in relationship to the deep neural network.

Disclosed herein are related to a method, a system, and a non-transitory computer readable medium storing instructions for operating a group of central plant equipment to serve thermal energy loads of a building or building system. In one approach, a base capacity of one or more devices of the group of central plant equipment is identified. A change in requested load allocated to the one or more devices crossing the base capacity at a crossover time may be detected. In one approach, an adjusted capacity of the one or more devices is set such that the adjusted capacity is offset from the base capacity before the crossover time and decays toward the base capacity during a decay period after the crossover time. The requested load allocated may be compared with the adjusted capacity after the crossover time. The group of central plant equipment may be operated according to the comparison.

A method of determining multimodal polyethylene quality comprising the steps of (a) providing a multimodal polyethylene resin sample; (b) determining, in any sequence, the following: that the multimodal polyethylene resin sample has a melt index within 30% of a target melt index; that the multimodal polyethylene resin sample has a density within 2.5% of a target density; that the multimodal polyethylene resin sample has a dynamic viscosity deviation (% MVD) from a target dynamic viscosity of less than about 100%; that the multimodal polyethylene resin sample has a weight average molecular weight (M.sub.w) deviation (% M.sub.wD) from a target M.sub.w of less than about 20%; and that the multimodal polyethylene resin sample has a gel permeation chromatography (GPC) curve profile deviation (% GPCD) from a target GPC curve profile of less than about 15%; and (c) responsive to step (b), designating the multimodal polyethylene resin sample as a high quality resin.

Deep Learning is a candidate for advanced process control, but requires a significant amount of process data not normally available from regular plant operation data. Embodiments disclosed herein are directed to solving this issue. One example embodiment is a method for creating a Deep Learning based model predictive controller for an industrial process. The example method includes creating a linear dynamic model of the industrial process, and based on the linear dynamic model, creating a linear model predictive controller to control and perturb the industrial process. The linear model predictive controller is employed in the industrial process and data is collected during execution of the industrial process. The example method further includes training a Deep Learning model of the industrial process based on the data collected using the linear model predictive controller, and based on the Deep Learning model, creating a Deep Learning model predictive controller to control the industrial process.

Examples disclosed herein relate to an energy device including a memory, one or more processors, a transceiver, a display, and a dashboard. The memory includes one or more energy modules. The one or more processors are configured to communicate via the transceiver with one or more energy devices. The display is configured to display a dashboard of energy options based on one or more signals received from the one or more processors.

A server includes an interface, programmed to receive, from a vehicle, vehicle data indicative of vehicle status and user data indicative of usage of vehicle features by a user; and a processor, programmed to analyze the vehicle data and the user data using artificial intelligence (AI) logic to generate a comfort prediction for the user; and configure a comfort device associated with the user, external to the vehicle, using the comfort prediction.

A rotating equipment system with in-line drive-sense circuit (DSC) electric power signal processing includes rotating equipment, in-line drive-sense circuits (DSCs), and one or more processing modules. The in-line DSCs receive input electrical power signals and generate motor drive signals for the rotating equipment. An in-line DSC receives an input electrical power signal, processes it to generate and output a motor drive signal to the rotating equipment via a single line and simultaneously senses the motor drive signal via the single line. Based on the sensing of the motor drive signal via the single line, the in-line DSC provides a digital signal to the one or more processing modules that receive and process the digital signal to determine information regarding one or more operational conditions of the rotating equipment, and based thereon, selectively facilitate one or more adaptation operations on the motor drive signal via the in-line DSC.