Application of Regression and ANN Models for Heat Pumps with Field Measurements
Round 1
Reviewer 1 Report
I have to suggest the following revisions:
- Add comparative figures (with bar plots) by using data from tables 4 and 5. Make obvious the comparison.
- I think you can provide in the appendix the data (or some of the data) of the manufacturer and the experiment.
- Add an extra figure about the impact of the various paramours on the COP. It is important to have parametric work in order to show the physical meaning of the various cases. The present Journal is in the area of energy engineering and thus this fact is needed.
- Give the coefficients of the approximation formulas in any table.
- In the last paragraph of the introduction, state why your work is novel compared to others from the literature.
- Add some extra references in the first paragraphs of the introduction.
- Add a detailed nomenclature.
- Please check the last line of table 1. The “UT” has to have the “T” as a subscript.
Author Response
Please see the attachment
Author Response File: Author Response.docx
Reviewer 2 Report
Authors have used five different regression models and two neural network models to compare how accurately they evaluate the COP value of an existing heat pump system. In general, the paper is quite well-written. However, I have one major comment regarding the objective of the paper. In addition, I have few detailed comments.
Authors state in the beginning of introduction. The main drawback of thermodynamic steady state models is that they require a large number of parameters related to the refrigerant and the components of the heat pump which are not available in many cases. I don’t fully understand what this argument is based on? No reference either. According to my knowledge, log h,p diagrams are available for a lot of different refrigerants (r134a, r22, r32 etc.). I also assume that electronic versions are available. In addition, it is very simple to define COP values online if condenser/evaporator and compression powers are known. On the basis of Table 1, these values are known and are measured constantly. Basically, you don’t even have to know refringent values to define the COP. You just have to measure water temperatures and mass flow rates before and after the heat pump to get the COP.
In my opinion, authors should justify better in introduction why regression or ANN models are needed for heat pumps if COP values can be easily defined online based on measured data. Authors mention in introduction that accurate models are needed to optimize the operation of heat pumps better. Authors should define more precisely what performance parameters they try to optimize, and especially why regression and ANN models are necessary to reach this goal instead of thermodynamic models. The reasoning in lines 53-55 is not sufficient.
Detailed comments
L 77 Authors should define Gray-box model a bit better.
L 116. What is back-box model? Authors have not defined it earlier in the text. Or should this perhaps be a black-box model? If so, what is the difference between the black-box and gray-box model?
Section 2. What refrigerant is used in heat pumps in Fig. 3?
Table 1. Is it really so that volume/mass flow rates of water have not been measured, or are they just missing in Table 1?
L 361 Apparently reference is missing?
Authors have not published any numerical data on values listed in Table 1. If possible, authors could show an example of typical numerical values listed in Table 1, for example some temporary operational values of typical range of values. However, the reviewer understands well if the data is confidential and cannot be published.
Author Response
Please see the attachment
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
The paper has been revised in a proper way.