Efficient Design Research Articles

Building performance optimization through sensitivity Analysis, and economic insights using AI

Optimizing building designs for energy efficiency and occupant comfort presents significant challenges due to the complex and often conflicting requirements of various stakeholders. Consequently, this study conducts a multifaceted sensitivity and economic impact analysis that aims to improve building performance in terms of energy efficiency and occupant comfort by implementing machine learning techniques. Using a broad dataset comprising of 12,000 energy simulation runs for Tvedestrand Upper Secondary School in Norway, several machine learning models were employed with Multi-Layer Perceptron outperforming the others. In addition, several sensitivity analysis methods were used to explore the influence of individual parameters on building performance. The analysis reveals that ventilation rate, room depth, U-value of the facade, and heat gains significantly affect energy consumption. Economic impact analysis was also carried out to compare the cost-effectiveness of traditional HVAC systems with Building Management System (BMS) HVAC solutions. The BMS HVAC system shows significantly lower operational costs over time, with investment costs averaging around 1200 Norwegian kroner (NOK)/m2 and operational costs of approximately 150 NOK/m2 per year. Sensitivity analysis under different economic scenarios highlights the economic viability of the BMS HVAC system. This study identifies optimal building parameters that balance energy efficiency and thermal comfort, achieving total energy consumption between 11.05 and 22.51 kWh/m2 and zero discomfort hours (h > 26 °C). In sum, the findings offer valuable insights for stakeholders, enabling informed decisions about sustainable building design and energy efficiency improvements, ensuring both technical soundness and financial viability under a wide range of conditions, while using the tested tools.

Achieving 34 % efficiency with a dual-absorber solar cell design using CaRbCl3 and Ca3NCl3 perovskites

This study’s DFT calculations show that the direct bandgaps of perovskite CaRbCl3 and Ca3NCl3 are 1.386 eV and 1.430 eV, respectively, confirming their semiconducting nature. The study also provides density of states and dielectric constant data for these materials. In photovoltaic (PV) technology, perovskites are highly valued as solar absorber materials for their exceptional optical properties, low cost, high efficiency, and lightweight design. This study presents a structure (Al/FTO/CdS/dual absorber (CaRbCl3/Ca3NCl3)/V2O5/Au) incorporating a CdS electron transport layer (ETL) and a V2O5 hole transport layer (HTL) with a double perovskite absorber layer (DPAL) of CaRbCl3 and Ca3NCl3 using SCAPS-1D. The results show that the DPAL-based perovskite solar cell (PSC) outperforms single-layer PSCs, with significant efficiency gains when HTL is included. The study thoroughly examines the effects of layer thickness, doping levels, and defect densities on key electrical parameters such as VOC, JSC, FF, and PCE. It also explores J-V and QE characteristics and the impact of temperature. The ideal absorber thickness for achieving the highest PCE is 500 nm for CaRbCl3 and 600 nm for Ca3NCl3, combined with a doping density of 1 × 1017 cm−3, a defect density of 1 × 1011 cm−3 and interface defect density of 1 × 1010 cm−2. While single-absorber solar cells achieve maximum efficiencies of 24.40 % (CaRbCl3) and 24.02 % (Ca3NCl3), the DPAL setup reaches an optimized efficiency of 27.66 %, further increasing to 34.39 % with a VOC of 1.295 V, FF of 83.95 %, and JSC of 31.63 mA/cm2 when using V2O5 HTL. This research provides valuable insights and a practical strategy for developing cost-effective CaRbCl3/Ca3NCl3 thin-film solar cells.

Enabling scalable and adaptive machine learning training via serverless computing on public cloud

In today’s production machine learning (ML) systems, models are continuously trained, improved, and deployed. ML design and training are becoming a continuous workflow of various tasks that have dynamic resource demands. Serverless computing is an emerging cloud paradigm that provides transparent resource management and scaling for users and has the potential to revolutionize the routine of ML design and training. However, hosting modern ML workflows on existing serverless platforms has non-trivial challenges due to their intrinsic design limitations such as stateless nature, limited communication support across function instances, and limited function execution duration. These limitations result in a lack of an overarching view and adaptation mechanism for training dynamics, and an amplification of existing problems in ML workflows.To address the above challenges, we propose SMLT, an automated, scalable and adaptive serverless framework on public cloud to enable efficient and user-centric ML design and training. SMLT employs an automated and adaptive scheduling mechanism to dynamically optimize the deployment and resource scaling for ML tasks during training. SMLT further enables user-centric ML workflow execution by supporting user-specified training deadline and budget limit. In addition, by providing an end-to-end design, SMLT solves the intrinsic problems in public cloud serverless platforms such as the communication overhead, limited function execution duration, need for repeated initialization, and also provides explicit fault tolerance for ML training. SMLT is open-sourced and compatible with all major ML frameworks. Our experimental evaluation with large, sophisticated modern ML models demonstrates that SMLT outperforms the state-of-the-art VM-based systems and existing public cloud serverless ML training frameworks in both training speed (up to 8×) and monetary cost (up to 3×).

A novel tubular single-chamber microbial electrolysis cell for efficient methane production from industrial potato starch wastewater

The integration of microbial electrolysis cells (MEC) with anaerobic digestion (AD) shows great promise for enhancing methane production from high-COD wastewater. However, an efficient MEC-AD reactor design remains elusive. Here, a novel tubular single-chamber MEC-AD reactor was constructed to treat potato starch wastewater (COD over 20000mg/L). The concentric and compact design of the stainless-steel cathode and anode reduced internal resistance, resulting in enhanced methane production. Applying -0.2V vs. Ag/AgCl to the anode increased methane production by 1.73 times compared to the open circuit and halved hydraulic retention time. Moreover, the reactor achieved an average methane content of 82.57%, which was 23.89% higher than the open circuit. The reactor showed a total COD removal of 92.2%, which was 24% higher than the open circuit. Additionally, base consumption to maintain pH was reduced to one-sixth of that in conventional AD, preventing volatile fatty acid accumulation. Microbial analysis showed Geobacter (63.4%) and Methanobacterium (96.8%) were highly enriched in the anode and cathode biofilms, respectively. The proportion of fermentative bacteria also increased in the MEC-AD system. These results demonstrate the effectiveness of the tubular single-chamber MEC-AD reactor in enhancing methane production from potato starch wastewater, with strong potential for scale-up applications.

NiS2/FeS2 binary nanoparticles confined in interconnected carbon framework with regulated pore structure enabled by citric acid for enhanced sodium storage properties

Recently, pyrite iron disulfide (FeS2) has emerged as a promising anode candidate for sodium-ion batteries (SIBs) due to its high theoretical capacity, affordability, non-toxicity and abundant resource in nature. However, the utilization of FeS2 still confronts the challenges of inferior rate capability and cycling instability for sodium storage, stemming from its low electronic conductivity and substantial volume changes during cycling. Herein, to address these obstacles, NiS2/FeS2 binary nanoparticles encapsulated within a network of interconnected N-doped porous carbon framework (NiS2/FeS2@NPC) are prepared by a successive solid-state ball milling, carbonization and sulfurization strategy with coordination complex of nickel iron Prussian blue analogue (NiFe-PBA) as precursor. The introduction of citric acid plays a critical role for the formation of interconnected carbon framework with regulated internal porous structure, thereby achieving heterostructured NiS2/FeS2@NPC with modulated specific surface area and pore volume. The rational design of interconnected N-doped porous carbon framework and heterogeneous structure guarantees rapid ion/electron transport kinetics, alleviated mechanical stress, and enhanced structural integrity. Benefitting from these advantages, the optimal NiS2/FeS2@NPC-1 electrode exhibits a high reversible capacity (545 mAh/g at 1 A/g), superior rate capability (267 mAh/g at 5 A/g) and ultrastable long-term cycling performance (98.5 % retention over 1000 cycles at 5 A/g). This study presents a novel and efficient design approach for creating durable and high-rate anode materials based on metal sulfides for SIBs.

Automatic design generation of trusses from a reused steel stock library using graphic statics

Reuse of undamaged steel components in new structural design offers a means to reduce construction-related waste and carbon emissions. However, designing with reused members significantly differs from conventional methods as the design must conform to a so-called stock constraint. This work presents an automated stock-constrained design algorithm for planar trusses. To promote high efficiency of the resulting designs, graphic statics is used as the backbone of the algorithm. Additionally, optimization is used to reduce the wasted stock material when determining how to cut stock members to form the truss. To address the need for flexibility in reuse design, the underlying engine has inherent stochasticity and is fast (on the order of 1 min). It is herein executed 100 times per design case and shown to generate efficient design options that can either construct the same truss layout in multiple ways or provide the designer with different layout alternatives to select from. Carbon savings are estimated as 83%–90% for an obtained design case when compared to conventionally designed trusses made from recycled steel.

A Study on Series-Parallel Winding Changeover Circuit and Control Method for Expanding the High-Efficiency Operating Range of IPMSM for xEV Drive Systems

The motor characteristics control method using the winding changeover technique can improve the matching ratio between the most frequent operating point of electric vehicle (EV) and the motor’s high-efficiency operating point, thereby enhancing the overall average efficiency of the drive system. This technology reduces back electromotive force and winding resistance by adjusting the effective number of motor winding turns according to the EV’s operating speed, ultimately improving the average efficiency. In this paper, we propose a winding changeover circuit and control method that maximizes the average efficiency in the main driving regions to extend the driving range per charge and improve the fuel efficiency of EVs. The proposed circuit is constructed using thyristor switching devices, offering the advantage of relatively lower overall system losses compared to conventional circuits. Due to the characteristics of the thyristor switching devices used in the proposed circuit, seamless winding changeover is possible during motor operation. Additionally, no extra snubber circuits are required, and the relatively low switch losses suggest the potential for improved efficiency and lightweight design in EV drive systems. To verify the proposed winding changeover circuit and control scheme, experiments were conducted using a dynamometer with an 80 kW permanent magnet motor, inverter, and the developed prototype of the winding changeover circuit.

Dose Individualization for Phase I Cancer Trials With Broadened Eligibility.

Broadening eligibility criteria in cancer trials has been advocated to represent the intended patient population more accurately. The advantages are clear in terms of generalizability and recruitment, however there are some important considerations in terms of design for efficiency and patient safety. While toxicity may be expected to be homogeneous across these subpopulations, designs should be able to recommend safe and precise doses if subpopulations with different toxicity profiles exist. Dose-finding designs accounting for patient heterogeneity have been proposed, but existing methods assume that the source of heterogeneity is known. We propose a broadened eligibility dose-finding design to address the situation of unknown patient heterogeneity in phase I cancer clinical trials where eligibility is expanded, and multiple eligibility criteria could potentially lead to different optimal doses for patient subgroups. The design offers a two-in-one approach to dose-finding by simultaneously selecting patient criteria that differentiate the maximum tolerated dose (MTD), using stochastic search variable selection, and recommending the subpopulation-specific MTD if needed. Our simulation study compares the proposed design to the naive approach of assuming patient homogeneity and demonstrates favorable operating characteristics across a wide range of scenarios, allocating patients more often to their true MTD during the trial, recommending more than one MTD when needed, and identifying criteria that differentiate the patient population. The proposed design highlights the advantages of adding more variability at an early stage and demonstrates how assuming patient homogeneity can lead to unsafe or sub-therapeutic dose recommendations.

Теплоотдача при движении воды в трубе при температурах, близких к температурам кипения в переходном режиме

Currently, there is a problem of developing effective methods to calculate heat and mass transfer processes during steam condensation from a vapor-gas mixture in industrial devices. This is due to the need to create new reliable and highly efficient designs of heat exchangers for various purposes. The finite element method has been used in the ANSYS Fluent software package during the numerical simulation. An experimental plant is tested in the pipeline of an industrial enterprise at the production site of the Technopolis KHIMGRAD industrial park (Kazan-city). The applicability of the Lee model to solve problems of water flow in a pipeline with partial evaporation is proved. This model allows you to accurately account for the processes of evaporation and condensation, which is important for the design of cooling and air conditioning systems. The authors have designed a three-dimensional model of the flow area of the experimental module and a CFD model to calculate temperatures, phase velocities and their concentrations, considering the peculiarities of the ongoing processes of heat and mass transfer. A slight effect on the heat flow of a 180° rotation during the flow of a liquid medium in the range of the Reynolds numbers 1800–2600 is shown. The developed model is verified with the results obtained at the experimental plant. It is established that the model reproduces experimental data well. An analysis of studies of heat and mass transfer processes during the fluid flow in channels in the presence of phase transitions is carried out. The problems of calculating the parameters of heat transfer during the movement of a fluid in a transient mode are revealed, especially at values of the Reynolds number close to 2000. The experimental plant diagram is presented. It is found that in the range of Reynolds numbers of 2600–3600, the calculated temperature values differ from the temperature values obtained experimentally by less than 0,42 %. During the partial transition of water into steam in the range of Reynolds numbers of 1800–2600, the deviations were less than 6 %, which confirms the adequacy of numerical modeling of calculating the heat transfer process in a pipe. It is shown that considering the influence of temperature on surface tension, the coefficient of thermal conductivity of water and the coefficient of dynamic viscosity of water is important to obtain accurate results. The proposed approach can be used to optimize the operating parameters of condensers and evaporators, as well as to accurately calculate the heat transfer coefficient and determine the areas of vapor formation. It will improve the energy efficiency of industrial installations and reduce the cost of equipment operation.

Photoacoustic Resonators for Non-Invasive Blood Glucose Detection Through Photoacoustic Spectroscopy: A Systematic Review

Diabetes mellitus is a prevalent disease with a rapidly increasing incidence projected worldwide, affecting both industrialized and developing regions. Effective diabetes management requires precise therapeutic strategies, primarily through self-monitoring of blood glucose levels to achieve tight glycemic control, thereby mitigating the risk of severe complications. In recent years, there have been significant advancements in non-invasive techniques for measuring blood glucose using photoacoustic spectroscopy (PAS), as it shows great promise for the detection of glucose using the infrared region (e.g., MIR and NIR) of light. A critical aspect of this method is the detection of the photoacoustic signal generated from blood glucose, which needs to be amplified through a photoacoustic resonator (PAR). In this work, an overview of various types of PARs used for non-invasive glucose sensing is reviewed, highlighting their operating principle, design requirements, limitations, and potential improvements needed to enhance the analysis of photoacoustic signals. The motivation behind this review is to identify and discuss main parameters crucial to the efficient design of PARs used in non-invasive glucose detection, which will be helpful for furthering the basic understanding of this technology and achieving the highly sensitive PAR required for non-invasive glucose monitoring.

A hybrid equivalent circuit model and plane wave spectrum method for decoupling surface designs in multi‐band shared‐aperture antenna arrays

AbstractShared‐aperture antenna arrays are attractive for 5G base stations, owing to their compact size and multi‐band frequency coverage. Decoupling surfaces are effective for suppressing crossband mutual coupling. However, their designs typically rely on full‐wave simulations of scattering parameters, which may not ensure a good radiation performance and is time‐consuming. A systematic methodology based on the hybrid equivalent circuit model and plane wave spectrum method is proposed for decoupling surface designs. The decoupling surface is represented by its equivalent circuit and the spectral expression converts the electromagnetic field into a circuit problem. Both the scattering parameter and radiation pattern can be solved with the circuit, offering an efficient design method that ensures consistent radiation patterns between the model and full‐wave simulations. A double‐layer decoupling surface with wide stopbands, high transmission property, and sharp roll‐off transition was realised. Furthermore, to reduce the overall profile height, the reflection phase of the decoupling surface was tuned. A triple‐band shared‐aperture base station antenna array was developed. Simulated and measured results demonstrate the array works properly in the 0.69–0.96 GHz, 1.7–2.7 GHz, and 3.3–3.8 GHz frequency bands with stable radiation patterns, and crossband mutual coupling is well suppressed, validating the effectiveness of the proposed methodology.

Performance prediction and manipulation strategy of a hybrid system based on tubular solid oxide fuel cell and annular thermoelectric generator

Abstract Tubular solid oxide fuel cells (TSOFCs) generate high-grade waste heat during operation, but the existing waste heat recovery technologies designed for flat solid oxide fuel cells cannot be directly applied to TSOFC due to the geometry mismatch. To efficient harvest the waste heat, a new geometry-matching hybrid system including TSOFC and annular thermoelectric generator (ATEG) is synergistically integrated to evaluate the performance upper limit. A mathematical model is formulated and verified to describe the hybrid system by considering various thermodynamic-electrochemical irreversible effects. Key performance indicators are established to assess the potential performance. Calculations show that the peak power density and corresponding efficiency of the proposed system are enhanced by 20.39 % and 13.89 %, respectively, compared to a standalone TSOFC. Furthermore, the exergy destruction rate is reduced by 7.04 %. Extensive sensitivity analyses indicate that higher operating temperatures enhance the system’s performance, while larger electrode tortuosity negatively affects it. Additionally, various optimization paths of ATEG are explored to improve the system performance, including considerations such as the number of thermocouples, leg radial width, leg thickness, or annular shape parameter. The three-objective optimization yields an efficient design solution for the entire system, offering valuable insights for its design and operation.

Design of high efficiency and wideband dual-functional metasurface for polarization conversion and asymmetric transmission

A wideband and high-efficiency bi-functional metasurface for polarization conversion and asymmetric transmission (AT) based on the gratings/ polarization converter/gratings structure is proposed. The proposed metasurface structure consists of a two double-split ring structure array sandwiched by two orthogonal metallic sub-wavelength gratings. Using two orthogonal metallicsub-wavelength gratings, the proposed metasurface reveals excellent AT effect and nearly perfect polarization conversion performance for linearly polarized waves. The proposed metasurface achieves a polarization conversion ratio (PCR) of above 0.99 from 3.7 GHz to 20.2 GHz with a relative bandwidth (RBW) of 138% and an AT parameter of over 0.9 from 5.8 GHz to 17 GHz withan RBW of 99.5%. Moreover, the proposed metamaterial maintains high PCR and AT performance for a wide range of incident angles. Due to its excellent features, such as the wideband and high efficiency of both PCR and AT performances, the proposed bi-functional metasurface can be useful for promising applications such as polarization imaging. radar, radiometer, and transmission.array antennas.

Development of a Design Procedure Combining Topological Optimization and a Multibody Environment: Application to a Tram Motor Bogie Frame

Nowadays, it is essential to find increasingly rapid and efficient design strategies. This approach becomes crucial in the railway industry, where components must be verified according to multiple reference standards, both structurally and dynamically. In this context, the present research activity aims to develop a fast and effective desin procedure based on European reference standards. The goal was to develop the geometry of a motor bogie frame for a tram vehicle, integrating three fundamental tools for development: finite element simulations and topological structural optimization, a Write Computer Aided Design (CAD) environment, and a multibody environment. Their integration could enhance design accuracy, streamline the traditional design workflow, and support innovation. The optimization process involved the introduction of complex technological constraints, directing the geometry toward production by casting. A tool was developed to automate running dynamics simulations and the output of results for immediate verification of the entire vehicle performance. Finally, the new geometry was tested both structurally and dynamically. The mass was reduced by approximately 7% while ensuring satisfactory mechanical performance. The maximum value of stress was reduced by about 16%. The dynamic performance showed negligible variation, confirming the encouraging outcomes to make this procedure increasingly effective and reliable.

Accelerating spin Hall conductivity predictions via machine learning

AbstractAccurately predicting the spin Hall conductivity (SHC) is crucial for designing novel spintronic devices that leverage the spin Hall effect. First‐principles calculations of SHCs are computationally intensive and unsuitable for quick high‐throughput screening. Here, we have developed a residual crystal graph convolutional neural network (Res‐CGCNN) deep learning model to classify and predict SHCs solely based on the structural and compositional information. This is enabled by having access to 9249 instances of SHCs data and incorporating extra residual networks into the standard CGCNN framework. We found that Res‐CGCNN surpasses CGCNN, achieving a mean absolute error of 115.4 (ℏ/e) (S/cm) for regression and an area under the receiver operating characteristic curve of 0.86 for classification. Additionally, we utilized Res‐CGCNN to conduct high‐throughput screenings of materials in the Materials Project database that were absent in the training set. This led to the prediction of several previously unreported materials displaying large SHCs exceeding 1000 (ℏ/e) (S/cm), which were validated through first‐principles calculations. This study represents the inaugural endeavor to construct a machine learning model capable of effectively capturing the intricate nonlinear relationship between SHCs and crystal structure and composition, serving as a useful tool for the efficient screening and design of materials exhibiting high SHCs.

Mesostructured emerging 316L-17-4PH composite steel parts by material extrusion for the property maximization through an efficient microstructural design

Mesostructured emerging 316L-17-4PH composite steel parts by material extrusion for the property maximization through an efficient microstructural design

Mechanical and Dimensional Stability of Gelatin-Based Hydrogels Through 3D Printing-Facilitated Confined Space Assembly.

Hydrogels have emerged as promising biomaterials for tissue regeneration; yet, their inherent swelling can cause deformation and reduced mechanical properties, posing challenges for practical applications in biomedical engineering. Traditional methods to reduce hydrogel swelling often involve complex synthesis procedures with limited flexibility. Inspired by nature's efficient designs, we present here the approach to improve hydrogel performance using 3D printing-assisted microstructure engineering. By utilizing polymerization-induced phase separation of hydrogel from copolymerization of gelatin methacrylate and hydroxyethyl methacrylate (poly(GelMA-co-HEMA)) in the confined space during vat photopolymerization (VPP) 3D printing, we replicate the cuttlebone-like microstructure of hydrogels with enhanced mechanical properties and swelling resistance. We demonstrate here a 4-fold increase in elastic modulus compared to bulk polymerization of poly(GelMA-co-HEMA), together with improved mechanical and dimensional stability. This method offers promising opportunities for practical biomedical and tissue engineering applications, overcoming previous limitations in the design and performance.

PB-MNSGA-II algorithm and its application for multi-objective optimization design of under-vehicle equipment

Structural damage caused to under-vehicle equipment under adverse working conditions affects the safe operation of high-speed electric multiple units (EMUs). To address the low optimization efficiency and take into account vibration fatigue in traditional under-vehicle equipment design, herein, a multi-objective optimization design method (PB-MNSGA-II) that combines particle swarm optimization-back propagation (PSO-BP) neural network and modified non-dominated sorting Genetic Algorithm II (MNSGA-II) was proposed. It can help engineers choose better equipment component thickness and achieve efficient and accurate design. First, MNSGA-II, the core content of PB-MNSGA-II, was proposed, which improved four aspects of NSGA-II: Latin hypercube sampling population initialization, adaptive crossover and mutation probability, mutation interference of Levy flight strategy, and optimal elite strategy. Then, zero-ductility transition (ZDT) series functions were employed to test MNSGA-II, thereby verifying its convergence and computational efficiency. Subsequently, utilizing the frame structure of the traction transformer (TTR) as an illustrative example, the PB-MNSGA-II approach was employed to enable multi-objective optimization. During this optimization process, the focus was on random vibration fatigue damage, thus enhancing the rationality of the optimization outcomes. After structural optimization, the mass of the TTR decreased by 20.1 kg, and the maximum vibration fatigue damage value decreased from 1.09 to 0.808, verifying the effectiveness and feasibility of the proposed method. The findings of this study provide important insights for the multi-objective optimization design of under-vehicle equipment under complex operating conditions.

Interpretable Learning of Accelerated Aging in Lithium Metal Batteries.

Lithium metal batteries (LMBs) with high energy density are perceived as the most promising candidates to enable long-endurance electrified transportation. However, rapid capacity decay and safety hazards have impeded the practical application of LMBs, where the entangled complex degradation pattern remains a major challenge for efficient battery design and engineering. Here, we present an interpretable framework to learn the accelerated aging of LMBs with a comprehensive data space containing 79 cells varying considerably in battery chemistries and cell parameters. Leveraging only data from the first 10 cycles, this framework accurately predicts the knee points where aging starts to accelerate. Leaning on the framework's interpretability, we further elucidate the critical role of the last 10%-depth discharging on LMB aging rate and propose a universal descriptor based solely on early cycle electrochemical data for rapid evaluation of electrolytes. The machine learning insights also motivate the design of a dual-cutoff discharge protocol, which effectively extends the cycle life of LMBs by a factor of up to 2.8.

Scalable protein design using optimization in a relaxed sequence space.

Machine learning (ML)-based design approaches have advanced the field of de novo protein design, with diffusion-based generative methods increasingly dominating protein design pipelines. Here, we report a "hallucination"-based protein design approach that functions in relaxed sequence space, enabling the efficient design of high-quality protein backbones over multiple scales and with broad scope of application without the need for any form of retraining. We experimentally produced and characterized more than 100 proteins. Three high-resolution crystal structures and two cryo-electron microscopy density maps of designed single-chain proteins comprising up to 1000 amino acids validate the accuracy of the method. Our pipeline can also be used to design synthetic protein-protein interactions, as validated experimentally by a set of protein heterodimers. Relaxed sequence optimization offers attractive performance with respect to designability, scope of applicability for different design problems, and scalability across protein sizes.