An Italian company, serving oil and gas, water transmission, desalination, power generation fields with engineered pump solutions tailored on customers requirements, has developed a tool aims at supporting the hydraulic design of centrifugal pumps with extraordinary performances using Artificial Intelligence (AI) approach or advanced Machine Learning (ML) methods.
Target of the system is to provide a geometric setup of the pumps that achieve the extremely aggressive rated performances, in terms of efficiency or suction capability.
The software will be trained with the data of an internal database where all geometric information and the performance data of the best hydraulics of the company are collected.
The tool can be applied to all pump types managed by the companies. On the first phase the tool will be developed for a pilot pump type (i.e. single stage-volute pump), then the methodology will be extended to other pump families.
The core of the tool can be developed by following one or both of the following methods:
- to use Machine Learning approaches with advanced methodologies
- to use innovative Artificial Intelligence algorithms (i.e. processing images of flow field within the pump).
The code will be supported with several Computaional Fluid Dynamics (CFD) analyses of complete pumps that will be used for both the training and verification phase.
Our aim is to implement an innovative approach for the hydraulic selection and design, allowing to perform a significant optimization of the process in terms of efficiency and power consumption.
Universities or AI software suppliers are possible partners for the development of the tool.
Advantages and innovation
The expected advantages are:
- shortening the design activity of pump hydraulic;
- optimization of hydraulic selection to match the client s requirements;
- improvement of pump efficiency with lower power consumption and minor environment impacts.
Technical Specification or Expertise Sought
Our aim is to implement an innovative approach for the hydraulic selection and design, which allows to perform a significant optimization of the process in terms of efficienty ad power consumption.
The preferred partner shall have the following referenced skills:
1) Good experience with programming languages.
2) Proved experience in feeding and verifying Artificial Intelligence (AI) / Machine Learning (ML) softwares.
3) Analysing the available data on the pump hydraulic performances.
4) Designing the the mathematical models and developping the relevant Artificial Intelligence (AI) / Machine Learning (ML) solutions/softwares.
5) Testing the developped models and improving them on the basis of the tested data.