Analysis and Optimization of Truck Windshield Defroster

Round 1

Reviewer 1 Report

The article presents design changes to duct paths in a commercial truck cabin leading to defrosting patterns on the windshield. The CFD methodology for producing outcomes are appropriate, but the method of analysis and reporting are inadequate unless major changes are made. The following are Major and Minor revisions requested by the reviewer:

Major Changes:

• There is no overarching approach to the introduction. The conclusion indicates both design and optimization, but it is unclear how those two tasks are different from one another. Also, it mentions evaluation of the performance but unclear whether the evaluation is based on time, location, temperature or other quantifiable properties.
• The methodology of CFD and theoretical equations is adequate, but the dimensions of the duct and temperature of the air are not specified. Furthermore, the thickness of the frost on the windshield is unknown to determine the amount of latent heat of fusion is necessary melt the ice for initial breakthrough.
• ALL figures with contour plots have a colorbar but no units are presented neither in the image nor in the caption. It is unclear whether the reader is viewing velocity, temperature, ice dimension, etc because the descriptions discuss different features within the results.
• The results indicate continuous optimization in Line 190, but the evaluation criteria are not identified whether it is temperature, area, perimeter, etc for determining defrosted zones. While the qualitative approach mentioned in Section 3. Model validation is OK for preliminary evaluation, the scientific evaluation of approaching optimal is not reported in Section 4 (i.e. a trendline of performance as function of design parameters is not provided).
• Figures 6 and 9 present one dimensional configuration of the vanes within the duct claiming that optimization is achieved. However, no performance plots of defrosting performance as a function of angle and/or position of the vane are provided. An evaluation at 25 minutes shown in Figure 11 indicates adequate performance for visibility, but does not confirm optimal nor any efficiency criteria. Clarifying the evaluation criteria mentioned in the first bullet point is critical to confirm the outcomes achieved.

Minor Changes:

• The paragraph in lines 33-38 is wandering and unclear at that point in the manuscript.
• The external citations read more as a list of references with the following sentence summarizing individually rather than a connected research thread that is being addressed by this submitted manuscript. Greater connectivity of the external sources would be beneficial.
• The inlet port for the duct in Figures 2, 6, & 9 are unclear and should be specified with annotation or quiver arrows.
• The theory section changes from Continuity to Continuous equation in EQ (1) and fails to specify what the variable P is in EQ (2).
• The results compare an "equal pressure loss method" to "minimum pressure loss method" as part of the design mechanism. Is this missing a citation or other standard design principle for these different descriptions?
• The sentence starting "Streamline design ..." on lines 189-190 is a fragment.
• The grey lines overlayed on the windshield appear in Figures 4, 5, 8, 10, 11 appear to be important since they are repeated, but the purpose is not specified anywhere. Also, Figure 11(a) has some extra diagonal lines that are unspecified.

Overall, the design improvement of the duct does appear to change performance. However, the claims of optimal performance are unclear to the reader unless the different input design parameters that are unspecified are reported with respect to the efficiency measure, which is also unspecified. Visual, qualitative inspection of the defrosted results after 25 minutes is an inadequate measure; something that reports area, perimeter, temperature as a function of time will confirm the optimal design was achieved for the changes in duct design features shown in Figures 6 and 9.

Author Response

Replies to the reviewers #1:

 

Major Changes:

Point 1: There is no overarching approach to the introduction. The conclusion indicates both design and optimization, but it is unclear how those two tasks are different from one another. Also, it mentions evaluation of the performance but unclear whether the evaluation is based on time, location, temperature or other quantifiable properties.

 

Response 1: In the section “introduction”, we have clearly described the current research problems in the defrosting of windshield and the current research work that investigates different defrosting or defogging performance for different types of vehicles. We have also highlighted the novelty of this paper in the last paragraph in Introduction (Lines 77 to 86, page 2).

The design and optimization are inter-connected. The optimization is part of the design in this paper. We first evaluated the defrosting performance of the windshield of a medium-size truck of Model N800 using CFD numerical method based on the commercial design. Then the defrosting duct is optimized to achieve the best defrosting performance on the windshield. As indicated in the paper, the evaluation of the defrosting performance is based on the latest Chinese national standard (GB11555-2009) which is when the defrosting system works for 25 minutes, more than 80% of the defrosting area on the windshield should complete defrosting. Therefore, the performance evaluation is based on the defrosting effect of a fixed time as stated in the manuscript (Lines 136 to 138, page 4 and Lines 177 to 178, page 6).

 

Point 2: The methodology of CFD and theoretical equations is adequate, but the dimensions of the duct and temperature of the air are not specified. Furthermore, the thickness of the frost on the windshield is unknown to determine the amount of latent heat of fusion is necessary melt the ice for initial breakthrough.

 

Response 2: As the reviewer suggested, we have now added the dimension of the defrost duct in Figure 2. Among them, the total width and height of the defrosting duct are mainly marked in Figure 2, and the size of the air inlet of the defrosting duct is highlighted in Figure 7. The heating air temperature is introduced according to the HVAC heating curve as shown in Figure 4. Fig. 4 is a new figure that has now been added in the paper and shows the HVAC heating curve, with an initial temperature of 253K. The initial thickness of the frost layer is set to 1 mm, which has been added in the manuscript (Line 142, Page 4).

 

Point 3: ALL figures with contour plots have a colorbar but no units are presented neither in the image nor in the caption. It is unclear whether the reader is viewing velocity, temperature, ice dimension, etc because the descriptions discuss different features within the results.

 

Response 3: In response to the reviewer's opinion, we have added the unit in all these figures with contour plots.

Point 4: The results indicate continuous optimization in Line 190, but the evaluation criteria are not identified whether it is temperature, area, perimeter, etc for determining defrosted zones. While the qualitative approach mentioned in Section 3. Model validation is OK for preliminary evaluation, the scientific evaluation of approaching optimal is not reported in Section 4 (i.e. a trendline of performance as function of design parameters is not provided).

Response 4: Thanks for your comment, in fact, there is no easy or simple corresponding function or curve between the combination of angle and position and the defrosting effect. As stated in the paper, we used pressure loss reduction simulation (Lines 187 to 188, page 6). The pressure loss largely affects the air velocity and flow distribution that determine the defrosting performance on the windshield. We adjusted the plate angles and positions to minimize the pressure loss and then evaluate the defrosting performance. For the best efficiency standard, we use to compare the defrosting area of the windshield at 25 minutes to determine the defrosting efficiency. By comparing the defrosting effects of different angles and positions, the best efficiency scheme is finally determined.

Minor Changes:

Point 1: The paragraph in lines 33-38 is wandering and unclear at that point in the manuscript. The external citations read more as a list of references with the following sentence summarizing individually rather than a connected research thread that is being addressed by this submitted manuscript. Greater connectivity of the external sources would be beneficial.

Response 1: We have revised lines 33-38. It now has a good connection between the sentence and paragraphs (Lines 31 to 39, page 1).

 

Point 2: The theory section changes from Continuity to Continuous equation in EQ (1) and fails to specify what the variable P is in EQ (2).

 

Response 2: We have made the continuity equations consistently in the text and EQ (1). In EQ (2), P is the pressure, the unit is Pa. We have now added the explanation to the manuscript (Lines 129 to 130).

Point 3: The results compare an "equal pressure loss method" to "minimum pressure loss method" as part of the design mechanism. Is this missing a citation or other standard design principle for these different descriptions?

Response 3: Both “equal pressure loss method” and “minimum pressure loss method” are commonly used in the design of air-conditioning ducts. The "equal pressure loss method" is an air duct calculation method based on the premise that the air duct per unit length has equal frictional resistance loss, so that the pressure loss of each air outlet of the defrost air duct is basically the same. According to the simulation results of the original defrosting duct, it can be found that the pressure loss of the outlet of each defrosting duct is basically the same, so it can be indicated that the original defrosting duct is designed according to the "equal pressure loss method". In the optimization model, we used the "minimum pressure loss method", the pressure loss of each air outlet is no longer the same, ensuring that the pressure loss at each outlet of the defrost air duct becomes the smallest.

Point 4: The sentence starting "Streamline design ..." on lines 189-190 is a fragment.

Response 4: As you suggested, we removed this statement to minimize the confusion.

Point 5: The grey lines overlayed on the windshield appear in Figures 4, 5, 8, 10, 11 appear to be important since they are repeated, but the purpose is not specified anywhere. Also, Figure 11(a) has some extra diagonal lines that are unspecified.

Response 5: The gray lines on figures 4, 5, 8, 10 and 11 (new figures 5, 6, 9, 11 and 12) represent defrosting areas A and B. As shown in figure 1, the defrosting area A represents the defrosting area in front of the driver, and the defrosting area B represents the total defrosting area of the front windshield. Figure 12(a) has some extra diagonal lines to assist in dividing the defrost area. We have added the explanation to the manuscript (Lines 90 to 92, Page 2).

Point 6: Overall, the design improvement of the duct does appear to change performance. However, the claims of optimal performance are unclear to the reader unless the different input design parameters that are unspecified are reported with respect to the efficiency measure, which is also unspecified. Visual, qualitative inspection of the defrosted results after 25 minutes is an inadequate measure; something that reports area, perimeter, temperature as a function of time will confirm the optimal design was achieved for the changes in duct design features shown in Figures 6 and 9.

Response 6: We have now clearly indicated the optimization function in the manuscript (Lines 187 to 190, page 7).

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors challenged an important issue in automobile technology and produced interesting computational results.
However, as a technical or scientific paper, many essential points are missing.
Please refer to the following comments.

2. Mathematical Model
The method of evaluating frost thickness is essential.
However, nothing is explained about it.
If heat transfer analysis was conducted, its details should be clarified.

No optimization strategy is explained.
How did the authors determine the shape of the deflector in Fig.6 and the angle shown in Fig.9(b)?
These are the most important points of the study, which the authors are supposed to discuss in details.

3. Model validation
Only qualitative comparison was made for validation.
Detailed quantitative comparison is required.
The agreement of the defrosting patterns between the experiment and simulation only indicates that the spots of flow impingement against the wind shield were well-predicted by CFD.

Figures (lack of explanations)
L173:"the two triangle zones as highlighted in the figure"
No such triangle zones can be seen in Fig.5.

Fig.6: More explanations are required about Fig.6.
Readers would be unable to understand what part of the system is shown in the figure.

Fig.9 is incomprehensible.
The figure seems to show some cut-plane; what plane in the system is depicted?
What do the authors mean to express with shading of the gray color?
What do the solid lines stand for?

Comments on specific sentences
L166: "This indicates that the design of the defrost duct is based on the equal pressure loss method in this truck."
This sentence implies that the original model was designed by the strategy of "equal pressure loss".
However, it is unlikely that the computational results with comparable pressure losses at all the exits were due to the design strategy.
The reviewer would advise the authors to reconsider the interpretation of the results.
Air must have the approximately equal total pressure across the inlet plane. On the other hand, the air originating from the inlet eventually reaches one of the exits to the cabin space.
Therefore, the observation that the total pressure loss is comparable regardless of the route simply reflects the fact that the pressure is comparable at all the exits.
Indeed, you can control the flow rate at each exit by some geometry change. However, the resultant total pressure loss would not differ greatly among the exits (as shown in Fig.7).

L181:
Although the authors pointed out that "the low air volume flow rate and the low speed at the air outlets CL and CR" were responsible for the poor defrosting performance in the original model, no quantitative information is shown about the flow rate or flow speed at the exits in the improved model with the optimized deflector.

 

Author Response

Replies to the reviewers #2:

Point 1: Mathematical Model. The method of evaluating frost thickness is essential. However, nothing is explained about it. If heat transfer analysis was conducted, its details should be clarified. No optimization strategy is explained. How did the authors determine the shape of the deflector in Fig. 6 and the angle shown in Fig.9(b)? These are the most important points of the study, which the authors are supposed to discuss in details.

Response 1: In this paper, evaluation of the defrosting performance on the windshield is based on the Chinese standard (GB11555-2009). In this standard, the frost thickness is 1 mm and the criterial for evaluating the defrosting performance is based on the defrosting areas on the windshield after the defrosting time of 25 mins. Therefore, the frost thickness is not investigated. We have now added this explanation in the manuscript (Lines 136 to 138, lines 142 to 143 on page 4 and lines 177 to 178 on page 6). The main optimization parameter of the defrosting duct is the different angles and positions of the deflector which affect the pressure loss, air velocity and flow distribution. In our simulation, we adjusted the angles and position to achieve the minimum pressure loss along the duct. This minimum pressure loss method directly links with the air velocity and flow distribution on the windshield which are the two key parameters affecting the defrosting performance. Once the optimal angle for the minimum pressure is selected, we evaluated the defrosting performance and compared it with the initial commercial design. This is our optimization approach. We have now clearly specified in the paper (Lines 187 to 190 on page 6).

 

Point 2: Model validation

Only qualitative comparison was made for validation. Detailed quantitative comparison is required. The agreement of the defrosting patterns between the experiment and simulation only indicates that the spots of flow impingement against the wind shield were well-predicted by CFD.

Response 2: Figure 5 shows a qualitative comparison between the simulation results and experimental data for a medium-size truck of Model N800. By comparison, the experimental results not only indicate the spots of flow impingement against the windshield, but also the defrosting effect (that is the distribution of the defrosting area shape on the window) of the defrosting 15 minutes and 20 minutes is the same as the CFD simulation results. For the comparison of the defrosting area results at 20 minutes, the defrosting area of the CFD simulation result accounts for 46% of the total windshield area. The defrosting area of the experiment is calculated by drawing the defrosting area diagram, and the defrosting area of the experimental results accounts for the total windshield area is about 48%. This can also be considered as quantitative effect. However, there is no way to make the quantitative comparison as the temperature of the frost is constant during the defrosting period.

Point 3: Figures (lack of explanations)：L173:"the two triangle zones as highlighted in the figure". No such triangle zones can be seen in Fig. 5.

Response 3: As shown in Figure 6 (Old Fig. 5), the two triangles are highlighted by the black circles. Because these two places on the two sides of the windshield limits both driver’s vision, the defrosting effect is not good. The shape is similar to an inverted triangle. We mark it with a black circle and call it the triangle area. We added an explanation to the manuscript (Line 180, page 6).

Point 4: Fig.6: More explanations are required about Fig.6. Readers would be unable to understand what part of the system is shown in the figure.

Response 4: This suggestion has been well-taken. In order to show the position change of the defrosting air duct deflector more clearly, Figure 7 (old figure 6) is a cross-sectional view of the central area of the defrosting duct (outlet CC). It also indicates the main dimensions of the defrosting air duct and the position and angle of the deflector. We have now added more explanation to the manuscript about this figure (Lines 191 to 193, page 6).

 

Point 5: Fig.9 is incomprehensible. The figure seems to show some cut-plane; what plane in the system is depicted? What do the authors mean to express with shading of the gray color? What do the solid lines stand for?

 

Response 5: In order to better show the position of the deflector of the defrosting duct, Figure 10 (old figure 9) is a sectional view of the thickness of the center of the defrost air duct. Since the defrost air duct is rendered after specifying the material on the 3D software, it will appear gray Shadow. The solid line represents the position of the deflector. We have now added the explanation to the manuscript (Lines 227 to 230 on page 8).

Point 6: L166: "This indicates that the design of the defrost duct is based on the equal pressure loss method in this truck." This sentence implies that the original model was designed by the strategy of "equal pressure loss". However, it is unlikely that the computational results with comparable pressure losses at all the exits were due to the design strategy. The reviewer would advise the authors to reconsider the interpretation of the results. Air must have the approximately equal total pressure across the inlet plane. On the other hand, the air originating from the inlet eventually reaches one of the exits to the cabin space. Therefore, the observation that the total pressure loss is comparable regardless of the route simply reflects the fact that the pressure is comparable at all the exits. Indeed, you can control the flow rate at each exit by some geometry change. However, the resultant total pressure loss would not differ greatly among the exits (as shown in Fig.7).

Response 6: Thanks for your suggestion. We have revised this part accordingly (Line 172, page 6).

Point 7: L181:Although the authors pointed out that "the low air volume flow rate and the low speed at the air outlets CL and CR" were responsible for the poor defrosting performance in the original model, no quantitative information is shown about the flow rate or flow speed at the exits in the improved model with the optimized deflector.

Response 7: As shown in Table 2, the air flow distribution and flow rate have been reported for each outlet. It obviously shows that the air flow and air velocity at CL and CR are lower. After the optimization, we have compared the flow distribution at each outlet before and after optimization in Table 3. Through the comparison of the flow distribution, we can see that the flow distribution of the optimized defrosting duct at outlets CL and CR increased by 47% and 22%, respectively (Lines 250 to 251, page 9).

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Figures 5, 6, 9 and 12 show the distributions of frost thickness.
The method of evaluating the frost thickness is still unclear, even after the authors added some explanations.
The authors claim "evaluation of the defrosting performance on the windshield is based on the Chinese standard (GB11555-2009)" in the response to the reviewer.
However, details of the method used in the study should be described in the manuscript, because not all readers are familiar with the standard in question.
Please make sure that readers would be able to reproduce the same results by following the procedures explained in the paper.

Author Response

Replies to the reviewers（Round 2） 

Point 1: Figures 5, 6, 9 and 12 show the distributions of frost thickness. The method of evaluating the frost thickness is still unclear, even after the authors added some explanations. The authors claim "evaluation of the defrosting performance on the windshield is based on the Chinese standard (GB11555-2009)" in the response to the reviewer.

However, details of the method used in the study should be described in the manuscript, because not all readers are familiar with the standard in question.

Please make sure that readers would be able to reproduce the same results by following the procedures explained in the paper.

Response 1: As we described in the manuscript, the evaluation of the defrosting performance on the windshield is based on the Chinese standard (GB11555-2009). In this standard, the initial frost thickness on the windshield is 1 mm. If the frost area is larger than 80% after defrosting the windshield for 25 minutes, then it could justify that the defrosting system meets the requirement in the standard. As for how the distribution of frost thickness is simulated, we have now added this part in the manuscript. This paper uses the enthalpy method to solve the transient problem of windshield defrosting. The cyclic solution of the equation can provide the distribution of the air temperature and heat flow on the windshield. In each step, the total energy from heat conduction, convection and radiation (usually negligible) is calculated and compared with the melting heat required for the frost layer to melt in the grid cell. When the total energy in the grid cell is the same as the heat of melting, the frost layer begins to melt. As the total energy in the grid cell is enough to melt all the frost into water, it is considered that there is no frost left in the grid cell. In this paper, the ice-water solid-liquid mixture is treated as a fluid, and the liquid phase fraction β is defined according to the solid temperature T_S and the liquid temperature T_L to determine the phase change status of each cell. This model is simplified, in the phase change process, the molten water moves slowly, the static ice-water mixture is directly used to approximate the simulation. The heat transfer of the heating curve (Figure 4) is introduced to solve the energy equations (7) ~ (10) of the temperature field to obtain the frost layer distribution on the windshield.

                               H=h+ΔH                   (7)

                        h=h_ref + ∫T_ref C_pdT           (8)

Where H is enthalpy, J/kg; h_ref is reference enthalpy, J/kg; T_ref is reference temperature,℃; C_p is constant pressure specific heat, J/(kg·K); ΔH is latent heat, J/kg.

Define the liquid phase fraction β as:

If ,T<T_S,β=0.

If ,T<T_L,β=1.

If ,T_S<T<T_L,β=(T-T_S)/(T_L-T_S).

The energy equation for the defrost model：

            ∂(ρH)/∂t + ∇·(ρvH) =∇·(k₁∇T)+S₁         (9)

Where v is the fluid velocity, m/s; k₁ is the thermal conductivity, W/(m·K);

                  S1=[(1+β)²/(β³+δ)]A_mф                (10)

Where β is liquid phase fraction, δ is a very small constant(0.001) to avoid division by 0, A_m is the area constant, ф is the solved turbulence (k, ε, ω, etc.).

We have now added this part in the revised manuscript.

 

Author Response File: Author Response.pdf