Con esta herramienta te facilitamos un acceso a todas las ofertas y demandas de tecnología europeas y a búsquedas de socios para participar en propuestas europeas de I+D publicadas en la red Enterprise Europe Network, pudiendo filtrar los resultados para facilitar las búsquedas más acordes con tus necesidades.

¿Quieres recibir estos listados de oportunidades de colaboración en tu correo de forma periódica y personalizada? Date de alta en nuestro Boletín

Para optimizar los resultados de la búsqueda, se recomienda utilizar términos en inglés.

Software de optimización de procesos industriales

Resumen

Tipo:
Oferta Tecnológica
Referencia:
TOES20161129001
Publicado:
23/12/2016
Caducidad:
23/12/2017
Resumen:
Una pyme española especializada en eficiencia energética ha desarrollado un nuevo software que incluye herramientas conocidas de extracción de datos e inteligencia artificial para optimizar procesos industriales (desde grupos de compresores hasta unidades de producción de hidrógeno y líneas de distribución de gas). Este software contribuye a la eficiencia energética de los procesos y a la reducción de los costes energéticos y emisiones de CO2. A pesar de las numerosas aplicaciones potenciales de las técnicas empleadas por esta herramienta, el software es especialmente interesante para aplicaciones en procesos multivariables complejos, como plantas petroquímicas. Se buscan empresas con el fin de establecer acuerdos comerciales con asistencia técnica.

Details

Tittle:
Software tool for industrial process optimisation
Summary:
A Spanish SME has developed a new software tool, comprising well-known data mining techniques and artificial intelligence, for the optimization of industrial processes (from compressor groups to production hydrogen units and gas distribution lines) and therefore contribute to the energy efficiency of the processes and the reduction of the energy costs and CO2 emissions. They are looking for companies to formalise commercial agreements with technical assistance.
Description:
A Spanish company working in the field of energy efficacy has developed a software tool, based on data mining and artificial intelligence, for modelling and optimizing industrial processes that contributes to energy reduction cost and CO2 emissions reduction of the industrial site or the process where it´s been applied. Although the numerous potential applications of the techniques employed by this tool, this software is particularly interesting for applications in complex multivariable processes such as petrochemical process plants.
Accurate understanding of the influence of variables and in so far as they do to the performance of a process isn´t known, only artificial intelligence offers answers to the modelling tasks.
Data mining techniques starts from historical databases of the process, which must meet two basic requirements: extension and sufficiently rich. Accordingly, the database must collect an acceptable number of records, covering a significant operational variability of the different operation modes of the process.

Variables of influence identification
For variables of influence (VoI) on variables to model, this tool has a powerful filtering module and feature selection, which allows for: (i) Remove anomalous records, (ii) Detect correlation (either linear or not) between variables in order to modelling and (iii) select the relevant and non-redundant set of VoIs. Among the techniques used for feature selection, this soft uses filters and wrappers: the first are simpler and faster to apply since they are based on different correlation coefficients (Person, Spearman, DRC among others); while the second group comprises brute force methods, which perform the calculation using model-based classifiers (Forward -Y- Backward Selection) techniques.

Modelling
The modelling technique used by this tool is the neural network, whereby; a neural structure previously created by the user can be trained using various algorithms to reproduce certain behaviours. Because they are stochastic methods, the tool allows for the application of bootstrapping, which permits an evaluation of the random effect of methods and thus offering a guarantee on the modelling results.
Upon training processes is complete, the information of the performance of the models created through statistical variables as MAE and MSE is presented, as well as graphical representations of comparison between the actual and modelled values.

Optimization
Once verified the predictive ability of the models previously created, the optimization problem arises: Is it possible to find new situations that report extra benefits compared to the figure in the current operating situation? To answer this question, this solution offers two reliable alternative optimization algorithms with virtually universal applications, such as simulated annealing and genetic algorithm. In general, objective functions can exhibit complex topologies in the considered optimization space, thus requiring the use of powerful optimizers in order to avoid stacking on local optima.
The software optimizer allows for solving multi-objective problems both off and on-line. It is also possible to modify the main characteristics of each optimization algorithm. The definition of the optimization space is performed by selecting first optimization variables between the non-redundant variables of influence; and secondly specifying restrictions for them.

This company is interested in establish commercial agreements with technical assistance with industrial companies (SMEs and big companies) to optimize their processes and increment the energy efficiency and contribute to the reduction of the energy costs.
Advantages and Innovations:
* All the mathematical techniques employed are well-known and its utility has been demonstrated for a wide variety of applications.
* Data mining techniques included allow for removing wrong registers, thus improving accuracy in simulations and optimization results.
* The utilization of artificial neural networks may by-passes problems related to reconciliation.
* Artificial neural networks allow for modelling a whole system or unit without knowing all the physics included in them. In consequence, a sufficiently rich database is needed.
* Advanced optimization methods, likes genetic algorithms and simulated annealing; have replaced traditional ones, unable to solve problems with big number of variables and non-linear behaviours involved.
* The simulation and optimization tool is pre-configured, just requiring few changes for any industrial plants.
* The user´s interface has been developed for easy use and quick configuration. New operating restrictions can be included in optimization scenes.
Stage of Development:
Already on the market
IPs:
Secret Know-how

Partner sought

Type and Role of Partner Sought:
SMEs and big industries working in the field of petrochemical, fertilizer, chemical, agri-food and all those activity fields operating with large, complex systems of air compressors, boilers, hydrogen network, etc. that can not be optimized by traditional methods such as linear programming.

Client

Type and Size of Client:
Industry SME <= 10
Already Engaged in Trans-National Cooperation:
No
Languages Spoken:
English
Spanish

Keywords

Technology Keywords:
01003006 Computer Software