Short Term Active Power Load Prediction on A 33/11 kV Substation Using Regression Models

Round 1

Reviewer 1 Report

Paper 1 216 362

1. Brief summary of the paper:

The paper describes the problem of active power load prediction in medium voltage substation with the use of various regression models – simple linear, multiple linear and polynomial regression. The results of analysis with final conclusions were presented.   The proposed and described regression models predict the active power load with a good accuracy, better than other existing methods. 

2. Strong and weak points:

Paper is well prepared and is almost in accordance with Energies journal template. The survey of the literature sources was made.  

The paper in its introduction contains the description of electric power prediction tools used for forecasting the load and balance between it and power generation. The short, medium and long term prediction types are described and discussed. The obtained results are clearly described and given in schemes, figures and tables.   

I cannot see the weak points of the paper.   

3. Some minor/major recommendations/remarks should be taken into consideration for paper improvement, namely:

1. Lines 6; 92; 109; 193 and 342 – is “33/11KV” – there should be the space/distance between the numerical value of voltage and the unit, as well as “K” should be as lower case letter – like: “33/11 kV”.  This same in the title of the paper.
2. Line 7 – the name of Kakathiya University should start with upper case letter “K” and “U”.  
3. Lines 39; 203 and many others – is “….Table.1…” – should be “…Table 1…”, without the dot. Please correct this in the whole text.
4. Line 85 – there is “….e Levenberg- ….” – should be “…a…”.  
5. Figure 1 and others – all figures should be centered, as well as a caption on a single line should be centered – according to the ENERGIES magazine template demands. This same with tables. Please correct this in the whole text.
6. The references list should be shifted to left margin of the page – see the ENERGIES magazine template

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

This paper demonstrates use of linear/multilinear/polynomial regression with time correlation for predicting load data on a particular application.

While the results seem reasonable, the novelty of the approach may leave somewhat to be desired. The authors describe some very capable tools for capturing system behavior (ARIMA, neural nets, etc) but use significantly more basic methods in their application. While the fitting technique used is interesting, it does not strike me as state-of-the-art or expanding the field.

Perhaps if the authors were to describe why their approach is preferable in some way to the advanced methods described in their introduction, there could be a justification for using their methodology, and it could provide benefit to the field of predictive signal analysis as applied to energy engineering.

The grammar is overall fine, but some errors are notable enough that review by a technical editor would be recommended.

The results obtained are described somewhat heuristically despite using mean square error predictions; what degree of error is acceptable for the target applications? How does it compare with existing predictive tools? With the prevalence of data science tools today, there should be some justification of why this particular application stands out (speed, accuracy, simplicity, etc).

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The revisions made in response to the first review dramatically change my understanding of what the authors were hoping to achieve, and puts the methods and results in perspective. I believe myself and readers will always be interested in using simple models when they can perform well compared to more complex techniques. Thank you for your revision.