A hybrid rule-based and data-driven approach to product design

Hybrid Modelling Engineering Procedures?for Ultimate Reliability and Accuracy on Product Design

In most application fields in the engineering, physical models are expressed as a set of governing constitutive relations, spatio-temporal relations, and/or dynamical systems. Data-driven discovery in these application areas are specifically constructed to discover constitutive relations and differential governing equations in a wide variety of fields, conservation laws or propagation of nonlinear waves , spatio-temporal dynamics , development of predictive procedures for structural dynamics for nanoscale flow. In particular, data-driven techniques may be extremely important in complex areas of engineering allowing for unveiling unknown mechanisms. Thus, there are increasingly funding initiatives to obtain new methods, software tools and training within the framework of Big Data analysis.

A hybrid rule-based and data-driven approach to?product design

The difficulties in considering all aspects of engineering modeling and science with data-driven procedures, and the interest in taking advantage of the learnings from the classical approach, is currently motivating a mixed approach in which data-driven modeling is guided by some physical insigh. The purpose of developing mixed approaches is to improve the reliability of the obtained relations through fundamental principles, like conservation laws .

A hybrid rule-based and data-driven approach to product design

Data-driven procedures focus on data and try to extract variables and relations directly from raw data, giving frequently more accurate responses without the use of classical analytical laws and equations. However, many open questions remain, and in some occasions, drawbacks have been found as the lack of fulfillment of some physical principles. Then, new physics-based data-driven procedures are getting in.

Implementation of quantum hybrid models

The reliability-based multi-design point methodology is proposed to acquire the appropriate key design parameters by comprehensive reliability analysis for multiple operating conditions of interest. This methodology establishes corresponding thermodynamics-based simulation model of aircraft engine with uncertainty component performance; It divides the whole flight profile into multiple critical operating conditions and simultaneously regard them as the on-design points, and then respectively determine their overall performance requirements;?

This methodology generates a certain number of training scenarios and train the neural network for establishing reliability prediction surrogate models of each concerned operating condition; It constructs objective function by the attained surrogate models to facilitate the comprehensive performance reliability analysis in optimization design.

Advantages of using quantum hybrid models

The proposed design method is applied on the cycle design of the propulsion plant performance model could empower propulsion plant designers, with .Hybrid quantum algorithm which integrates the Accelerated Optimization algorithm, based pre-training technique into the network training procedure.

It respectively reduced the average error and maximum error of Artificial Neural Network. Utilization of the ANN surrogate models facilitate the reliability-based cycle design optimization, which replaces the time-consuming probabilistic analysis based on Monte Carlo simulationOptimization design solution of presented methodology reasonably increases the aero engine performance redundancy to precisely reach the expected reliability of all concerned operating conditions.This methodology is universal and can be easily applied to other types of propulsion plant. It facilitates the application of the reliability-based multi-design point method for candidates of future aviation propulsion, which are noted for excellent multi-mission adaptability.

Challenges of using quantum hybrid model

Novel aero engine mechanical structure becomes more complex, the following overall reliability problems would undoubtedly be even worse than that of the conventional gas turbine engine. Benefits of using quantum hybrid models aero engine conceptual design is already an extremely complicated problem, which involves the interaction of each component and the coupling of multiple disciplines. When the existence of uncertainty factors cannot be ignored, solving this problem becomes more difficult. Therefore, the traditional conceptual design method is facing challenges and it is worthwhile devoting much effort to this. We have to bring computing solutions to a problem involving airframe loads, mass modelling and structural analysis. Our target should be preserving structural integrity while optimising weight. Weight optimisation is key to low operating costs and reduced environmental impact. Structural integrity can be demonstrated by simulating key flight occurrences through life cycles of design, required by air worthiness regulations.

?Further remarks using quantum hybrid models

This methodology addresses the limitation of traditional deterministic design method that determines the key design parameters by subjectively setting the performance redundancy. Therefore, the overall performance redundancy could be set at a reasonable level so that contributes to the technical risk management and cost control of aero engine manufacturing.

Applications

Autonomous vehicles need to model the behavior of surrounding human driven vehicles to be safe and efficient traffic participants. Existing approaches to modeling human driving behavior have relied on both data-driven and rule-based methods. While data-driven models are more expressive, rule-based models are interpretable, which is an important requirement for safety-critical domains like driving. However, rule-based models are not sufficiently representative of data, and data-driven models are yet unable to generate realistic traffic simulation due to unrealistic driving behavior such as collisions. Methedology proposed combines rule-based modeling with data-driven learning. While the rules are governed by interpretable parameters of the driver model, these parameters are learned online from driving demonstration data using particle filtering. Results shown that driver models based on hybrid rule-based and data-driven approach can accurately capture real-world driving behavior.

Hybrid Additive Manufacturing

Additive manufacturing (AM) suffers from residual stress formation and part distortion. One way to improve AM is to use optimal process parameters that yield residual stress distribution that is less detrimental. A finite element analysis (FEA) based model can be used as a design tool to determine optimal process parameters. Being a design tool, such a model should not only predict the residual stress formation and part distortion accurately but should be capable of performing the simulation in a reasonable time as well. However, the traditional?FEA?models are either accurate but computationally very expensive as in the case of process-based models that captures the evolution of the temperature and stress/strain field or very fast but often inaccurate as in the case of inherent strain based (IST) based models which comprise a series of quasi-static stress-analyses. In this study, we demonstrate that a model that is both accurate and computationally less expensive can be developed by adopting a hybrid approach in which the spatial and temporal resolution of the field solutions over different regions of the part are selectively controlled depending on the expected levels of stress gradients in these regions. The new model, called Additive FEA Hybrid (AFEA-H) model, uses a previously reported Additive FEA (AFEA) framework?to combine a process-based model that has a high spatial and temporal resolution with an IST based model that has a low resolution. Numerical simulations performed for a few example problems demonstrate that the AFEA-H method can enhance the computational efficiency substantially without compromising solution accuracy.