Mills gather enormous amounts of data in various forms. These data are often stored in separate databases, silos. Moreover, these data are underutilized, especially process data from manufacturing. Software tools typically used in mills are inadequate in bringing right information in the right format at the right time for the right people. This leads into symptoms such as quality claims, waste, slow troubleshooting, and increased costs through higher variation in the process.
‍
Quva® Flow analytics software is dedicated to solving these obstacles and help mills to take full advantage of automated utilization of their own data, resulting in more stable production. In practice, Quva® Flow is comprised of three layers, which form a seamless flow of information from raw data to actionable information for decision support. The layers are seamless mill data integration, AI enhanced analytics engine, and easy-to-use visualization. Quva® Flow is a modern software-as-a-service (SaaS) platform that is easily accessible and future-proof.

Core Functionality

Quva® Flow is designed to respond to critical user needs in a manufacturing environment. Quva® Flow is for you when you are looking for Improved quality & customer satisfaction, less waste for sustainability impact, data driven process and product development or shared visualization and access to core metrics

Quality Discovery

With Quva® Flow Quality Discovery-tool user can study root causes for interesting phenomena. Tool will browse through selected group of products and as end result provide listing of most interesting anomalies. There are many different use cases for this tool. It can be even said that only restriction is user's own imagination.
Most common use case and studied phenomena is to compare good and bad quality products/reels. You can also compare reels which are produced in different timeframes or compare reels produced by different shifts

Quality Score Card

Alarms View
With Quva® Flow’s Quality Card, process and quality engineers get a fast and comprehensive view on all deviations in any reel produced across quality and process data.
Quality Card highlights anomalies in all reel related data to mill personnel enabling fast identification and reaction to issues concerning product quality.

Monitoring

Monitoring view provides easily approachable product specific visual meters backed by artificial intelligence and alarms to help operators to better produce the wanted quality in a proactive manner.
With monitoring view user is able to monitor real time values for parameters which can be selected either manually or based on analytical suggestions.

How is Quva® Flow different?

Unlike other analytics companies talking about industry 4.0 solutions and how their products are leveraging AI, Quva® Flow stands out for four key reasons:

High ROI through significant decrease in claims and waste (up to 20 %) combined with reasonable pricing

Proven advanced analytics and data models for process manufacturing, especially paper and cardboard production

Holistic solution for several mill roles such as operator, engineer, and manager

Fast implementation at the mill
typically taking less than 2 months.

How Does Quality Discovery Work?

In Select Sources-view, from Analysis Source box user defines needed parameters for root cause analysis: dates, production line and products to be studied.
Grouping assistance: User will see boxplots for this variable when doing selections for groups A and B.

Reels which are matching for done selections are listed in below table.

In this view user will choose reels which are in next phase used for comparison by analytics.
Group A is chosen by activating green pen and then from below dialog selecting single reel(s) by individual clicks or multiple reels by painting gesture.
In similar manner Group B is chosen with red pen.
Filter-functionality below can also be used for selecting green and red groups.

This image shows how filter can be used for example when selecting Group A to include all reels with QA_status value P1 and Group B to include reels with QA_status value P2.

In Analyse and Discover-view user has listed results of analytic calculations in order of importance.

By clicking individual results user can see measured values for that specific measurement with same sample selections, and study if it can possibly have some affect on interested phenomena.
User can pin interesting results. Pinned results will be shown on top of result listing even when moving between result pages.

Calculation result is automatically saved for later usage as own web page. This link can also be shared for example to colleague.

Comprehensive data integrations

Quva® Flow was developed specifically to process manufacturing environments. Therefore, there are several data sources that we have already integrated the software with. If we are missing one of your systems, rest assured, we haven’t yet encountered a data source or a system to which we have not been able to integrate. Most frequently used data integrations include:

Honeywell PHD

ABB RTDB

Osisoft PI

Valmet DNA

Tieto TIPS

SAP

Gain competitive edge leveraging ML/AI in industry context

If you want to run your business based on up-to-date facts, then you definitely want to make your data work for you. Quva® Flow’s analytics engine is tailored to automatically refine meaningful information for decision-support in complex settings with high data volumes. Our analytics algorithms are optimized to time series data, which are ample in the world of process manufacturing. Artificial intelligence enhanced analytics engine is highly configurable comprising a variety of techniques from real-time statistical process control to machine learning and deep neural networks. You can rest assured that your organization’s information needs will be solved.

Role based visualization

World-class analytics gets you far, but something else is also needed for you to reach your goals. On top of its constantly running analytics engine, Quva® Flow provides mill personnel with real-time visual information about process deviations enabling them to tackle issues more thoroughly. Because of its scalability and configurability, it suits various uses from identifying issues in real-time at factory floor to AI assisted troubleshooting and mill level KPI drill-downs. Unlike traditional tools, Quva® Flow provides different views for different roles in the mill from operators to engineers and managers as useful information for one role might be useless for others.

Quva® Flow’s Features

Alarms View
Important deviation information on all process and quality data of products (e.g. reels, cooking batches) at a glance in an easy-to-use visual format.

Capability View
Interactive tool to identify and monitor mill-level key performance indicators such as amount of waste per day and drill-down to factory floor level deviations, which might explain the high-level issues.

Discovery View
User assisted machine learning to take troubleshooting from days of work into a few data-driven clicks..

Analysis & Reporting
A simple tool to make fast and informative graphs of selected variables from all production data regardless the source (e.g. process, quality, and other data).

Prediction View
Advanced machine learning to create soft sensor for production and e.g. predict end quality measurements through process data in real-time.

Monitoring View
Easily approachable product specific visual meters backed by artificial intelligence and alarms to help operators to better produce the wanted quality in a proactive manner.

Logbook
Innovative feature to combine domain knowledge with automated data analytics of sensor and measurements data

How Does Quva® Flow Quality Card Work?

With Quva® Flow's Quality Card, process and quality engineers get a fast and comprehensive view on all deviations in any reel produced across quality and process data. Quality Card highlights anomalies in all reel related data to mill personnel enabling fast identification and reaction to issues concerning product quality.

On top of the screen, user can choose timeframe where reels are wanted to be studied from. Number of alerts (Quality alerts, Critical process alerts) for reel are shown with icon and traffic light indicators. Wider the icon, more there is variability in particular reel. When user clicks any reel in the list, new view opens up with reel specific information

In reel specific view all variables are listed. Variables are shown in descending order based on variables alarm score. Variables are categorized based on colors for alerts (red), warning and green. For each variable and row, boxplot shows measured values. Also limits are visible. User can click any interesting variable for more detailed information

In variable specific view, user can see detailed alarm and measurement data for particular variable. Also performance and statistical scores are listed. In this view user has good visibility for historical data of same variable when making same particular product.

QUVA® FLOW