SUCCESS STORIES

PROBLEMS

In the manufacturing process of parts for the automotive sector, the assembly by welding of certain mechanical components can be affected by deformations that could be detrimental to quality.

Until then, the result could only be determined by manual inspection, resulting in a significant loss of raw material and time.

RESULTS

• Deformation prediction through the Digital Twin
• Reduction of time and raw material in the testing process of the parts.
• Higher quality of the assembly result by obtaining more robust parts thanks to the predictive model.

SOLUTION

Predictive model of part deformation:

• Investigation of previous parts (pre- and post-assembly) to determine the influence of the parameters involved on the final finish.
• Development and training of a predictive model to estimate the deformations of the given part under input conditions.
• Establish acceptance thresholds to translate these deformations into quality metrics.

TOOLS

PROBLEMS

• Data from different equipment involved in the drug manufacturing steps are manually sorted in Excel tables.
• In the absence of an integrated data collection study, data needs to be exploited and processed manually, and human errors can occur.

SOLUTION

Creation of a platform that receives data from the manufacture of medicines, displaying them in real time and providing different types of graphs and statistical data. of medicines, displaying them in real time and providing different types of graphs and statistical data in order to check the evolution of processes and quality controls.

RESULTS

• Reliable and quality data collection
• Statistical reports and graphs to improve the work of pharmaceutical production sites
• Better results at the end of quality processes, considerably reducing waste.

TOOLS

PROBLEMS

Hospitals do not have an optimal method to transform all the data already recorded into relevant information, so the ability to identify patients at risk of non-adherence to their treatment increases considerably.

RESULTS

• Increasing the quality of information for prescribing treatment.
• Increased capacity and agility in decision-making.
• Reduction in the number of unscheduled consultations and visits or emergencies.
• Reduction in the number of hospital admissions.

SOLUTION

It is proposed to create a predictive model to predict the risk of non-adherence to oral antineoplastic therapy. The aim is to identify, from the outset, those patients on whom it will be necessary to focus interventions and resources individually, without having to wait for possible clinical and/or economic consequences.

TOOLS

OBJECTIVE

The client requires a solution that allows him to automate the reading of license plates and the identification of vehicles (make/model) in order to be able to invoice the corresponding tariff.

RESULTS

Development of a tool that by means of Artificial Vision-Deep Learning (convolutional neural networks), allows the reading of license plates.

Development of a vehicle category classification system.

Some of the advantages gained are:

• Task automation
• Improving the quality of information
• Optimisation of invoicing conditions

SOLUTION

• Computer vision
• Neural networks (CNN)
• Machine Learning: object classification and detection algorithms.

TOOLS

OBJECTIVE

The customer has several hydraulic presses for the manufacture of parts in the automotive sector, each of which has a sensor to monitor its correct operation.

Another of our customers, with a similar case, has a combined cycle electric generator that monitors the air fuel flow parameters as well as the electrical parameters of the power generated and has a series of alarms that warn if anything is outside the operating limits considered as normal.

The need for these customers lies in anticipating any problems that might occur with the machines, with the intention of preventing unscheduled downtime.

SOLUTION

It is proposed to develop a data acquisition system for the more than 700 sensors, centralised in a platform that allows us not only to measure their operation in real time, but also to develop a predictive model.

RESULTS

• Reduce the number of breakdowns by 55%.
• Downtime reduction (MTTR = 60%)
• Reduction of spare parts stock
• Increased life of parts
• Increase in production (30%)

TOOLS

OBJECTIVE

The customer requires a solution to optimise the costs and maintenance time of the Spanish railway infrastructure.

RESULTS

• Enormous reduction in the working time of the maintenance team required for track maintenance.
• Reduction of the cost associated with it

SOLUTION

Artificial vision model using neural networks that allows:

• Detect and analyse each frame automatically
• Detect elements that are in a state of deterioration.
• Identify anomalous or non-infrastructure items
• Establish a register of those points on the road where the deteriorated elements are located.

TOOLS

OBJECTIVE

The customer requires a tool that allows it to exploit the data from the Data Link of the air traffic control system to facilitate the analysis and monitoring of the data.

Therefore, it is proposed to automate the processing and exploitation of log data from the air traffic control system to monitor the occurrence and frequency of different types of messages.

SOLUTION

The main objective of the project is the analysis, design, development and implementation of a data processing and exploitation tool that facilitates the reading and analysis of air traffic control system messages, ensuring their quality control.

RESULTS

• Time and cost savings in data collection (centralised, accurate, quality and traceable).
• Optimise decision making, through an intuitive and robust exploitation layer.
• Focus the effort on the business and not on technology or on manual or semi-manual procedures.
• Provide users with the self-service capability to perform their own analysis.

TOOLS

OBJECTIVE

The main objective of the project is the construction of an information extraction process, mainly from the ERP, for the monitoring of commercial information from the second-hand vehicle department and its subsequent consultation, ensuring the quality control of the information.

SOLUTION

To cover the main needs of information analysis and automatic generation of information, guaranteeing scalability and future maintenance.

RESULTS

• Optimisation of the selling price of used vehicles.
• Estimation of the depreciation of a given asset in order to know when to sell it.
• Establish a sales scenario simulation model.

TOOLS