Finding Events with Sabisu Machine Learning

As you’ll have seen from previous blog posts, we’re enthusiastic about the applications for using machine learning with industrial data.

Machine Learning identifies features more reliably than other methods with fewer false positives/negatives; ideal when trying to find recurring behaviour which needs early action, e.g., asset failures, golden batches, repeated maintenance, non-conformances, etc.

As Sabisu Machine Learning is nearly ready for general release, here’s an update on what you can expect in the first release.

Training made easy

Machine learning is an ‘umbrella term’ that covers a myriad of Sabisu algorithms and techniques.  They’re all statistical methods which “learn” from data; the more data you have, the better they will perform.

Sabisu makes this easy with Events, which allow users to mark a section of data, e.g., decoke, heat exchanger failure, pump failure, etc.

This provides a library of identified features in your data. It’s exactly what the algorithms need to identify thousands of events at the touch of a button.

Sabisu Event

An Event shown on a Sabisu chart.

After marking the Event, Sabisu will ask whether it should find similar occurrences of this behaviour.

Sabisu will identify every single similar event in the date range you choose. Machine learning ensures that the match is good – we’ve optimised our machine learning algorithms to outperform other statistical techniques.

Event Identification

After an Event has been tagged, you are given the option to find similar behaviour in your data.

Each of these matches is marked as a new Event so you can quickly move between them, looking for similarities or root causes.

All the usual collaboration capabilities are in play; highlight similarities in Notes, allocate an Action to investigate or analyse further, upload files and so on.

How does Sabisu ML work?

To analyse large quantities of data Sabisu utilises cloud technology, seamlessly passing data to our cloud data-lake; a highly versatile & scalable structured and unstructured data store.

Distributed processing is used to to build a model using unsupervised learning (see our previous blog post). This is used to identify past occurrences of the chosen event throughout the entirety of the supplied data. The results of this operation are stored in our distributed cloud data warehouse, allowing Sabisu to display them to you through pipelines, widgets, and events.

EventIdentification Schematic

This schematic shows how your event and historical data are passed from your Sabisu unit to our data lake, through a Sabisu analytics cluster where machine learning is used to identify similar events in historical data, with the results stored in a parallelised distributed data warehouse for rapid access.

What’s next?

This is just the first step.

Currently in beta testing Sabisu is using machine learning to detect when an identified event is about to occur (‘incipient events’) or has just started to occur (‘triggering events’) so that you can take early or pre-emptive action.

An obvious application of this is in the case of asset failures where early warning can be invaluable in reducing downtime and optimising maintenance.

Machine learning is also being used to de-noise and clean data intelligently before it’s committed to the data-lake.

Very soon, Sabisu will provide users greater control over how machine learning models are constructed and used with customised data classification to further tune model behaviour.

Machine learning is becoming an invaluable tool to help keep processes running in the best possible way, reducing wastage and maximising quality through the identification of recurring good or bad behaviour as soon as possible.

Contact Us

We’re always interested in hearing from you with any comments or suggestions, feel free to get in touch.

Posted in Operational Excellence, Operational Intelligence, Project Intelligence | Tagged , , | Leave a comment

Drive usage & target improvement with Community Reporting

At the Sabisu User Group in January customers highlighted Community Reporting as eing extremely beneficial, helping them manage their teams and improving productivity. Clearly this was something we could implement that all Sabisu platform users would find valuable.

You can access the Community Reports in a couple of ways:

  • From a Community Page, select the page menu, then ‘Manage this page’, ‘Community Reports’
  • Find the Community in your list of Communities and navigate to the ‘Community Reports’ tab

Currently 5 reports are available by default – if there are any others you may find useful please let us know:

Platform Usage – This shows the total number of community members and the number of unique user log in’s (how many people logged in on this day)

Most Popular Pages- Shows the top 5 most visited

Least Popular Pages – Shows the top 5 least visited

Most Frequent Users- Shows the top 5 most frequent

Least Frequent – Shows the top 5 least frequent

You can also select a date period in which you wish to view the data, so you can track activity over the week, month or year.

Contact Us

We’re always interested in hearing from you with any comments or suggestions, feel free to get in touch.

Posted in Release Notes | Tagged , | Leave a comment

A Simple Process for Complex Data: Dynamic Queries in the Pipeline

Recently we introduced Self Service Queries in the Sabisu Pipeline, allowing you to choose exactly which columns and data components you need in your Pipeline and how that data should be filtered.

This week we’ve made them even better by adding a step by step guide to make it easier to set one up and get going:

  • Start by selecting your data source (e.g., an MS Excel upload, SQL table,  IP.21 or PI tag)
  • Select the columns that you want out of the data source, e.g., time stamp, value, units
  • Set up any filters you may need such as data from the last 24 or 48 hours, or only get rows of data that contain certain information selected by you.

This makes it easy enough for any user to find the data they need and use it in a Pipeline.

Contact Us

We’re always interested in hearing from you with any comments or suggestions, feel free to get in touch.

 

 

 

Posted in Release Notes | Tagged , , , | Leave a comment

Total Process flexibility with new Actions triggers

We’ve updated the triggers in Sabisu Processes to allow conditional initiation of Actions; the next Action in a process can be initiated either once all or selected previous Actions have been completed.

When configuring each Action in a Process, users choose which preceding Actions initiate the one they’re configuring. There are two Trigger options:

  • All means the current Action is initiated by Sabisu when all the Actions selected are completed.
  • Any means the current Action is initiated by Sabisu when any one of the Actions selected are completed.

Sabisu looks after all aspects of initiating the Action; creating, allocating, reminding and setting completion dates.

You’ll notice that you can select Actions to be triggered from any of the preceding steps in the Process. This allows you to run several tasks in parallel so you don’t need to wait for one step to begin before you start another.

Work continues on Sabisu Processes; we’re currently working on integrating Standard Operating Procedures into the platform.

Contact Us

We’re always interested in hearing from you with any comments or suggestions, feel free to get in touch.

Posted in Release Notes | Tagged , , | Leave a comment

Actions goes mobile

You can now create Sabisu Actions on your mobile device – any time, any place.

Perhaps you’re walking round the plant and notice a wall that needs repainting or some scaffolding without the correct safety tags. Simply create an Action straight away for a quick resolution.

Maybe you’re out of the office and need a reminder to update documents? Create an Action for yourself and Sabisu will ensure you don’t lose track.

If you have no internet connection  don’t worry; Sabisu Actions will save and synchronise as soon as a connection is re-established.

If you don’t have it already, download it now from the Google play store or get in touch if you’d like it deployed to your plant mobile devices.

Contact Us

We’re always interested in hearing from you with any comments or suggestions, feel free to get in touch.

Posted in Release Notes | Tagged , | Leave a comment

Workflow Management with Sabisu Processes

Sabisu Processes are a focus for the platform development team right now. Processes are chains of Actions which trigger in sequence – each one being created and allocated after the previous one is complete, allowing you to build any process, however complex.

They also work seamlessly with Sabisu Pipelines so you trigger a process based on criteria set by you, e.g., when a particular cost centre breaches a limit, or when a process parameter falls below a certain limit.

Once you’ve set up a Process you can edit or trigger it on demand and Sabisu will take care of the rest by sending out Actions as and when they are required.

processes

To set a Process to trigger from a Pipeline you will need to first create the Process (see above) then select the correct Process within the Pipeline (see below):

process with pipeline

It’s remarkably easy to get started with Processes and, of course, because Sabisu uses Actions to deliver them users can access and update them on any Android mobile device.

Contact Us

We’re always interested in hearing from you with any comments or suggestions, feel free to get in touch.

Posted in Release Notes | Tagged , , | Leave a comment

Cloud first: The path to disintermediation in a conservative market

Never say never

Sabisu was designed as a hybrid-cloud solution. We wanted the ease of deployment of a SaaS solution as that would in turn allow us that degree of agility missing from many existing vendors in the manufacturing systems sector. Cloud also gave us a conduit for delivering data to external third parties.

“No one will ever connect their processes to the cloud. ”

However it seemed unlikely that customers would ever want to move their process data to the cloud. Before 2011, the oil & gas/petrochems/manufacturing guys were actively against it on security grounds. There was a slight thawing up to 2015, where the position was that the benefits, costs and risks were neutral at best, so why bother?

Rolling the snowball

We couldn’t say there was a single reason to move to the cloud – there are lots of good reasons. We started with a single reason, then added another, then another and so on, like rolling a snowball causes it to grow.

For us, the first snowflake was collaboration. It’s been important to Sabisu since day #1. Collaboration between organisations needs an intermediary, so that’s some internet servers – or, well, the cloud.

Seamless data exchange was the second factor. The third was the need for a truly zero-touch SaaS implementation so customers could get up and running very quickly.

Then there was the need for elastic compute to support analytics. Next it’s machine learning. Then it will be Artificial Intelligence.

These capabilities demand technology that can’t be found on-premise. The snowball rolls and grows and suddenly the cloud looks like a smart move.

Everyone’s on this journey – some faster than others.

Non-industrial sectors now lead

Oil & gas, petrochemicals and manufacturing can be very conservative. The change has happened. Everyone is now looking outside their on-premise environment.

The two key catalysts are (i) tech marketing and (ii) non-industrial innovation.

Tech marketing has created a permissive environment regarding the cloud. This is your standard ‘Crossing the Chasm’ model; everyone watches the techies, visionaries and early adopters, then pick it up when it feels right to you, depending on how conservative you/your organisation is (hence early/late majorities, laggards).

For example, major capital projects: industrial projects are more likely to be early adopters than public sector projects. Or take petrochemicals plants; new plants may be early adopters, some running plants will be laggards.

That will change, of course, when the cloud capabilities mature into relevant killer applications for those sectors. That’s coming.

Small manufacturers who couldn’t previously afford historians, or complex MES systems will be winners – and therefore, early adopters.

The other change is that everyone is now used to their personal data living somewhere out of their reach. They’re used to easy, SaaS applications that put user experience first

Non-industrials are rinsing out the tech way before industrials. In the old days, tech would filter from high end industry out to consumers/retail (e.g., lasers, strain gauges). Not these days – and definitely not with the cloud.

For example, Google pioneered technology 10 years ago (e.g., MapReduce, PageRank) which industrials have not yet got their heads around. Strava takes billions of lines of process (time-series) data every day and drives analytics from it.

One step ahead

We’re seeing customers following now – but really, only just now. Our job is to stay ahead, ready for them.

The first step is to make it easy to get the data to the cloud. We’ve done it with Bridge and Appliance and others are no doubt doing the same.

There are of course rules preventing use of the cloud in many organisations. They’re being ignored.  Users will break the rules to get what they want. Always have and always will.

Getting the data is the first step. Once you have it, you need to do something with it.

That’s where the SaaS capabilities are crucial. Self-service is essential as any implementation overhead will delay deployments, probably fatally.

Our approach has been to build SaaS capabilities into the platform:

  • Pipelines, which provide an easy way to create charts, initiate processes in response to changing data
  • Analytics that are powerful but easy enough for everyone to use
  • Workflows and Actions which hook users and communities together to get work done
  • Communities, so users can share data and administer groups of like-minded individuals
  • Data Objects, so users don’t have to know anything about the data to use it

What about conservative customers?

We totally understand that some customers are not ready to go to the cloud.

Sabisu has on-premise Units at a number of customers which will be supported for the foreseeable future. It’s a success.

Indeed, the Unit will benefit from some changes that the cloud implementation will drive, moving to a micro-service, highly scalable architecture but with the confines of a virtual private cloud.

Ultimately there will always be some users who want a hardware platform with a server on top that they understand and feel comfortable managing.

We feel most new customers will be going to the cloud and will be ever more comfortable doing so.

Contact Us

We’re always interested in hearing from you with any comments or suggestions, feel free to get in touch.

 

Posted in Uncategorized | Leave a comment

No SQL required: Allowing end-users to use the data-lake

A feature always worth examining in more detail is the self-service querying of data objects in Sabisu Pipelines

It’s really easy for end-users to get value out of their data using Sabisu. Most users will never see the query they create as Sabisu hides the detail from all but the most technical.

When you select a Sabisu Data Object (SDO) it will show both pre-defined queries and all available tables, views and tags that a user may want to use in a specific, custom query.

Pre-defined queries are already parameterised so you need only to choose the key values you need, e.g., the tags and retrieval period. We estimate 90% of users will use pre-defined queries.

If you’re one of the 10% then custom queries will do what you need. Custom queries start by looking at the SDO to return the structure, allowing you to choose exactly which columns and data components you need in your Pipeline. You can also choose exactly how you want to filter your data.

Dynamic queries

You can check that the right data has been returned, adjust your filters and returned parameters to get precisely the data you need.

From that point you have your data and it’s just a case of building a Pipeline as  you would with any other data.

Contact Us

We’re always interested in hearing from you with any comments or suggestions, feel free to get in touch.

Posted in Release Notes | Tagged , , | Leave a comment

Sabisu Actions meets Android

Our latest deployment is a mobile version of Sabisu Actions available for all Android devices.

You can download it for free from Google Play Store. You’ll now be able to complete Sabisu Actions whenever and wherever you may be, even if you’ve no internet connection. If you complete an Action the update will push to the Sabisu platform as soon as internet connectivity has been restored.

actions in play

Once you’ve installed Sabisu Actions on your mobile device you’ll need to log in with your Sabisu account details. Any action assigned to you or a Community you’re in will appear in the app so you can start completing them as soon as you’re logged in.

Sabisu Actions is also available for installation on Plant Mobile devices such as the Getac z710. Please contact us if you would like to install Sabisu Actions on your intrinsically safe device.

MOBILE ACTIONS on android

Keep an eye out for the next release of Sabisu Actions which will allow you to create an Action on your mobile device.

Contact Us

We’re always interested in hearing from you with any comments or suggestions, feel free to get in touch.

 

Posted in Release Notes | Tagged , , , , | Leave a comment

Sabisu Bridge: Securely connecting your enterprise to the cloud

Last year we released Sabisu Go and Sabisu Bridge to help users easily connect enterprise data to the Sabisu cloud. They support the Sabisu Pipeline which gives end-users total control over their data, enabling them to pull together charts and templated reports easily.

Sabisu Bridge is well named; it’s a bridge between customer data and our cloud data-lake. After a 2 minute installation Bridge will pull data out on a configurable schedule, e.g every minute we get the last 2 minutes worth of data. This gives Bridge enhanced fault tolerance in the instance of connection failure; data is cached and once a connection is re-established the data will be pushed through.

Bridge High Level Overview

Bridge makes it easy to connect to external systems like OSISoft PI or AspenTech IP.21 but any data is in play; it’s also used to connect to SAP, MS Access, Oracle Primavera and others.

Transmission and writing to the cloud data-lake is very quick indeed thanks to a totally distributed, parallel processing architecture; customers are writing 120,000 rows/second of process data with ease.

When Sabisu Bridge has transmitted the data to the cloud it’s available immediately for you to query using the Sabisu Pipeline Self Service Query Generator.

In November last year, we introduced Sabisu Pipeline which gives end-users total control power their data, enabling them to pull together charts and templated reports from historians like IP.21 and OSISoft PI.

bridge pipelines

Bridge can also be used to connect data to an on-premise Unit – it’s an easy and quick way to extract data into Sabisu without traditional IT integration.

Currently you need a Sabisu Developer to set you up with Sabisu Bridge and carry out a few minutes of configuration. Over the coming weeks we’ll be releasing a revision which eliminates this configuration so that anyone can set up & run Bridge.

Get in touch if you’d like to join our beta user community.

 

Contact Us

We’re always interested in hearing from you with any comments or suggestions, feel free to get in touch.

 

 

Posted in Release Notes | Tagged , , , | Leave a comment