Since its foundation in 2002, PEPITe’s core business is Advanced Analytics for industrial process optimization. Experience has taught us that optimization, scalable and sustainable results require the right tool, support, knowledge transfer and capability building. That is why PEPITe is dedicated to fostering client capabilities in Artificial Intelligence and more specifically in Advanced Analytics.
Starting with a complex AI optimization project can be challenging for several reasons. Firstly, not every plant is ready for AI technology. We recommend assessing a plant’s digital maturity to define the appropriate implementation roadmap. Secondly, an AI project without knowledge transfer leaves the plant with a one-off solution, without the skills or tools to maintain or scale that solution.
This is why PEPITe’s methodology goes hand in hand with our technology, DATAmaestro. By increasing our clients’ skills, they can become completely independent to implement and maintain AI optimization projects. Additionally, we transfer the skills to help make plants faster and more efficient in solving industrial problems, troubleshooting process issues and helping them to deliver impact quickly in the plant.
Without this emphasis on capability building, clients run the risk of carrying out yet another Pilot Project, without scalability.
DATAmaestro is designed by engineers for engineers as a self-service analytics tool. Thanks to a low-code interface and automated data collection and preparation, DATAmaestro helps remove barriers for engineers to harness their data. It can also help reduce time spent on integration and management of specific models developed by data scientists or machine learning engineers. >
Creating impact with data is not just an IT question, as having the right tool is not everything. It’s important to know how to create value with that data and with that tool.
As a result of 20 years of experience, PEPITe has developed OPTImaestro, a participative methodology to deploy successful and sustainable AI projects in industry. Engineers and operators hold a wealth of process knowledge about their plant, while IT systems have accumulated a huge amount of sometimes unused data. OPTImaestro combines both these sources of information to create value from data.
As part of our approach, PEPITe teaches their clients the keys to following the OPTImaestro methodology to become leaders of their own optimization projects.
OPTImaestro methodology:
Throughout the project, the plant learns how to implement an optimization project and how to use DATAmaestro, PEPITe's advanced analysis tool.
The Digital Data Diagnostic helps gauge the technical and economical feasibility of AI projects in a plant, while building awareness of these opportunities. This diagnostic helps identify the required data sources and connection options to collect data from those data sources.
Based on the results from the Diagnostic, DATAmaestro can be deployed on premises or in the cloud and connected to plant data sources.
Users are given access to detailed modules of DATAmaestro functionalities stepping through an Analytics workflow to be followed online at their own pace. The goal of this e-Learning is to provide an elementary knowledge of Advanced Analytics and DATAmaestro’s functionalities.
Additionally, participants can follow several predefined modules organized through live webinars. They cover the entire methodology to successfully implement advanced analytics projects with DATAmaestro (16 hours). These sessions go a little bit deeper than e-Learning and encourage interaction to consolidate knowledge.
Once the system is deployed and client teams have followed the initial training sessions, they are ready to begin their first Analytics project. PEPITe provides coaching to help end users in their Advanced Analytics projects. These sessions focus on a specific problem the client wishes to solve with DATAmaestro, including: opportunity sizing, root cause analysis, models building, deployment and maintenance, etc.)
The coaching is entirely tailored to the client’s needs.
At the end of this pilot project, the client has the skills and tools in place to continue to maintain and replicate analytics projects.
If companies want to stay competitive they need to learn how to scale Advanced Analytics. PEPITe ensures that they obtain convincing and scalable results by increasing their skills in a quick and accessible manner.
Phase 1: PROCESS AND BUSINESS UNDERSTANDING
The first phase of any analytics project is to understand the process and business challenges.
Phase 2: DIGITAL DATA DIAGNOSTIC
What data is available, what are the plant’s Key Performance Indicators (KPIs) and what are the process improvement
opportunities?
The Digital Data Diagnostic helps gauge the technical and economical feasibility of AI projects in a plant. Digital maturity is evaluated, data availability is highlighted, roadblocks are identified. Opportunities often include improvements in yield, energy efficiency, throughput, product quality, equipment availability?
Phase 3: FLASH ANALYSIS
What are the potential savings from an AI project? The Flash Analysis is a preliminary analysis on historical production
data to quickly quantify the potential improvements and main sources of variability that might explain low performance.
The goal is to refine the expected savings and prepare the brainstorming sessions with plant staff.
Phase 4: WORKSHOP WITH OPERATORS
How can we involve operators to ensure a successful project? At this stage, brainstorming workshops are organized with
production and maintenance teams (operators, engineers, managers). These sessions are designed to pinpoint root causes
underlying a particular performance problem and to understand operating constraints.
Experience has shown that this phase plays a key role in the successful implementation and engagement of plant staff to trust data-driven and AI based decision support tools.
Phase 5: ADVANCED ANALYTICS AND MODELINGWhich parameters are key to maintain optimal performance? Thanks to advanced analytics tools based on artificial intelligence:
- explore past process variability
- identify key variables
- detect root causes to process issues
- understand best operating conditions
- predict performance depending on plant conditions
- prescribe set points to maintain optimal performance
Phase 6: IMPLEMENTATION
What tools, dashboards or reports are required to facilitate and support operators in maintaining optimal performance?
Ensure the right information gets to the right person at the right time by deploying real-time models in dashboards to
help monitor performance. Teach plant staff to take the appropriate actions based on live dashboards and reports.
Phase 7: REVIEW
How can I ensure long-term follow-up and sustainable results? Following implementation, a follow-up review period
ensures that models are performing correctly and that the reports and dashboards are understood and well used. This is
the beginning of a continuous improvement cycle.