Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “MPC”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 145 records · Page 8
All sourcesNASA NTRSDOE OSTI
NASA NTRS1990

Observations of surface-brightness fluctuations in Virgo

VRI photometric observations of surface-brightness fluctuation in 13 Virgo Cluster galaxies, M32, and NGC 3379, obtained using an 800 x 800 CCD array on the 4-m telescope at KPNO during March 1989, are reported. The data-reduction techniques are described in detail, and the results are presented in tables and graphs. Particular attention is given to distance estimates derived from the I-band fluctuations: it is shown that NGC 4365 is behind the cluster, while NGC 4468 and NGC 4489 are in the foreground. The average distance to the cluster is found to be 17 + or - 1 Mpc, with H0 = 78 + or - 6 km/sec Mpc; the corresponding values from a direct isochrone calibration are 21 Mpc and 64 + or - 5 km/sec Mpc.

Tonry, John L.↗
NASA NTRS1993

Cluster evolution as a probe of primordial density fluctuations

Although COBE's detection of large angle microwave background anisotropies fixes the amplitude of density fluctuations on length scales k exp -1 approximately = (300-6000) h(exp -1)Mpc, what is crucial for the level of large scale clustering is the amplitude of density fluctuations on scales (5-50) h(exp -1)Mpc. The level of dynamical clustering is usually parameterized by the size of the mass fluctuations in 8 h exp -1 Mpc spheres, sigma sub 8. For the cold dark matter model, COBE gives sigma sub 8 approximately = 1, while models with extra large scale power give sigma sub 8 approximately = 1/2. The most massive clusters of galaxies (greater than or approximately = 10 exp 15 solar mass) form from rare 'peak patches' found in the initial mass density distribution. Their abundance as a function of redshift is a sensitive probe of the wave number band k(exp -1) approx. (3-8) h(exp -1)Mpc, hence of sigma sub 8, and so cluster evolution can discriminate among models allowed by the COBE results. We use our Hierarchical Peaks Method, which accurately reproduces the results of P3M N-body simulations, to calculate the evolution of cluster x-ray flux counts, luminosity, and temperature functions as a function of sigma sub 8 for CDM models and those with more large scale power. We find that the EMSS and Edge et al. cluster samples support sigma sub 8 in the range from approx. 0.6-0.9, and that models with more large scale power (and hence flatter fluctuation spectra in the cluster regime) fit the x-ray bright end better.

Bond, J. Richard↗
NASA NTRS1994

A hydrodynamic approach to cosmology: The mixed dark matter cosmological scenario

We compute the evolution of spatially flat, mixed cold and hot dark matter models containing both baryonic matter and two kinds of dark matter. Hydrodynamics is treated with a highly developed Eulerian hydrodynamic code (see Cen 1992). A standard particle-mesh (PM) code is also used in parallel to calculate the motion of the dark matter components. We adopt the following parameters: h equivalent to (sub 0)/100 km/s Mpc(exp -1) = 0.5, OMEGA(sub C) = 0.3, and OMEGA(sub B) = 0.06, with amplitude of the perturbation spectrum fixed by the Cosmic Background Explorer Satellite (COBE) Dark Matter Radiation (DMR) measurements (Smoot et al. 1992) being sigma (sub 8) = 0.67. Four different boxes are simulated with box sizes of L = (64, 16, 4, 1) h(exp -1) Mpc, respectively, the two small boxes providing good resolution but little valid information due to absence of large-scale power. We use 128(exp 3) approximate 10(exp 6.3) baryonic cells, 128(exp .3) cold dark matter particles, and 2 x 128(exp 3) hot dark matter particles. In addition to the dark matter we follow separately six baryonic species (H, H(+), He, He(+), He(++), e(-)) with allowance for both (nonequilibrium) collisional and radiative ionization in every cell. The background radiation field is also followed in detail with allowance made for both continuum and line processes, to allow nonequilibrium heating and cooling processes to be followed in detail. The mean final Zeldovich-Sunyaev y parameter is estimated to be y Bar = (5.4 + or - 2.7) x 10(exp -7) below currently attainable observations, with a rms fluctuation of approximately delta bar y = (0.6 + or - 3.0) x 10(exp -7) on arcminute scales. The rate of galaxy formation peaks at an even later epoch (z approximate 0.3) than in the standard (OMEGA = 1, sigma sub 8 = 0.67) cold dark matter (CDM) model (z approximate 0.5) and, at a redshift of z = 4, is nearly a factor of 100 lower than for the CDM model with the same value of sigma sub 8. With regard to mass function, the smallest objects are stabilized against collapse by thermal energy: the mass-weighted mass spectrum has a broad peak in the vicinity of M(sub B) = 10(exp 9.5) solar mass with a reasonable fit to the Schechter luminosity function if the ratio of baryon mass to blue light is approximately 4. In addition, one very large PM simulation was made in a box with size (320 h(exp - 1) Mpc) containing 3 x 200(exp 3) = 10(exp 7.4) particles. Utilizing this simulation we find that the model yields a cluster mass function which is about a factor of 4 higher than observed, but a cluster-cluster correlation length marginally lower than observed, but that both are closer to observations than in the (COBE) normalized CDM model. The one-dimensional pairwise velocity dispersion is 605 + or - 8 km/s at 1/h separation, lower than that of the DCM model normalized to COBE, but still significant higher than observations (Davis & Peebles 1983). A plausible velocity bias b(sub v) = 0.8 + or - 0.1 on this scale will reduce but not remove the discrepancy. The velocity auto-correlat ion function has a coherence length of 40/h Mpc, which is somewhat lower than the observed counterpart. In all these respects the model would be improved by decreasing the cold fraction of the dark OMEGA(sub CDM)/ (OMEGA(sub CDM) + OMEGA(sub HDB). But formation of galaxies and clusters of galaxies is much later in this model than in COBE-normalized CDM, perhaps too late. To improve on these constraints a larger ratio of OMEGA(sub CDM)/ (OMEGA(sub CDM) + OMEGA(sub HDM)) is required than the value of 0.67 adopted here. It does not seem possible to find a value for this ratio which would satisfy all tests. Overall, the model is similar both on large and intermediate scales to the standard CDM model normalized to the same value of sigma(sub B), but the problem with regard to late formation of galaxies is more severe in this model than in that CDM model. Adding hot dark matter, significantly improves the ability of the COBE-normalized CDM scenario to fit existing observations, but the model is in fact not as good as the CDM model with the same sigma(sub 8) and is still probably unsatisfactory with regard to several critical tests.

Cen, Renyue↗
NASA NTRS1994

Internal velocity and mass distributions in simulated clusters of galaxies for a variety of cosmogonic models

The mass and velocity distributions in the outskirts (0.5-3.0/h Mpc) of simulated clusters of galaxies are examined for a suite of cosmogonic models (two Omega(sub 0) = 1 and two Omega(sub 0) = 0.2 models) utilizing large-scale particle-mesh (PM) simulations. Through a series of model computations, designed to isolate the different effects, we find that both Omega(sub 0) and P(sub k) (lambda less than or = 16/h Mpc) are important to the mass distributions in clusters of galaxies. There is a correlation between power, P(sub k), and density profiles of massive clusters; more power tends to point to the direction of a stronger correlation between alpha and M(r less than 1.5/h Mpc); i.e., massive clusters being relatively extended and small mass clusters being relatively concentrated. A lower Omega(sub 0) universe tends to produce relatively concentrated massive clusters and relatively extended small mass clusters compared to their counterparts in a higher Omega(sub 0) model with the same power. Models with little (initial) small-scale power, such as the hot dark matter (HDM) model, produce more extended mass distributions than the isothermal distribution for most of the mass clusters. But the cold dark matter (CDM) models show mass distributions of most of the clusters more concentrated than the isothermal distribution. X-ray and gravitational lensing observations are beginning providing useful information on the mass distribution in and around clusters; some interesting constraints on Omega(sub 0) and/or the (initial) power of the density fluctuations on scales lambda less than or = 16/h Mpc (where linear extrapolation is invalid) can be obtained when larger observational data sets, such as the Sloan Digital Sky Survey, become available.

Cen, Renyue↗
NASA NTRS2000

Predictive and Neural Predictive Control of Uncertain Systems

Accomplishments and future work are:(1) Stability analysis: the work completed includes characterization of stability of receding horizon-based MPC in the setting of LQ paradigm. The current work-in-progress includes analyzing local as well as global stability of the closed-loop system under various nonlinearities; for example, actuator nonlinearities; sensor nonlinearities, and other plant nonlinearities. Actuator nonlinearities include three major types of nonlineaxities: saturation, dead-zone, and (0, 00) sector. (2) Robustness analysis: It is shown that receding horizon parameters such as input and output horizon lengths have direct effect on the robustness of the system. (3) Code development: A matlab code has been developed which can simulate various MPC formulations. The current effort is to generalize the code to include ability to handle all plant types and all MPC types. (4) Improved predictor: It is shown that MPC design using better predictors that can minimize prediction errors. It is shown analytically and numerically that Smith predictor can provide closed-loop stability under GPC operation for plants with dead times where standard optimal predictor fails. (5) Neural network predictors: When neural network is used as predictor it can be shown that neural network predicts the plant output within some finite error bound under certain conditions. Our preliminary study shows that with proper choice of update laws and network architectures such bound can be obtained. However, much work needs to be done to obtain a similar result in general case.

Kelkar, Atul G.↗
NASA NTRS2020

A Large, Deep 3 Deg2 Survey of Hα, [OIII], and [OII] Emitters from LAGER: Constraining Luminosity Functions

We present our measurements of the Hα, [OIII], and [O II] luminosity functions as part of the Lyman Alpha Galaxies at Epoch of Reionization (LAGER) survey using our samples of 1577 z=0.47Hα-, 3933 z=0.93 [O III]-, and 5367 z=1.59 [O II]-selected emission line galaxies in a 3 deg2 single, CTIO/Blanco DECam pointing of the COSMOS field. Our observations reach 5σ depths of 8.2 × 10−18 erg s−1 cm−2 and comoving volumes of (1−7) × 105 Mpc3 making our survey one of the deepest narrow-band surveys. We select our emission line galaxies via spectroscopic confirmation, photometric redshifts, and colour–colour selections. We measure the observed luminosity functions for each sample and find best fits of φ* = 10−3.16+0.09 −0.09 Mpc−3and L* = 1041.72+0.09 −0.09 erg s−1 for Hα, φ* = 10−2.16+0.10−0.12 Mpc−3 and L* = 1041.38+0.07 −0.06 erg s−1 for [O III], and φ* = 10−1.97+0.07−0.07 Mpc−3 and L* = 1041.66+0.03−0.03 erg s−1 for [OII], with α fixed to −1.75, −1.6, and −1.3, respectively. An excess of bright >1042 erg s−1 [O III] emitters is observed and may be due to active galactic nucleus (AGN) contamination. Corrections for dust attenuation are applied assuming AHα = 1 mag. We also design our own empirical restframe g − r calibration using SDSS DR12 data, test it against our z = 0.47 Hα emitters with zCOSMOS 1D spectra, and calibrate it for (g − r) between −0.8 and 1.3 mag. Dust and AGN corrected star formation rate densities (SFRDs) are measured as log10ρSFR/(Msun yr−1 Mpc−3)=−1.63 ± 0.04, −1.07 ± 0.06, and −0.90 ± 0.10 for Hα, [OIII], and [O II], respectively. We find our [O III] and [O II] samples fully trace cosmic star formation activity at their respective redshifts in comparison to multiwavelength SFRDs, while the Hα sample traces ∼70 per cent of the total z = 0.47 SFRD.

galaxies↗
NASA NTRSDate not supplied

Model Predictive Control Strategies for Turbine Electrified Energy Management

The increasing electrification of aircraft propulsion systems is leading to new control architectures being developed to address integration between electric machines and gas-based turbine engines. For hybrid-electric propulsion systems, current conceptual architectures often couple electric machines with the shafts of gas turbine engines and introduce energy storage. Leveraging the electrical power system of hybridized engines, Turbine Electrified Energy Management (TEEM) is a recent control approach that improves transient operability in an effort to enable more efficient and lighter weight turbomachinery. This study seeks to expand TEEM’s application beyond traditional proportional-integral (PI) control by presenting linear model predictive control (MPC) schemes to execute the TEEM concept. Through constraint selection and cost function design, transient operability goals for TEEM are considered with no external logic or saturation. Unique to the designs are the use of a washout filter, which simplifies transient detection and motor activation logic. The proposed architectures are implemented with both centralized MPC and distributed MPC approaches, and comparisons are drawn to a benchmark PI controller simulated on a nonlinear turbofan engine model at one ground condition and one cruise condition. Performance is evaluated using compressor maps, stall margin performance, and two novel metrics: transient stack usage and transient excursion integral. Results reveal the linear MPC scheme performs comparably to the baseline controller and can be implemented in at least two distinct configurations with potential for further modifications, thus establishing the groundwork for future investigations.

Model Predictive Control↗
NASA NTRSDate not supplied

Model Predictive Control Strategies for Turbine Electrified Energy Management

The increasing electrification of aircraft propulsion systems is leading to new control architectures being developed to address integration between electric machines and gas-based turbine engines. For hybrid-electric propulsion systems, current conceptual architectures often couple electric machines with the shafts of gas turbine engines and introduce energy storage. Leveraging the electrical power system of hybridized engines, Turbine Electrified Energy Management (TEEM) is a recent control approach that improves transient operability in an effort to enable more efficient and lighter weight turbomachinery. This study seeks to expand TEEM’s application beyond traditional proportional integral (PI) control by presenting linear model predictive control (MPC) schemes to execute the TEEM concept. Through constraint selection and cost function design, transient operability goals for TEEM are considered with no external logic or saturation. Unique to the designs are the use of a washout filter, which simplifies transient detection and motor activation logic. The proposed architectures are implemented with both centralized MPC and distributed MPC approaches, and comparisons are drawn to a benchmark PI controller simulated on a nonlinear turbofan engine model at one ground condition and one cruise condition. Performance is evaluated using compressor maps, stall margin performance, and two novel metrics: transient stack usage and transient excursion integral. Results reveal the linear MPC scheme performs comparably to the baseline controller and can be implemented in at least two distinct configurations with potential for further modifications, thus establishing the groundwork for future investigations.

Model Predictive Control↗
NASA NTRS2023

Model Predictive Control Strategies for Turbine Electrified Energy Management

The increasing electrification of aircraft propulsion systems is leading to new control architectures being developed to address integration between electric machines and gas-based turbine engines. For hybrid-electric propulsion systems, current conceptual architectures often couple electric machines with the shafts of gas turbine engines and introduce energy storage. Leveraging the electrical power system of hybridized engines, Turbine Electrified Energy Management (TEEM) is a recent control approach that improves transient operability in an effort to enable more efficient and lighter weight turbomachinery. This study seeks to expand TEEM’s application beyond traditional proportional-integral (PI) control by presenting linear model predictive control (MPC) schemes to execute the TEEM concept. Through constraint selection and cost function design, transient operability goals for TEEM are considered with no external logic or saturation. Unique to the designs are the use of a washout filter, which simplifies transient detection and motor activation logic. The proposed architectures are implemented with both centralized MPC and distributed MPC approaches, and comparisons are drawn to a benchmark PI controller simulated on a nonlinear turbofan engine model at one ground condition and one cruise condition. Performance is evaluated using compressor maps, stall margin performance, and two novel metrics: transient stack usage and transient excursion integral. Results reveal the linear MPC scheme performs comparably to the baseline controller and can be implemented in at least two distinct configurations with potential for further modifications, thus establishing the groundwork for future investigations.

Model Predictive Control↗
DOE OSTI2023

Adaptive Data-Driven Model Predictive Control for Heat Pipe Microreactors

To establish a technical basis for self-regulating microreactors, a model predictive control (MPC) system is investigated to proactively respond to anomalies and disturbances in anticipation of potential deviations from operating setpoints. Due to the difficulty of developing a physics-based surrogate model that can accurately match plant data in various operating conditions, machine learning algorithms are used in MPC, which allow for learning from both simulation and operation data, thus efficiently describing the targeted transient with arbitrary accuracy. However, one of the biggest concerns in applying ML algorithms like artificial neural networks (ANNs) is that the predictive capabilities of ANN are limited by training data. If there are gaps between the training and target domain, the accuracy of an ANN can degrade significantly when it is used to predict unseen data. To improve the predictive capability of ANN and enable a confident use of data-driven MPCs outside the training data, this study proposes an adaptive data-driven MPC framework. The system will monitor the discrepancy between plant responses and surrogate predictions, fine-tune the ANN-based surrogate when a large discrepancy is detected, and continue MPC operation with updated surrogates. The framework is demonstrated on a point kinetic model for microreactors. The hyperparameters of the update strategy, including layers to update, error thresholds, learning rate discount, and number of data points used for fine-tuning, are optimized so the simulated microreactor is able to follow changes in setpoint with the smallest of deviations.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2023

Adaptable Data Driven Model Predictive Control for Heat Pipe Microreactors

To establish a technical basis for self-regulating microreactors, a model predictive control (MPC) system is investigated to proactively respond to anomalies and disturbances in anticipation of potential deviations from operating setpoints. Due to the difficulty of developing a physics-based surrogate model that can accurately match plant data in various operating conditions, machine learning algorithms are used in MPC, which allow for learning from both simulation and operation data, thus efficiently describing the targeted transient with arbitrary accuracy. However, one of the biggest concerns in applying ML algorithms like artificial neural networks (ANNs) is that the predictive capabilities of ANN are limited by training data. If there are gaps between the training and target domain, the accuracy of an ANN can degrade significantly when it is used to predict unseen data. To improve the predictive capability of ANN and enable a confident use of data-driven MPCs outside the training data, this study proposes an adaptive data-driven MPC framework. The system will monitor the discrepancy between plant responses and surrogate predictions, fine-tune the ANN-based surrogate when a large discrepancy is detected, and continue MPC operation with updated surrogates. The framework is demonstrated on a point kinetic model for microreactors. The hyperparameters of the update strategy, including layers to update, error thresholds, learning rate discount, and number of data points used for fine-tuning, are optimized so the simulated microreactor is able to follow changes in setpoint with the smallest of deviations.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2022

Robust constrained tension control for high-precision roll-to-roll processes

Tension control is critical for maintaining good product quality in most roll-to-roll (R2R) production systems. Previous work has primarily focused on improving the disturbance rejection performance of tension controllers. Here, a robust linear parameter-varying model predictive control (LPV-MPC) scheme is designed to enhance the tension tracking performance of a pilot R2R system for deposition of materials used in flexible thin film applications. The performance of a tension controller may degrade due to disturbances associated with model uncertainties and the slowly-changing dynamics in R2R systems. We introduce a method that separately treats these two sources of disturbance. The controller utilizes an incremental model to eliminate the errors caused by the mismatch between the nominal model and the actual system. A tube-based MPC formulation combined with scheduled parameters adequately updates models and corrects for the time-varying dynamics. Constraints on the rated motor torque are incorporated in the MPC to maintain the controller reliability and avoid machine failures. We illustrate the operation of our control algorithm through simulation of an actual R2R system. The controller outperforms the benchmarks in terms of fast transient response and offset-free tension tracking. Furthermore, it also demonstrates immunity from variations due to parametric uncertainties.

42 ENGINEERING↗
DOE OSTI2021

Feed-Forward Compensation for Model Predictive Control in Tri-port Current-Source Medium-Voltage String Inverters for PV-Plus-Storage Farms

The novel tri-port current-source medium-voltage string inverter (TCS-MVSI) is a promising candidate for large-scale PV-plus-storage (PVS) farms owing to its galvanic isolation, easy storage integration, soft-switching capability across entire load range, controlled low dv/dt and EMI, benign fault tolerance, etc. Due to its low inertia feature, traditional PI-based control cannot manage large transients effectively. Instead, a model-based predictive control (MPC) is proposed to achieve robust and stable operation. As is well known, the control performance of MPC is compromised by the sampling and computational delay during implementation significantly, if not well addressed. This paper analyzed and quantified these delays and then proposes feed-forward compensation (FFC) for the MPC method to compensate the delays and the large parameter variations due to the low-inertia nature. In addition, this method also compensates for the high dc-link ripple within each switching cycle, a unique issue for low-inertia converters. Here, the proposed method requires low computational effort, allowing it to be extended for multiple ports. The effectiveness of the proposed method has been validated in experiments. In 10 kW test, the proposed method decreases the average dc-link current by 17%, leading to ~20% conduction losses and ~0.5% increase in converter efficiency. In addition, the peak dc-link current also decreases by 15%, resulting in reduced transformer size. As a result, an increased power density can be achieved with the proposed method. Similar improvements have been observed across the power range from 2 kW to 10 kW.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2022

Data-Enabled Predictive Control for Building HVAC Systems

Model predictive control is widely used as a control technology for the computation of optimal control inputs of building heating, ventilating, and air conditioning (HVAC) systems. However, both the benefits and widespread adoption of model predictive control (MPC) are hindered by the effort of model creation, calibration, and accuracy of the predictions. In this paper, we apply the data-enabled predictive control (DeePC) algorithm for designing controls for building HVAC systems. The algorithm solely depends on input/output data from the system to predict future state trajectories without the need for system identification. The algorithm relies on the idea that a vector space of all input–output trajectories of a discrete-time linear time-invariant (LTI) system is spanned by time-shifts of a single measured trajectory, given the input signal is persistently exciting. Closed-loop simulations using EnergyPlus are performed to demonstrate the approach. The simulated building modeled in EnergyPlus is a modified commercial large office prototype building served by an air handling unit-variable air volume HVAC system. Temperature setpoints of zones are used as control variables to minimize the HVAC energy cost of the building considering a time-of-use electricity rate structure. Furthermore, sensitivity analysis is conducted to gain insights into the effect of parameter tuning on DeePC performance. Simulation results are used to illustrate the performance of the algorithm and compare the algorithm with model-based MPC and occupancy-based setpoint controller. Overall, DeePC achieves similar performance compared to MPC for lower engineering effort.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2025

A Robust Data-Driven Approach for Mechanical Serial Sectioning

Mechanical serial sectioning (MSS) provides detailed microstructural information across large length scales. By repeatedly removing thin layers of material and imaging the exposed surface, a 3D representation of a specimen’s internal structure can be constructed, enabling failure analysis and feature identification that are otherwise inaccessible via conventional 2D or nondestructive evaluation techniques. Achieving consistent and accurate material removal can be challenging due to system variability, requiring an experienced operator to manually adjust parameters, prolonging data collection times and necessitating post-processing routines to standardize the data. Here, to address these challenges, this paper presents the employment of a one-step model predictive control (MPC) framework tailored to a run-to-run (R2R) controller. The R2R-MPC controller automates the parameter selection process, improving the consistency of material removal through iterative feedback for disturbance rejection and accurate tracking of the target removal rate. Using a data-driven approach, the controller robustly adapts to changing material characteristics. The effectiveness of the R2R-MPC controller is demonstrated through simulation and experimental results and compared to previous data collection procedures.

3D Materials Science↗
DOE OSTI2024

HHL algorithm with mapping function and enhanced sampling for model predictive control in microgrids

Here, this paper presents a refined quantum Harrow Hassidim Lloyd (HHL) algorithm for microgrid control. The first novelty of the developed method is that a mapping shift function enables the original HHL algorithm to handle general linear equations with non-singular and indefinite matrix. Second, a method of Matrix Extension for Amplifying Sampling Probabilities of Intended Solution (ME-ASPI) is proposed to design the reformulated linear algebraic equations, allowing for improved sampling efficiency of the quantum tomography in the refined HHL algorithm. Then, we applied the method to solve the model predictive control (MPC) problem in nonlinear dynamical microgrids. Specifically, with the ME-ASPI method, the refined HHL algorithm can effectively obtain the intended partial optimal control inputs for MPC. The optimization of quadratic programming problem in each time step of MPC is transformed into a linear system problem, which is addressed by the proposed quantum solver through using only partial information, with the time complexity improved from $\mathscr{O}(\mathscr{N}^{2.37286})$ classically to $\mathscr{O}(\mathscr{N}^{2} log \mathscr{N}$ x $p$ log $p)$ in quantum. Numerical examples have validated the effectiveness of the refined HHL algorithm with the proposed mapping function and the ME-ASPI method. By leveraging quantum properties, the proposed method provides a hybrid quantum–classical framework for microgrid control. This generic method can also potentially tackle many other challenges in analyzing and controlling general complex engineered systems.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

Deep reinforcement learning control for co-optimizing energy consumption, thermal comfort, and indoor air quality in an office building

With the recent demand for decarbonization and energy efficiency, advanced HVAC control using Deep Reinforcement Learning (DRL) becomes a promising solution. Due to its flexible structures, DRL has been successful in energy reduction for many HVAC systems. However, only a few researches applied DRL agents to manage the entire central HVAC system and control multiple components in both the water loop and the air loop, owing to its complex system structures. Moreover, those researches have not extended their applications by incorporating the indoor air quality, especially both CO2 and PM2.5concentrations, on top of energy saving and thermal comfort, as achieving those objectives simultaneously can cause multiple control conflicts. What's more, DRL agents are usually trained on the simulation environment before deployment, so another challenge is to develop an accurate but relatively simple simulator. Therefore, we propose a DRL algorithm for a central HVAC system to co-optimize energy consumption, thermal comfort, indoor CO2 level, and indoor PM2.5 level in an office building. To train the controller, we also developed a hybrid simulator that decoupled the complex system into multiple simulation models, which are calibrated separately using laboratory test data. The hybrid simulator combined the dynamics of the HVAC system, the building envelope, as well as moisture, CO2, and particulate matter transfer. Three control algorithms (rule-based, MPC, and DRL) are developed, and their performances are evaluated on the hybrid simulator environment with a realistic scenario (i.e., with stochastic noises). The test results showed that, the DRL controller can save 21.4 % of energy compared to a rule-based controller, and has improved thermal comfort, reduced indoor CO2 concentration. The MPC controller showed an 18.6 % energy saving compared to the DRL controller, mainly due to savings from comfort and indoor air quality boundary violations caused by unmeasured disturbances, and it also highlights computational challenges in real-time control due to non-linear optimization. Finally, we provide the practical considerations for designing and implementing the DRL and MPC controllers based on their respective pros and cons.

Guo, Fangzhou↗
DOE OSTI2022

Model-based predictive control of multi-zone commercial building with a lumped building modelling approach

Here this study investigates the applicability of a lumped building modeling approach to model-based predictive control (MPC) to alleviate the complex modeling process of the grey-box multi-zone building model. Based on experimental data, two building models were estimated in this study. The detailed model as a reference case and a lumped model were estimated with decentralized and conventional approaches, respectively. Then, simulations were performed with two boundary conditions, including the comfort bound and electricity cost structure. The performances of the MPC with the detailed and lumped models were analyzed compared to the feedback control. More savings was achieved with a larger comfort bound and more aggressive electricity cost structure. The savings potential of the proposed lumped model approach was not as high as that of the detailed model. However, the proposed method yields good control performance, whose savings was approximately 8.6% over that of feedback control. These results suggest that the proposed method can be used to facilitate MPC implementation in multi-zone building applications.

42 ENGINEERING↗