Prediction using Regression

[Ricky Joe Rescar]

Hi Everyone,

 

I have a case scenario where we would like to predict the number of days before a parameter reaches a set limit. For example, I wanted to see in my trend a dash line where my Tank Level would reach 80%. Correct me if I am wrong but the Predictive Tool for Seeq uses co-variance signals to generate a forecasted signal. What I wanted to achieve is to forecast the predictive value of a signal in the next succeeding days or hours.

 

Am I able to do so in SEEQ?

 

Regards,

Ricky

[Guest Jon Peterson]

Hi Ricky,

Thanks for your posting. Yes, Seeq can do this. The trick is a signal for time, such as hours in service, that signal would be a covariant in the regression. Presumably, the regression would be done as follows:

level would be the response variable (or signal to model)

covariants (independent variables) would be time, and flow rate or rates (if you have it). 

We have a heat exchanger fouling example in this webinar: https://info.seeq.com/webinar-how-to-use-advanced-analytics-in-the-oil-gas-industry. Slightly different, but it is a good example of projecting the model into the future. 

The key is, for any model to project in the future, you need some independent variable that projects into the future. The easy one is time. But you can imagine more sophisticated approaches, for example, based on production planning information in an ERP system, lookup or calculate presumed consumption of raw materials or the rate of production. 

I just wanted to quickly point you to the webinar so maybe you can get started. But, we'll get a Seeq Analytics Engineer to add to this. 

Regards,

Jon 

 

[Ricky Joe Rescar]

Hi Jon,

 

Thanks for giving the link for the webinar. I loved the webinar as it gave me insights about what other projects we can work at using SEEQ. I would also wait for the Analytics Engineers inputs,

 

Ricky