Optimal Computing

Unleash the full power of your design workflow

Participation to the UMons Workshop IA

In this application, an IA edge system for intelligent home is developped. The goal is to develop a system capable to identify a user allowed to use other services available inside the house. The 2 services are: a system capable to detect fire or start of fire inside the house, while the second one allows to identify objects inside the house: glasses, wallet, remote control and keys. The system is also capable to send notification when a person is identified or when a fire is detected. Finally, the application was deployed on Jetson Xavier. This was a excellent experience working with 4 team mate in the context of the Workshop IA organized by the university of Mons (UMons) in the Hands On IA program. A unique human and technical experience in a friendly atmosphere. Thanks to Sandrine, Didier, Marcel and Jonathan we got the 'Performance Award' in this challenge

Pump Multi-Disciplinary Optimization

In this application, the goal is to perform the optimization of a radial pump. The goal is to modify 13 parameters of the geometry to increase the pump efficiency while keeping the mechanical stresses below a maximum value and using constraints on the mechanical vibration mode of the pump. The challenges solved in this application are:

We could achieve high robustness of this complex simulation chain
The optimization using 13 parameters in a few design cycles (less than 50 simulations of the whole pump performance)
The difficulties of optimizing with the vibration modes introduced a strong local minimum in the design space that could only be solved thanks to our extremely powerful optimization technique.
The efficiency is increased from 86.4 % to 88.5 % while respecting 30 mechanical constraints

VERTICAL AXIS WIND TURBINE OPTIMIZATION

In this application, a vertical axis wind turbine was optimized. The fluid calculation is performed using OpenFoam, while the geometry is defined with CAESES and the optimization is performed using Xtreme. The goal is to optimization and maximie the Performance coefficient of the wind turbine by modifying 4 design variables: the profile chord, the profile thickness, the profile incidence angle and the rotational speed. The fluid simulation is performed using OpenFoam on a 2D unsteady simulation as show on the video below.

Xtreme Software: New Version 2.8.0

30th December 2019:

Today, Optimal Computing releases a new version of Xtreme 2.8.0

This version includes many improvements to the graphical user interface
Regarding the numerical optimizer, several improvements have also been included in the sequence optimization algorithm and in particular to its parallel version

Cost Optimization of BIPV application

In this application, the goal is to find the cost-optimal set up of Building Integrated Photovoltaic panels (BIPV) onto an existing building. BIMSolar for the simulation of solar irradiance and Xtreme for numerical optimization are used. A cost model is developed to compute the cost of any BIPV setup on that building. A numerical model was also developed to compute the energy flux in and out of the building including batteries. The design variables are the number of photovoltaic modules on each facade of the building. We were able to define the best BIPV set up and the optimal battery size as a function of the investment budget.

The Artificial Intelligence Revolution

Artificial intelligence is capable to find the solutions that are not possible using human reasoning or any other mathematical algorithms. This can be the case when trying to predict the future from past data. A typical example is when trying to predict the energy consumption for today+1, today+2, and today+3 as we did it for a residential house. We were able to predict the energy consumption with an incredible level of accuracy using 2 years of measured data

Xtreme Software: New Version 2.7.12

30th October 2018:

Today, Optimal Computing is proud to release its new software version of Xtreme 2.7.12

This version includes many improvements among with improvement on the restart capabilities of the optimization algorithm.

A new optimization algorithm allowing to perform sequence optimization. The sequence optimization provides incredible performance compared to other optimizers. It allows the user to easily set up the problem and to include all the constraints usually found during the optimization application. This new algorithm comes in addition to the existing continuous and discrete numerical optimization algorithms.

Xtreme Software: New Version 2.7.2

30th December 2017:

Today, Optimal Computing is proud to release its new software version of Xtreme 2.7.2 (Free trial available)

A new graphical user interface is available providing easier access to project configuration for both the stand-alone graphical user interface and the Microsoft Excel AddIns
Regarding the numerical optimizer, several improvements have also been included in this release.
  - Improvement of the management of uncomputable functions
  - Improved files management of simulations and directories
  - Improved graphical presentation of the optimization results

Xtreme Software: New Version 2.6.2

18th August 2017:

Today, Optimal Computing is proud to release its new software version of Xtreme 2.6.2 (Free trial available)

Several improvements to the Python API provided in the Xtreme software package
Migration to Microsoft Visual 2015 for the C++ API

Xtreme Software: New Version 2.5.12

6th March 2017:

Today, Optimal Computing is proud to release its new software version of Xtreme 2.5.12 (Free trial available)

Significant improvements to the speed of the neural network and genetic algorithm allowing to significantly increase the number of design variables than can be handled by the software.

Optimization of a residential house

28th December 2016:

We publish an article showing the application of our optimization software Xtreme to the design and optimization of a residential house. This article presents multi-objective optimization where the goal is to find the optimal Pareto front which minimizes both the heating energy needs and the construction cost. A real residential house to be constructed has been used as a use case. 2 simulations software have been used to compute the energy demands of the house: the PHPP and the TRNSYS software. In this paper, the results obtained with the PHPP software are presented. The cost evaluation function was developed in this project and the optimization was performed using XTREME, a commercial optimization software.

RESIDENTIAL HOUSE COST AND ENERGY OPTIMIZATION USING MULTI OBJECTIVES PARETO FRONT METHOD

Summary

This article presents the multi-objective optimization of a residential house. The goal is to find the optimal Pareto front which minimizes both the heating energy needs and the construction cost. A real residential house to be constructed has been used as a use case. 2 simulations software have been used to compute the energy demands of the house: the PHPP and the TRNSYS software. In this paper, the results obtained with the PHPP software are presented. The cost evaluation function was developed in this project and the optimization was performed using XTREME, a commercial optimization software.

SYMBOLS

PHPP                Passive House Planning Package (http://passive.de)
TRNSYS            Transient System Simulation Tool (http://www.trnsys.com)
XTREME           Numerical Optimization Software (http://www.optimalcomputing.be)
U                      Thermal conductance (W/m2K)
g or SHGC        Solar Heat Gain Coefficient (%)

INTRODUCTION

Reducing the C02 production around the world and in all sectors is now recognized as mandatory objectives in the near future. Among the biggest consumer sector is the householder's sector, which in Europe represents around 24.8% of the energy use [1].
Therefore, since several years the construction of low energy or even passive building attracts a lot of interest from many stakeholders: private person to construct their personal house, public authorities, European Commission.
Energy consumption targets have been set for the passive house but the extra cost required to reach this passive standard is often a barrier for individual persons. It is difficult to evaluate precisely the return on investment period for such a passive building.
This paper has 3 objectives:
  1.     Demonstrate how numerical optimization techniques in general and Pareto Front optimization, in particular, can be used to find the set of best solutions for typical houses depending on their energetic performance.
  2.     Deduce from this Pareto front, the return on investment time of a standard residual house in terms of energy demands and construction cost.
  3.     Deduce important best practices depending on the target heat consumption (standard house, low energy house, passive house.) The outcome of this study will be to demonstrate that the return on investment time for several energy targets can be evaluate using Pareto front optimization techniques.
This paper first presents the house used as a baseline in this study. Then, the next 2 sections present the software used to simulate the energy performance and the cost function developed to calculate the construction cost. The fourth section presents the optimization software used as well as how it was coupled to the simulation software. Finally, the optimization results are presented and analyzed.

Pareto Front in the space Energy Demand - Cost difference

Final Residential House

Download the full paper here

Xtreme Software: New Version 2.4.5

26th December 2015:

Today, Optimal Computing is proud to release its new software version of Xtreme 2.4.5 (Free trial available)

The major new feature is the management of uncomputable accurate functions

Xtreme Software: New Version 2.4.1

24th September 2015:

Today, Optimal Computing is proud to release its new software version of Xtreme 2.4.1 (Free trial available)

Main new feature: flexible response surface viewer
The Xtreme software allows you to solve design challenges in many engineering disciplines (aerodynamic, structural mechanics, virtual manufacturing, noise, vibration, and many more), as well as in fields such as finance, scheduling, manufacturing, building. Virtually any design problem that can be modeled using your own or commercial simulation tools or even Microsoft Excel can be solved using Xtreme.
Xtreme is based on very advanced optimization techniques relying on artificial intelligence techniques such as genetic algorithms and artificial neural networks.
Xtreme is the most efficient numerical optimization technique you will find on the market reducing by a factor of 10 to 25 the required number of expensive simulations to reach the optimum.
The software is very simple to use thanks to huge development/research efforts performed to avoid expert parameters to be used and tuned. As a consequence, you can concentrate your effort on your discipline and let the optimizer find the best solution for you.