Sequential Quadratic Programming Research Articles

On optimizing the sensor spacing for pressure measurements on wind turbine airfoils

Abstract. This research article presents a robust approach to optimizing the layout of pressure sensors around an airfoil. A genetic algorithm and a sequential quadratic programming algorithm are employed to derive a sensor layout best suited to represent the expected pressure distribution and, thus, the lift force. The fact that both optimization routines converge to almost identical sensor layouts suggests that an optimum exists and is reached. By comparing against a cosine-spaced sensor layout, it is demonstrated that the underlying pressure distribution can be captured more accurately with the presented layout optimization approach. Conversely, a 39 %–55 % reduction in the number of sensors compared to cosine spacing is achievable without loss in lift prediction accuracy. Given these benefits, an optimized sensor layout improves the data quality, reduces unnecessary equipment and saves cost in experimental setups. While the optimization routine is demonstrated based on the generic example of the IEA 15 MW reference wind turbine, it is suitable for a wide range of applications requiring pressure measurements around airfoils.

Self-learning Markov prediction algorithm based aging-oriented gradient drop power control strategy for the transient modes of fuel cell hybrid electric vehicles

The power transients caused by switching from drive mode to brake mode in fuel cell hybrid electric vehicles (FCHEV) can result in significant degradation cost losses to the fuel cell. To address this issue, this study proposes a self-learning Markov prediction algorithm based aging-oriented gradient drop power control strategy. First, a real-time self-learning Markov predictor (SLMP) based on the traditional offline training Markov improvement is designed to predict the demand power and combined with the sequential quadratic programming (SQP) optimization algorithm to solve for the inner optimal demand power based on its global cost function minimization characteristic. On this basis, the fuel cell gradient drop power (FGDP) strategy is proposed to optimize the operating state of the vehicle powertrain under vehicle mode switching. This involves establishing a power gradient drop step based on considering the fuel cell hydrogen consumption cost and its lifetime degradation cost to further obtain the outer fuel cell demand power at the optimal step. And three execution modes are designed to trigger the FGDP strategy. Finally, by combining the above efforts, the SLMP-FGDP optimization control strategy is constructed. The numerical verification and hardware in loop experiments results show that the proposed improved SLMP can predict the vehicle demand power more accurately. Compared with the non-FGDP system, the SLMP-FGDP strategy can effectively near-eliminate the fuel cell power transient due to any braking scenario, thus effectively controlling the fuel cell lifetime degradation cost in a lower range and realizing a reduction of up to 52.21% of the fuel cell usage costs without significantly sacrificing the hydrogen fuel economy.

Development of an efficient design optimization strategy for thick-walled cylinders treated with combinations of autofrettage, shrink-fit and wire-winding processes

Shrink-fit, wire-winding, and autofrettage processes are commonly utilized to enhance fatigue strength and durability of thick-walled cylinders across various mechanical applications. In this study, a novel practical design optimization methodology has been developed to determine the optimal configuration of a thick-walled cylinder, incorporating different combinations of shrink-fit, wire-winding, and autofrettage techniques. The objective is to identify the optimal layer thickness, shrink-fit interference, conventional autofrettage pressure, and reverse autofrettage pressure, if applicable, to maximize the compressive residual stress and minimize the tensile residual stress, thereby extending fatigue lifetime of the cylinder. First, different configurations of thick-walled cylinders, subjected to various combinations of reinforcement processes, are identified. A dataset of residual hoop stress profiles through the cylinder thickness is subsequently generated for these configurations based on the same manufacturing process. Neural network regression is effectively utilized to construct a single fitting function for the residual hoop stress profiles. A parametric study is performed to determine the optimal training functions, activation functions, and hyperparameters, achieving a remarkable agreement with the dataset, indicated by a coefficient of determination of over 0.97. A combination of Genetic Algorithm and Sequential Quadratic Programming algorithms is utilized to determine the accurate optimal values. Fatigue life analysis is subsequently conducted to estimate the fatigue lifetime of the optimal configuration. Results suggest that the optimal configuration, involving conventional autofrettage of the inner layer followed by shrink-fitting with a virgin layer and wire-winding the entire assembly, achieves a maximum fatigue life of 88 × 10⁶ cycles under cyclic pressure load of 300 MPa.

Optimizing Air Contactor Design for Efficient Carbon Capture in Direct Air Capture Systems: A Multi-Disciplinary Approach

As carbon dioxide (CO2) emissions continue to rise, direct air capture (DAC) has emerged as an important technology for removing CO2 from ambient air to combat climate change. The air contactor design is a critical factor impacting DAC system performance and costs. This research describes a strategy for improving the design of a single air contactor unit to increase CO2 capture flux while minimizing costs using optimization techniques grounded in experimental data .The research focuses on a 100 m2 slab air contactor modeled utilizing cooling tower and scrubber principles. The system includes the chemical, mechanical, and structural engineering disciplines. The mass transfer coefficient, air velocity, packing depth, and capital cost per frontal area are four major design characteristics that were maximized.The model integrated requirements for low air velocity, pressure drop, and energy consumption, as well as adequate gas-liquid interaction via structured packing selection and depth.The design space was explored using a multi-objective optimization strategy based on gradient-based and heuristic methods. Initial sampling via a design of experiments methodology demonstrated response surface trends in the objective function across the design space. The response surfaces mapped the optimization objective as a function of the design variables. Using gradient information, sequential quadratic programming efficiently located an optimal design vector that met all constraints on the design variables and feasibility requirements. The constraints included bounds on mass transfer coefficient, air velocity, packing depth, and capital cost per frontal area. The optimized air contactor design achieved a CO2 capture flux of 10,440 kg/m2-yr at a cost of $ 0.068 per kg CO2 captured. The developed optimization framework maximizes contactor performance while decreasing operational costs. A multi-objective genetic algorithm approach was used to identify the trade-offs between the two competing objectives of maximized CO2 capture and minimized cost by expanding the formulation. The Pareto-optimal front of the optimization results offered numerous optimized solutions with specific DAC system setups. This methodology, as well as the optimum configurations identified, serve as the foundation for constructing larger-scale DAC systems, in future. Ongoing research aims to improve solution approaches by including advanced optimization software packages such as DAKOTA, which offers global optimization algorithms, as well as to boost model accuracy via integrating additional choice variables. Future optimization and modeling studies will include contactor interactions and the modeling of a full DAC facility to boost confidence that a globally optimal design has been established.Overall, this study illustrated an efficient optimization strategy for air contactor design and contributes to the improvement of sustainable carbon capture technology, which is crucial to controlling rising global CO2 levels. Contactor configuration optimization enables superior DAC system designs at lower costs, hence promoting the urgent deployment of negative emissions solutions.

Seismic Design of Structures by Sequential Quadratic Programming with Trust Region Strategy and Endurance Time Method

Seismic Design of Structures by Sequential Quadratic Programming with Trust Region Strategy and Endurance Time Method

A two-timescale neurodynamic approach to robust distributed model predictive control for nonlinear systems

In this paper, a robust distributed model predictive control strategy is proposed for nonlinear dynamical systems affected by bounded disturbances. Grounded on sequential quadratic programming, a two-timescale recurrent neural network paradigm is deployed for solving minimax optimization problems. Sufficient criteria for the two-timescale recurrent neural network are delineated, which specifically guarantee its stability and optimality. Under these conditions, it is demonstrated that the recurrent neural network converges to optimal solutions, thereby enhancing the robustness of the control strategy. The simulation results exhibit significantly enhanced convergence in comparison to the mono-timescale recurrent neural network.

Delay aware resource allocation in ORAN through network optimization

AbstractA multi variable resource allocation problem is investigated in network environments, specifically focusing on the consideration of quality of service in open radio access network. The main objective is to minimise the combined latency of various servers while complying with network limitations. The delay of each server is represented by a non‐linear function that has exponentially based. This characteristic inherently brings non‐convexity into the objective function. In contrast, the constraints comprise various linear combinations of network variables, including resource block allocations, power consumption, and number of virtual machines. The purpose of these constraints is to guarantee that the allocation of resources adheres to practical limitations and upholds fairness among servers. Nevertheless, the inclusion of a non‐convex objective function significantly adds complexity to the optimisation problem and non‐convex behaviour, requiring specialised algorithms and techniques to identify solutions. Subsequently, the Lagrange multiplier method has been used to solve this problem mathematically. Numerically, three algorithms have been utilised and compared to solve the problem, these are active‐set, interior point and sequential quadratic programming. Note that the total delay as an objective function is based on the total power consumption of the servers. Previous to optimising the total delay, a delay model is proposed and compared with two research works that are based on experimental and real time data. The proposed model shows data matching with the other works and permits for more adaptation/integration with any other works that uses different servers’ characteristics and network parameters.

Bio-inspired algorithm integrated with sequential quadratic programming to analyze the dynamics of hepatitis B virus

BackgroundThere are a variety of lethal infectious diseases that are seriously affecting people's lives worldwide, particularly in developing countries. Hepatitis B, a fatal liver disease, is a contagious disease spreading globally. In this paper, a new hybrid approach of feed forward neural networks is considered to investigate aspects of the SEACTR (susceptible, exposed, acutely infected, chronically infected, treated, and recovered) transmission model of hepatitis B virus disease (HBVD). The combination of genetic algorithms and sequential quadratic programming, namely CGASQP, is applied, where genetic algorithm (GA) is used as the main optimization algorithm and sequential quadratic programming (SQP) is used as a fast-searching algorithm to fine-tune the outcomes obtained by GA. Considering the nature of HBVD, the whole population is divided into six compartments. An activation function based on mean square errors (MSEs) is constructed for the best performance of CGASQP using proposed model.ResultsThe solution's confidence is boosted through comparative analysis with reference to the Adam numerical approach. The results revealed that approximated results of CGASQP overlapped the reference approach up to 3–9 decimal places. The convergence, resilience, and stability characteristics are explored through mean absolute deviation (MAD), Theil’s coefficient (TIC), and root mean square error (RMSE), as well as minimum, semi-interquartile range, and median values with respect to time for the nonlinear proposed model. Most of these values lie around 10−10–10−4 for all classes of the model.ConclusionThe results are extremely encouraging and indicate that the CGASQP framework is very effective and highly feasible for implementation. In addition to excellent reliability and level of precision, the developed CGASQP technique also stands out for its simplicity, wider applicability, and flexibility.

Stochastic neuro-swarming intelligence paradigm for the analysis of magneto-hydrodynamic Prandtl–Eyring fluid flow with diffusive magnetic layers effect over an elongated surface

In recent years, the integration of stochastic techniques, especially those based on artificial neural networks, has emerged as a pivotal advancement in the field of computational fluid dynamics. These techniques offer a powerful framework for the analysis of complex fluid flow phenomena and address the uncertainties inherent in fluid dynamics systems. Following this trend, the current investigation portrays the design and construction of an important technique named swarming optimized neuro-heuristic intelligence with the competency of artificial neural networks to analyze nonlinear viscoelastic magneto-hydrodynamic Prandtl–Eyring fluid flow model, with diffusive magnetic layers effect along an extended sheet. The currently designed computational technique is established using inverse multiquadric radial basis activation function through the hybridization of a well-known global searching technique of particle swarm optimization and sequential quadratic programming, a technique capable of rapid convergence locally. The most appropriate scaling group involved transformations that are implemented on governing equations of the suggested fluidic model to convert it from a system of nonlinear partial differential equations into a dimensionless form of a third-order nonlinear ordinary differential equation. The transformed/reduced fluid flow model is solved for sundry variations of physical quantities using the designed scheme and outcomes are matched consistently with Adam's numerical technique with negligible magnitude of absolute errors and mean square errors. Moreover, it is revealed that the velocity of the fluid depreciates in the presence of a strong magnetic field effect. The efficacy of the designed solver is depicted evidently through rigorous statistical observations via exhaustive numerical experimentation of the fluidic problem.

Local quadratic convergence of the SQP method for an optimal control problem governed by a regularized fracture propagation model

We prove local quadratic convergence of the sequential quadratic programming (SQP) method for an optimal control problem of tracking type governed by one time step of the Euler-Lagrange equation of a time discrete regularized fracture or damage energy minimization problem. This lower-level energy minimization problem % describing the fracture process contains a penalization term for violation of the irreversibility condition in the fracture growth process and a viscous regularization term. Conditions on the latter, corresponding to a time step restriction, guarantee strict convexity and thus unique solvability of the Euler Lagrange equations. Nonetheless, these are quasilinear and the control problem is nonconvex. For the convergence proof with $L^\infty$ localization of the SQP-method, we follow the approach from \cite{Troeltzsch:1999}, utilizing strong regularity of generalized equations and arguments from \cite{HoppeNeitzel:2020} for $L^2$-localization. 2020 Mathematics Subject Classification 90C55, 49M41, 49M15.

Life-cycle production optimization with nonlinear constraints using a least-squares support-vector regression proxy

When nonlinear constraints such as field liquid or water production rate, injection pressures, etc., as functions of time need to be honored in addition to linear ones, the life-cycle production optimization problem, a component of a closed loop reservoir management, becomes challenging and computationally expensive to perform using a high-fidelity reservoir simulator with the existing gradient-based methods using the adjoint or stochastic approximate gradient methods. Therefore, the objective of this study is to present computationally efficient methods for deterministic production optimization under nonlinear constraints using a kernel-based machine learning method where the cost function is the net present value (NPV). We use the least-squares support-vector regression (LSSVR) to approximate the NPV function. To achieve computational efficiency, we generate a set of output values of the NPV and nonlinear constraint functions, which are field liquid production rate (FLPR) and water production rate (FWPR) in this study, by running the high-fidelity simulator for a broad set of input design variables (well controls) and then using the collection of input/output data to train LSSVR proxy models to replace the high-fidelity simulator to compute NPV and nonlinear state constraint functions during iterations of sequential quadratic programming (SQP). To obtain improved (higher) estimated optimal NPV values, we use the existing so-called iterative sampling refinement (ISR) method to update the LSSVR proxy so that the updated proxy remains predictive toward promising regions of search space during the optimization. Direct and indirect ways of constructing LSSVR-based NPVs as well as different combinations of input data, including nonlinear state constraints and/or the bottomhole pressures (BHPs) and water injection rates, are tested as alternative feature spaces. The results obtained from our proposed LSSVR-based optimization methods are compared with those obtained from our in-house stochastic simplex approximate gradient (StoSAG)-based line-search SQP programming (LS-SQP-StoSAG) algorithm that uses directly a high-fidelity simulator to compute the gradients of the objective function and the nonlinear state functions with StoSAG for the Brugge reservoir model. The results show that nonlinear constrained optimization with the LSSVR ISR with SQP is computationally 3.25 fold more efficient than LS-SQP-StoSAG. In addition, the results show that constructing NPV indirectly using the field liquid and water rates for a waterflooding problem where inputs come from LSSVR proxies of the nonlinear state constraints requires significantly fewer training samples than the method constructing NPV directly from the NPVs computed from a high-fidelity simulator. LSSVR has advantages for its computational efficiency which is the main goal in our research, and robustness against overfitting especially for the cases where we have limited data. However, DNNs and random forest require large training size and require much more computational resources and longer training times. Also, Random Forests and Gradient Boosting Machines can be prone to overfitting and become computationally intensive.

An Obstacle Avoidance Trajectory Planning Methodology Based on Energy Minimization (OTPEM) for the Tilt-Wing eVTOL in the Takeoff Phase

Electric tilt-wing flying cars are an efficient, economical, and environmentally friendly solution to urban traffic congestion and travel efficiency issues. This article addresses the high energy consumption and obstacle interference during the takeoff phase of the tilt-wing eVTOL (electric Vertical Takeoff and Landing), proposing a trajectory planning method based on energy minimization and obstacle avoidance. Firstly, based on the dynamics analysis, the relationship between energy consumption, spatial trajectory, and obstacles is sorted out and the decision variables for the trajectory planning problem with obstacle avoidance are determined. Secondly, based on the power discretization during the takeoff phase, the energy minimization objective function is established and the constraints of performance limitations and spatial obstacles are derived. Thirdly, by integrating the optimization model with the SLSQP (Sequential Least Squares Quadratic Programming algorithm), the second-order sequential quadratic programming model and decision variable update equations are derived, establishing the solution process for the trajectory planning problem of the tilt-wing eVTOL takeoff with obstacle avoidance. Finally, the Airbus Vahana A3 is taken as an example to verify and validate the effectiveness, stability, and robustness of the model and optimization algorithm proposed. The validation results show that the OTPEM (obstacle avoidance trajectory planning methodology based on energy minimization) can effectively handle changes in the takeoff end state and exhibits good stability and robustness in different obstacle environments. It can provide a certain reference for the three-dimensional obstacle avoidance trajectory planning of Airbus Vahana A3 and other tilt-wing eVTOL trajectory planning problems.

Optimization on frequency constraints with FFT using automatic differentiation on hybrid ODE applications

PurposeThis study aims to optimize electrical systems represented by ordinary differential equations and events, using their frequency spectrum is an important purpose for designers, especially to calculate harmonics.Design/methodology/approachThis paper presents a methodology to achieve this, by using a gradient-based optimization algorithm. The paper proposes to use a time simulation of the electrical system, and then to compute its frequency spectrum in the optimization loop.FindingsThe paper shows how to proceed efficiently to compute the frequency spectrum of an electrical system to include it in an optimization loop. Derivatives of the frequency spectrum such as the optimization inputs can also be calculated. This is possible even if the sized system behavior cannot be defined a priori, e.g. when there are static converters or electrical devices with natural switching.Originality/valueUsing an efficient sequential quadratic programming optimizer, automatic differentiation is used to compute the model gradients. Frequency spectrum derivatives with respect to the optimization inputs are calculated by an analytical formula. The methodology uses a “white-box” approach so that automatic differentiation and the differential equations simulator can be used, unlike most state-of-the-art simulators.

Stabilized SQP Methods in Hilbert Spaces

Based on techniques by (S.J. Wright 1998) for finite-dimensional optimization, we investigate a stabilized sequential quadratic programming method for nonlinear optimization problems in infinite-dimensional Hilbert spaces. The method is shown to achieve fast local convergence even in the absence of a constraint qualification, generalizing the results obtained by (S.J. Wright 1998 and W.W. Hager 1999) in finite dimensions to this broader setting.

Design and Research of a Strain Elastic Element with a Double-Layer Cross Floating Beam for Strain Gauge Wireless Rotating Dynamometers.

Cutting force is one of the most basic signals that can reflect the information of the cutting process, so it is very necessary to study the strain elastic element of strain gauge wireless rotating dynamometers. This paper proposes a strain elastic element with a double-layer cross floating beam that can be applied to the strain gauge wireless rotating dynamometer, which can simultaneously obtain the four-component cutting force/torque information of FX, FY, FZ, and MZ. Based on the proposed strain elastic element, a compact strain gauge wireless rotating dynamometer is designed, which is composed of a tool holder, upper connection flange, strain elastic element, lower connection flange, tool base, and data acquisition and wireless transmission system. The static model of the double-layer cross floating beam on the strain elastic element is established by the segmented rigid body method, and the relationships between the material, force, structural parameters, and the strain and deformation of the floating beam are obtained. The static model is consistent with the finite element solution, which proves the rationality of the static model. Based on the established static model, the sequential quadratic programming algorithm is used to optimize the structural parameters of the double-layer cross floating beam to maximize the sensitivity of the floating beam. The overall structure of the strain elastic element is analyzed by finite element software, and the strain of the structure under simulation conditions is obtained, which provides a reference for subsequent calibration tests and circuit design. The calibration matrix and dynamic performance of the strain elastic element are obtained by the static calibration test, dynamic calibration test, and cutting test. The results show that the proposed strain elastic element has high sensitivity and low cross-sensitivity error, and can be applied to the strain gauge wireless rotating dynamometer to measure medium- and low-speed cutting forces.

An improved sequential quadratic programming method for identifying the total heat exchange factor of reheating furnace

Accurate identification of the total heat exchange factor (THEF), which serves as the foundation for online control data in the reheating furnace, holds significant importance for enhancing the digitization and intelligence level of the furnace, mitigating environmental pollution, and improving the economic performance of enterprises. For this purpose, based on the theory of inverse heat transfer problem, a solver combining Sequential Quadratic Programming and Broyden Combined Method (SQPBC) was developed to accurately identify high-dimensional and strongly nonlinear THEF in the reheating furnace. Integrating the efficient Broyden Combined Method (BCM) into Sequential Quadratic Programming (SQP) enables the rapid and accurate estimation of the Jacobian matrix to effectively enhance computational speed without compromising accuracy. Experimental data has been employed to validate the accuracy of the inversion results. The performance of SQPBC has been comprehensively analyzed through a series of numerical experiments, with a particular focus on investigating the impact of sampling frequency and sensor location on the inversion results. The findings reveal that after reducing the sampling period to 2 min, further decreasing the sampling period has a minor effect on the average relative error of the identification results; the closer the measurement points are to the surface of the slab, the more accurate the identification of THEF on the slab's surface.

An Effective Strategy for Achieving Economic Reliability by Optimal Coordination of Hybrid Thermal–Wind–EV System in a Deregulated System

This paper describes an effective operating strategy for electric vehicles (EVs) in a hybrid facility that leverages renewable energy sources. The method is to enhance the profit of the wind–thermal–EV hybrid plant while maintaining the grid frequency (fPG) and energy level of the EV battery storage system. In a renewable-associated power network, renewable energy producers must submit power supply proposals to the system operator at least one day before operations begin. The market managers then combine the power plans for the next several days based on bids from both power providers and distributors. However, due to the unpredictable nature of renewable resources, the electrical system cannot exactly adhere to the predefined power supply criteria. When true and estimated renewable power generation diverges, the electrical system may experience an excess or shortage of electricity. If there is a disparity between true and estimated wind power (TWP, EWP), the EV plant operates to minimize this variation. This lowers the costs associated with the discrepancy between actual and projected wind speeds (TWS, EWS). The proposed method effectively reduces the uncertainty associated with wind generation while being economically feasible, which is especially important in a deregulated power market. This study proposes four separate energy levels for an EV battery storage system (EEV,max, EEV,opt, EEV,low, and EEV,min) to increase system profit and revenue, which is unique to this work. The optimum operating of these EV battery energy levels is determined by the present electric grid frequency and the condition of TWP and EWP. The proposed approach is tested on a modified IEEE 30 bus system and compared to an existing strategy to demonstrate its effectiveness and superiority. The entire work was completed using the optimization technique called sequential quadratic programming (SQP).

Revealing the Invisible: Imaging Through Non-Radiating Subspace

This paper presents a novel modification to the subspace optimization method (SOM) for solving inverse scattering problems in diverse background media. By incorporating sequential quadratic programming (SQP), a fast and accurate optimization technique, our approach aims to improve the reconstruction of dielectric objects buried in complex environments. We investigate the influence of non-radiating (NR) subspace reconstruction on imaging quality by analyzing the induced current-exterior field mapping operator's singular values. The scattering formulation of direct problems is performed through numerical methods such as coupled dipole method, finite elements-boundary integral, and electric field integral equations. Radiating and non-radiating objective functions are defined and minimized using SQP to evaluate the effectiveness of the proposed method. Numerical experiments demonstrate significant enhancements in imaging quality, robustness, and computational efficiency compared to existing techniques. This modified SOM offers a promising avenue for precise and reliable reconstruction of electromagnetic scattering objects, with potential applications in various fields including medical imaging, remote sensing, and non-destructive evaluation.

Multiphoton parallel transmission (MP-pTx): Pulse design methods and numerical validation.

We propose and evaluate multiphoton parallel transmission (MP-pTx) to mitigate flip angle inhomogeneities in high-field MRI. MP-pTx is an excitation method that utilizes a single, conventional birdcage coil supplemented with low-frequency (kHz) irradiation from a multichannel shim array and/or gradient channels. SAR analysis is simplified to that of a conventional birdcage coil, because only the radiofrequency (RF) field from the birdcage coil produces significant SAR. MP-pTx employs an off-resonance RF pulse from a conventional birdcage coil supplemented with oscillating -directed fields from a multichannel shim array and/or the gradient coils. We simulate the ability of MP-pTx to create uniform nonselective brain excitations at 7 T using realistic and field maps. The RF, shim array, and gradient waveform's amplitudes and phases are optimized using a genetic algorithm followed by sequential quadratic programming. A 1 ms MP-pTx excitation using a 32-channel shim array with current constrained to less than 50 Amp-turns reduced the transverse magnetization's normalized root-mean-squared error from 29% for a conventional birdcage excitation to 6.6% and was nearly 40% better than a 1 ms birdcage coil 5 kT-point excitation with optimized kT-point locations and comparable pulse power. The MP-pTx method resembles conventional pTx in its goals and approach but replaces the parallel RF channels with cheaper, low-frequency shim channels. The method mitigates high-field flip angle inhomogeneities to a level better than 3 T CP-mode and comparable to 7 T pTx while retaining the straightforward SAR characteristics of conventional birdcage excitations, as low-frequency shim array fields produce negligible SAR.

Application of ANN control algorithm for optimizing performance of a hybrid ORC power plant

Hybrid systems for generating electricity from multiple sources are becoming an increasingly popular subject of analysis in science and industry. This paper presents a validated model of a hybrid ORC plant powered by solar and geothermal energy. A key challenge in optimizing the operating parameters over time was the variability of solar conditions, which was the main energy source of the system. The operation of the ORC plant is simulated using a complex model with Multiple Input Multiple Output (MIMO) variables, which is nonlinear. The input variables represent the system’s operational parameters, while the output variables describe the plant’s performance indicators. The main objective of this paper is to optimize the year-round performance of the ORC installation through different computational techniques. The first approach involves the application of the gradient-based optimization method that is known as sequential quadratic programming (SQP). With the use of SQP, two distinct simulation runs (SQP-N and SQP-Q/N) of the system are performed, each with a specific objective function to be optimized. The second approach is based on reinforcement learning principles and leverages the method known as Deep Deterministic Policy Gradient (DDPG) algorithm. The main advantage of DDPG over SQP is that DDPG does not require knowledge of the model. This improves the algorithm flexibility, making it well-adapted to fluctuating environmental conditions. Overall, three optimization runs (two using SQP, one using DDPG) have been performed, aiming at identifying the optimal year-round control strategy for the installation. The results revealed that under the control of DDPG, the hybrid system has produced the highest amount of electricity (4993.4 MWh), outperforming in this matter SQP-N and SQP-Q/N optimization variants by 16.83 % and 10.49%, respectively.