“…where ε is a vector for the minimum values of x var , and δ is a hyper-parameter of DeSearTAM to control search interval. After the evaluations of X i , the ranking of the candidate solutions is computed, where MCR-mod [38] is employed to handle (19) and the constraints of the given design problem.…”
Section: Parameter Optimization Methodsmentioning
confidence: 99%
“…Numerical design optimizations are one of the effective and helpful techniques to obtain acceptable designs with a relatively little effort [17], [18]. Thus, various optimization approaches for electric machines have been being studied such as parameter optimizations [19]- [22] and topology optimizations [23]- [27]. In these optimizations, design problems are generally expressed as the following optimization problem: (…”
The design problems of electric machines are actually treated as a kind of mixed-integer problem, because the machine shapes are defined by some integer variables, such as number of slots, and the other variables, such as the tooth width, which are here called the fundamental and shape variables, respectively. To automatically solve these design problems, this article presents an automatic design method by combining the reinforcement learning and evolutionary optimization. In the proposed method, the decision process is modeled as a tree structure to seek for the fundamental variables, which are determined as a result of the tree search depending on the value functions of the nodes. Then, the shape variables are estimated from the function of the fundamental variables. These functions are constructed based on the design data, to generate which the reinforcement learning and evolutionary optimization are employed. As a result, the proposed method can automatically be adapted to unexperienced design problems through the data generation and function learning. The proposed method is applied to a design problem of a linear induction motor. It is shown that the machine designs with the prescribed performance for given specifications are automatically obtained. Moreover, it is also shown that the acceptable candidate designs can immediately be generated when the given specification is similar to the previously-solved problems by utilizing the design data generated by the past explorations.
“…where ε is a vector for the minimum values of x var , and δ is a hyper-parameter of DeSearTAM to control search interval. After the evaluations of X i , the ranking of the candidate solutions is computed, where MCR-mod [38] is employed to handle (19) and the constraints of the given design problem.…”
Section: Parameter Optimization Methodsmentioning
confidence: 99%
“…Numerical design optimizations are one of the effective and helpful techniques to obtain acceptable designs with a relatively little effort [17], [18]. Thus, various optimization approaches for electric machines have been being studied such as parameter optimizations [19]- [22] and topology optimizations [23]- [27]. In these optimizations, design problems are generally expressed as the following optimization problem: (…”
The design problems of electric machines are actually treated as a kind of mixed-integer problem, because the machine shapes are defined by some integer variables, such as number of slots, and the other variables, such as the tooth width, which are here called the fundamental and shape variables, respectively. To automatically solve these design problems, this article presents an automatic design method by combining the reinforcement learning and evolutionary optimization. In the proposed method, the decision process is modeled as a tree structure to seek for the fundamental variables, which are determined as a result of the tree search depending on the value functions of the nodes. Then, the shape variables are estimated from the function of the fundamental variables. These functions are constructed based on the design data, to generate which the reinforcement learning and evolutionary optimization are employed. As a result, the proposed method can automatically be adapted to unexperienced design problems through the data generation and function learning. The proposed method is applied to a design problem of a linear induction motor. It is shown that the machine designs with the prescribed performance for given specifications are automatically obtained. Moreover, it is also shown that the acceptable candidate designs can immediately be generated when the given specification is similar to the previously-solved problems by utilizing the design data generated by the past explorations.
“…Similarly, multiobjective optimization has been used to help in the design of specific products or to support generic design methods. For instance, Bazzo et al (2016) proposed a bi-objective optimization model for the design of permanent magnet synchronous generators based on cost and lifetime energy production. Other studies have used multiobjective optimization for the design of specific products such as engine crankshaft bearings (Sun et al, 2016).…”
Section: Literature Review and Paper Contributionmentioning
Design is a process through which customer needs are transformed into product or service specifications, and then used to develop a model or prototype. The prototype is tested, and modifications are brought to it before the production process starts. Moreover, the design process may be divided into different stages, starting from the definition of the customer needs, going through the conceptual design phase and ending up with the detailed design. In this article, we address the conceptual design phase, where the customer needs are assumed to be known. The proposed approach considers, based on customer needs, primary and secondary design criteria. Each design criterion has a set of predetermined possible values (options) from which the designer may select. Making the best selection of all the design features while satisfying the customer needs in terms of cost, quality (customer preference) and environmental performance is a combinatorial problem and therefore a decision-making framework would be helpful for the designers. In this article, the design criteria are evaluated using fuzzy technique for order preference by similarities to ideal solution based on cost, quality and environmental sustainability. A multiobjective and a single-objective binary programming models are then developed and solved, and their optimal solutions are obtained. The multiobjective solutions provide the decision makers with the possible trade-offs, whereas the single-objective model solution can be used as a final decision-making tool. The proposed approach is implemented in a user-friendly software developed by the authors. A case study is conducted using a baby car seat for which three main and six secondary design criteria are considered. The obtained results show the effectiveness of the approach used.
“…In a more general sense, the recent reviews in [12,13] on optimisation trends for electric machines highlighted that the use of genetic algorithms for multi-objective design optimisation approaches is favourable for design problems with several performance measures. Moreover, multi-objective design optimisation is the most common approach for wind generator design, e.g., [14][15][16][17]. In [14][15][16][17], the actively controlled wind generators with maximum power point tracking (MPPT) were optimised for a specified drive-cycle to maximise annual energy production.…”
Small-scale uncontrolled passive wind generator systems are an attractive solution for rural energy generation because of the system’s reliability and low cost. However, designing these uncontrolled wind generators for good power matching with the wind turbine is challenging and often requires external impedance matching. In this paper, permanent magnet generators with different stator and rotor structures were investigated and designed to increase the generator’s synchronous inductance for a natural impedance matching. For the design methodology, multi-objective optimisation was used to design the generators for near-maximum turbine power matching, whereby internal impedance matching was reached as much as possible. It was shown that altering the placement and orientation of the permanent magnets in the rotor is a viable method to achieve the desired impedance matching; however, these generators do not have the best performance. It was found that the surface-mounted permanent magnet generator with semi-closed slots was the optimum topology. An optimised generator prototype was tested for the experimental validation. All designs were verified by comparing the results of 2D and 3D finite-element analysis.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.