Data-driven Model Predictive Control Research Articles

Minimizing total annualized cost per tonne of feed processed of a semicontinuous distillation process utilizing data-driven model predictive control

Semicontinuous distillation is a separation technique used to purify multicomponent mixtures with low to medium throughput. This research addresses the problem of designing a Data-driven Model Predictive Control (MPC) approach that enables minimizing the Total Annualized Cost (TAC) of the semicontinuous process per tonne of feed processed while maintaining the required product purity. In lieu of typically unavailable first principles models, the manuscript demonstrates the implementation of data-driven technique using data collected from an Aspen Plus Dynamics simulation as a test bed. A subspace model identification technique is adapted to develop a multi-model framework to capture the dynamic behavior of the process and then utilized within a Shrinking Horizon MPC (SHMPC) scheme, to achieve the required objective. The simulation results demonstrate a lowering of the TAC/tonne of feed by 11.4% compared to the traditional PI setup used in the previous studies.

Data-driven Koopman model predictive control for hybrid energy storage system of electric vehicles under vehicle-following scenarios

In the current studies on energy management strategy (EMS) for vehicle-following scenarios, the accuracy of vehicle state predictions based on mechanistic models is influenced by the time-varying conditions, affecting the optimization control performance. To address this issue, a data-driven Koopman model predictive control for hybrid energy storage system (HESS) of electric vehicles (EVs) in vehicle-following scenarios is proposed, combining the safety speed planning and energy management strategy. Firstly, a data-driven Koopman vehicle state prediction model is constructed in the upper layer for estimating parameters, such as road surface smoothness and slope. This model is then integrated into Model Predictive Control (MPC) to optimize the speed of the following vehicle. Subsequently, in the lower layer, utilizing the output from the upper layer and predicting load power, the load power is further allocated within the HESS. Simulation results demonstrate that in scenarios considering factors like slope, the hierarchical MPC with the Koopman model reduces energy consumption by 5.55% compared to the hierarchical MPC with mechanistic model.

Energy flexibility quantification of a tropical net-zero office building using physically consistent neural network-based model predictive control

Building energy flexibility plays a critical role in demand-side management for reducing utility costs for building owners and sustainable, reliable, and smart grids. Realizing building energy flexibility in tropical regions requires solar photovoltaics and energy storage systems. However, quantifying the energy flexibility of buildings utilizing such technologies in tropical regions has yet to be explored, and a robust control sequence is needed for this scenario. Hence, this work presents a case study to evaluate the building energy flexibility controls and operations of a net-zero energy office building in Singapore. The case study utilizes a data-driven energy flexibility quantification workflow and employs a novel data-driven model predictive control (MPC) framework based on the physically consistent neural network (PCNN) model to optimize the building energy flexibility. To the best of our knowledge, this is the first instance that PCNN is applied to a mathematical MPC setting, and the stability of the system is formally proved. Three scenarios are evaluated and compared: the default regulated flat tariff, a real-time pricing mechanism, and an on-site battery energy storage system (BESS). Our findings indicate that incorporating real-time pricing into the MPC framework could be more beneficial to leverage building energy flexibility for control decisions than the flat-rate approach. Moreover, adding BESS to the on-site PV generation improved the building self-sufficiency and the PV self-consumption by 17% and 20%, respectively. This integration also addresses model mismatch issues within the MPC framework, thus ensuring a more reliable local energy supply. Future research can leverage the proposed PCNN-MPC framework for different data-driven energy flexibility quantification types.

Real-life data-driven model predictive control for building energy systems comparing different machine learning models

By considering forecasts and exploiting storage effects, model predictive control can achieve significant energy and cost savings in the building sector. However, due to the high individual modeling effort, model predictive control lacks practical applicability. For that reason, data-driven process models, approximating the system behavior based on measurements, have become increasingly popular in recent years. Still, scientific literature lacks consent about the most promising model types and efficient workflows to integrate different machine learning models into a model predictive controller. With this work, we present a workflow to provide efficient model predictive controllers based on measurement data automatically. The main idea is to translate different machine learning models into optimization syntax to enable efficient optimization with full access to gradients. We currently consider artificial neural networks, gaussian process regression, and simple linear regression process models. We use a generic model ontology to automatize the controller generation further and test the methodology on two real-life use cases. The first use case is the application of five office rooms with smart thermostat valves. The second use case is a test hall with an air handling unit and a concrete core activation. Using only two days of initial training data, we deploy controllers based on the different model types for six weeks in the offices and apply online learning to improve the models continuously. We observe only minor differences in controller performance despite the artificial neural networks showing the highest prediction accuracy. The second use case shows that the simple linear models require less controller tuning effort. Thus, for practical applications, we recommend linear regression models.

Stability Margin of Data-Driven LQR and Its Application to Consensus Problem

In contrast with traditional control input design techniques based on mathematical models of the system, in data-driven control approaches, which have recently gained substantial attention, the controller is derived directly from the data that are collected from experiments or observations of the target system. In particular, several data-driven optimal control and model predictive control (MPC) techniques have been proposed. In this paper, it is shown that the recently proposed data-driven LQR (Linear Quadratic Regulator) has a stability margin that is the set of the uncertainties in the control input channels maintaining the closed-loop stability. As an application of the proposed stability margin of the data-driven LQR, the consensus problem is considered. Since the control design for the consensus of multi-agent systems can be reformulated into the robust stabilization of a linear system with uncertainty in the input channel, it is demonstrated that the derived stability margin can be used to design a controller for the consensus of multi-agent systems.

Data Driven Model Predictive Control for Hybrid Charging Stations using Ensemble Learning

Data Driven Model Predictive Control for Hybrid Charging Stations using Ensemble Learning

Data-Driven Nonlinear Model Predictive Control for Power Sharing of Inverter-based Resources

Data-Driven Nonlinear Model Predictive Control for Power Sharing of Inverter-based Resources

Identification for control approach to data-driven model predictive control

Identification for control approach to data-driven model predictive control

Identification for Control Approach to Data-Driven Model Predictive Control

Identification for Control Approach to Data-Driven Model Predictive Control

Data-Driven Model Predictive Control for Uncalibrated Visual Servoing

This paper addresses the image-based visual servoing (IBVS) control problem with an uncalibrated camera, unknown dynamics, and constraints. A novel data-driven uncalibrated IBVS (UIBVS) strategy is proposed, incorporated with the Koopman-based model predictive control (KMPC) algorithm and the adaptive robust Kalman filter (ARKF). First, to alleviate the need for calibration of the camera’s intrinsic and extrinsic parameters, the ARKF with an adaptive factor is utilized to estimate the image Jacobian matrix online, thereby eliminating the laborious camera calibration procedures and improving robustness against camera disturbances. Then, a data-driven MPC strategy is proposed, wherein the unknown nonlinear dynamic model is learned using the Koopman operator theory, resulting in a linear Koopman prediction model. Only input–output data are used to construct the prediction model, and hence, the proposed approach is robust against model uncertainties. Furthermore, with a symmetric quadratic cost function, the proposed approach solves the quadratic programming problem online, and visibility constraints as well as joint torque constraints are taken into account. As a result, the proposed KMPC scheme can be implemented in real time, and the UIBVS performance degradation which arises from the control torque constraints can be avoided. Simulations and comparisons for a 2-DOF robotic manipulator demonstrate the feasibility of the proposed approach. Simulation results further validate that the computation time of the proposed approach is comparable to the one of kinematic-based methods.

Quantitative comparison of reinforcement learning and data-driven model predictive control for chemical and biological processes

As manufacturing processes transition towards digitalization, data-driven process control is emerging as a key area of interest in future artificial intelligence technology. A crucial aspect in implementing data-driven process control is “What should we learn from the data?”. In general, the data-driven control method can be categorized into two main approaches: Learning the model and learning the value. To assist in selecting the more suitable approach, this paper applies six different control methods, with three falling under each approach, to three distinct manufacturing process systems. The simulation results indicate that the model-learning approaches display higher data efficiency and exhibit lower variance in total cost. These methods prove to be particularly advantageous for addressing the regulation problems. Conversely, value-learning approaches show competitive potential in closed-loop identification and in managing economic cost problems. The remaining challenges associated with each technique are discussed, along with practical considerations for their implementation.

Data-Driven Self-Triggered Control via Trajectory Prediction

Self-triggered control, a well-documented technique for reducing the communication overhead while ensuring desired system performance, is gaining increasing popularity. However, a majority of existing self-triggered control methods require explicit system models. An end-to-end control paradigm known as data-driven control designs control laws directly from data and offers a competing alternative to the routine system identification-then-control strategy. In this context, the present paper puts forth data-driven self-triggered control schemes for unknown linear systems using input-output data collected offline. Specifically, a data-driven model predictive control (MPC) scheme is proposed, which computes a sequence of control inputs while generating a predicted system trajectory. In addition, a data-driven self-triggering mechanism is designed, which determines the next triggering time using the solution of the data-driven MPC and the newly collected measurements. Finally, both feasibility and stability are established for the proposed self-triggered controller, which are validated using a numerical example.

Learning for Depth Control of a Robotic Penguin: A Data-Driven Model Predictive Control Approach

For bionic underwater robots, it is a great challenge for depth control due to model uncertainty and strong nonlinearity. To this end, we propose a data-driven model predictive control (MPC) approach using reinforcement learning (RL) for robotic penguin depth control. First, by imitating the underwater mode of the biological penguin, a robotic prototype with a tendon-driven head, two degrees of freedom wings, and a tendon-driven tail was designed. Then, a data-driven MPC framework is proposed considering the structure and motion properties of the robotic penguin. Specially, a data-based learning environment is constructed using a motion capture system, computational fluid dynamics, and a back-propagation neural network. Meanwhile, to maximize the benefits of the controller while ensuring safety and stability, a data-driven MPC using the RL scheme is applied to approximate the optimal policy. Combined with an appropriate reward design and periodic training, the closed-loop controller performance is significantly improved, and the validity of the proposed framework is finally tested by extensive simulations and experiments. Notably, this work will provide valuable insights into the learning-based motion control of bionic underwater robots.

A new model predictive control approach integrating physical and data-driven modelling for improved energy performance of district heating substations

District heating (DH) substations play a crucial role in ensuring the efficient and effective distribution of thermal energy necessary to provide space heating for buildings. However, optimizing their operation for energy savings while still ensuring indoor comfort poses significant challenges due to the complex dynamics of building demand and the inertia of building envelopes. To address these challenges, this study introduces a novel model predictive control (MPC) approach that combines a reduced-order physical model with a machine learning-based data-driven model to jointly optimize the operation parameters of a DH substation. In this approach, a reduced-order physical model is first used to capture essential operational principles and energy behaviors of the DH substations and generate candidate solutions for the control of the DH substations. Then, a data-driven model is constructed by integrating a Long Short-Term Memory model and a Back-propagation Neural Network, leveraging historical operational data of the DH substation concerned. The data-driven model is further formulated into a data-driven MPC framework to identify optimal control solutions from all candidates provided by the physical model. To evaluate the proposed approach, a data-driven surrogate model is developed using real operational data. Comparative analysis against the original fuzzy rule-based control strategy and a pure data-driven strategy demonstrates a substantial reduction in heat consumption of 4.77% and 19.47%, respectively. Moreover, compared with using a reduced-order physical model alone, this approach achieves additional benefits in reducing the energy consumption of the DH substation and minimizing indoor temperature fluctuations within the end-users.

RETRACTED: Data-driven model predictive control for real-time planned lead time optimization in a reconfigurable flow line

Mass customization has enabled smart manufacturing systems to develop reconfigurable machines and fixtures to fulfill customized products with high process variety. Coupled with these reconfigurable components, diversified design and operational problems should be dealt with such that the efficiency of reconfigurations can be maximized. This paper investigates a planned lead time (PLT) optimization problem after addressing the pre-determined reconfigurations and real-time disruptions in a flow line. A data-driven model predictive control (d-MPC) is developed for this problem such that the total expected cost for work-in-process inventory, final-product inventory, and tardiness penalty can be minimized. Considering the predictive multi-stage transitions and reconfigurable behaviors, d-MPC utilizes max-plus algebra to generate a time-varying state-space equation for a flow line. In this regard, a systematic method for permanent production loss identification can be formulated for real-time disruption diagnosis. This identification is useful in providing feedback signals for d-MPC and thus a closed-loop feedback control for PLT optimization can be achieved. A case study of an industrial welding line is reported to illustrate the operational benefits of d-MPC. The results demonstrate that d-MPC can efficiently cooperate with PLT optimization with machine reconfiguration such that the total expected cost can be minimized. Compared with continuous re-optimization, d-MPC adopts an event-driven mechanism to respond to real-time disruptions and can decrease the nervousness of flow lines with lower operational costs.

Driver-centric data-driven robust model predictive control for mixed vehicular platoon

Driver-centric data-driven robust model predictive control for mixed vehicular platoon

Grey-box model for model predictive control of buildings

Model predictive control (MPC) can improve energy efficiency and demand-side flexibility in buildings. Developing a grey-box model suitable for MPC is not straightforward, especially in buildings combining, not only ventilation and usual internal loads, but also Thermally Activated Building Structures (TABS) and large glass façades with external shading. To address these complexities, this paper presents a reduced order grey-box approach, considering all these elements. Various single zone model structures are compared, combining resistance–capacitance model, with finite difference or finite volume methods for modelling the TABS. The performance of these various model structures is evaluated using experimental data from a well-equipped living laboratory building. Additionally, the influence of technical parameters on the model's performance is investigated.The best model variant, with an enhanced glass façade model, achieves an accuracy of 0.25 °C of Mean Absolute Error over a year of simulation, on the 24 h zone temperature forecast compared to the measurement. This model has a small number of parameters (8), which are estimated with the least square non-linear method. The stability of the parameter values is analysed. The parameter identification requires only a small historical dataset of 1–2 weeks for startup and 2–4 weeks for training. This provides an adaptive model, in the sense that it is updated regularly (every day or week) based on recent measurement data. This data-driven evolving model is suitable across a wide range of applications involving data-driven Model Predictive Control (MPC) for buildings.

Data-driven learning-based Model Predictive Control for energy-intensive systems

Efficient control of energy-intensive systems is essential for reducing energy consumption and realizing sustainable development. However, considering the complex inter-dependent energy-consumption devices, numerous control parameters, and dynamic environments, the energy-efficient control of energy-intensive system is always challenging. To address such problems, this paper proposes a data-driven learning-based Model Predictive Control (MPC) method for the integrated control of various devices in energy-intensive systems. Specifically, a hybrid prediction model based on two variants of RNN is integrated with the MPC scheme to learn and predict the system dynamics based on massive time-series sensing data. Then an efficient tree-based prioritized group control model for heterogeneous devices is developed with a rolling optimization and feedback correction mode. A real-life case study is provided to evaluate the performance of the proposed method, which demonstrates its superiority over existing methods on saving the energy consumption.

An Integrated Control Approach for Shifting Process of Single-Axis Parallel Hybrid Electric Vehicle with a Multi-Speed AMT Gearbox

When a single-axis parallel hybrid electric vehicle (HEV) equipped with a multi-speed AMT gearbox is in its shifting process, the superposition of dynamic characteristics of multiple power sources and the intervention and withdrawal of AMT transmissions can easily cause significant vehicle longitudinal jerk. To achieve rapid and smooth output changes during the shifting process, this paper proposes an integrated multi-stage robust shifting control method for a single-axis parallel hybrid electric vehicles with a multi-speed AMT gearbox. First, models of key driveline components are constructed, and the shifting process is divided into five stages to provide a clear description of the control problem. Subsequently, we reproduce an integrated multistage robust control method to achieve favorable switching performance and control robustness under external disturbances. We propose a data-driven model predictive control strategy based on additional constraints in the torque unloading and recovery phases. Simultaneously, we present a joint control algorithm that integrates the optimal control and disturbance suppression in the speed synchronization phase. In addition, we develop a sliding mode auto-disturbance rejection control algorithm to achieve accurate position tracking of the shift actuator in the pickup and engage phases. Finally, simulations and bench tests are carried out to verify the effectiveness of the robust control method under different driving conditions. The results demonstrate that the proposed control method can not only coordinate the torque across different power sources and clutch while minimizing vehicle longitudinal jerk, shift time, and friction work, but also provides apparent robustness to model uncertainties and external disturbance. Therefore, the proposed method may offer a theoretical reference for the actual vehicle controller during shifting.

Safe operation of online learning data driven model predictive control of building energy systems

Model predictive control is a promising approach to reduce the CO2 emissions in the building sector. However, the vast modeling effort hampers the widescale practical application. Here, data-driven process models, like artificial neural networks, are well-suited to automatize the modeling. However, the underlying data set strongly determines the quality and reliability of artificial neural networks. In general, the validity domain of a machine learning model is limited to the data that was used to train it. Predictions based on system states outside that domain, so-called extrapolations, are unreliable and can negatively influence the control quality.We present a safe operation approach combined with online learning to deal with extrapolation in data-driven model predictive control. Here, the k-nearest neighbor algorithm is used to detect extrapolation to switch to a robust fallback controller. By continuously retraining the artificial neural networks during operation, we successively increase the validity domain of the artificial neural networks and the control quality.We apply the approach to control a building energy system provided by the BOPTEST framework. We compare controllers based on two data sets, one with extensive system excitation and one with baseline operation. The system is controlled to a fixed temperature set point in baseline operation. Therefore, the artificial neural networks trained on this data set tend to extrapolate in other operating points. We show that safe operation in combination with online learning significantly improves performance.