Experimental Study on the Effect of Heat-Retaining and Diversion Facilities on Thermal Discharge from a Power Plant

In order to reduce the influence of thermal discharge from the power plant on the surrounding water environment and the operation efficiency of the power plant, a distorted physical model was presented and applied to Huadian Kemen Power Plant for studying heat transport and analyzing the effects of heat-retaining and diversion facilities near the intake/outlet on the thermal discharge for six scenarios. Field investigations were also used to validate the model. This study is unique as it is the first to elaborate on the impact of heat-retaining and diversion facilities on thermal discharge. The results indicate that the construction of heat-retaining and diversion facilities can decrease the excess temperature at intake to meet the intake requirement and improve the distribution of low temperature rise, but the area of high temperature rise has an increase. When the heat-retaining wall and diversion dike were constructed, the maximum intake temperature rise of Phase III decreased significantly by 1.0–1.3 °C with an average decrease of 0.2 °C, and the maximum value of Phase I and II was reduced by 0.3 °C with little mean change. A comparative experiment with different construction heights was also conducted. Result analysis shows that when the crest elevation was reduced from 3 to 2 m, the influence on the intake temperature rise of Phase I and II could be ignored, and the average temperature rise of Phase III only had an increase of 0.1 °C, suggesting that constructions with 2 m play an effective role in reducing heat return to the intake.


Introduction
With the rapid economic development of coastal areas, more attention has been paid to ocean exploitation to address growing energy demands. Large power plants combined with new combustion/gasification technology [1][2][3] are often built in coastal areas and use the sea as cooling water, discharging a large amount of waste heat into the nearby seas [4]. The rise in water temperature and the water stratification caused by thermal discharge [5][6][7] have adverse impacts on the marine ecological environment [8][9][10][11]. Thermal pollution caused mainly by industrial use of water as a cooling agent has been regarded as a severe threat to ecological composition in coastal waters [12]. Meanwhile, a portion of the heated water directly returns to the intake, which negatively affects the unit operation efficiency [13,14]. The study on the hydraulic and thermal characteristics of thermal discharge is the basis to analyze the above problems.
The physical model is an important method to study thermal discharge characteristics. The mixing and dilution of heated water in near zones can be directly and truly reflected by experiments in the laboratory. The problem studied has occurred mainly in heat transport and the layouts of the power plant intake and outlet. Shawky et al. [15] tested two alternatives for outfall configuration, the surface open channel and multiport diffuser, to study the effect of thermal discharge by a scaled physical model. They indicated that the outfall of the diffusers could improve the mixing process and reduce thermal pollution to the allowable limits. Tian et al. [16] used the physical model with scale distortion to discuss the influence of intake retraction to bank on temperature rise. Intake retraction shortened the distance between intakes and outlets, the nearby water depth shoaled, the buoyant stratification caused by the temperature difference weakened and the direct recurrent of hot water increased, so that the intake temperature rise was more obvious, which is unfavorable to the plant. El-Ghorab [17] investigated the effects of effluents, excess temperature and flow regime on the mixing zone using a physical model, and established regression equations to predict the size of the mixing zone. The mixing zone area increases as the effluent increases and decreases as the flow in the river increases. It was found that the excess temperature at the outlet has more influence on the mixing zone. The work in [12] also provided a discussion on the effect of water discharge and the temperature at the outlet on the plume distribution. Zeng et al. [18] used the full-field physical model with scale distortion and the local physical model with a normal scale to analyze the detailed heat exchange phenomenon and discuss the effect of scale distortion. The results showed that compared with field measurements, the water temperature in coastal water was a little oversimulated near the surface and was a little undersimulated near the bottom of the heated-water layer by the full-field physical model. It was also shown that the model with scale distortion was applicable for simulating turbulent tidal flow and heat transport in coastal water in the case study. Shah et al. [19] used the physical model and numerical simulation to analyze the influence of flow rate and wind speed on the thermal plume. Shawky et al. [20] studied the effects of the main parameters such as the intake length, the distance between the intake and outlet, the water depth under the intake skimmer wall and the water depth just upstream the intake on the thermal characteristics using a physical model and numerical simulation. Two mathematical formulas indicating the relationship between these parameters and the hot water concentration at intake were deduced and verified.
At present, the majority of such studies have focused on the layout of the intake and outlet to reduce the impact of thermal discharge [21][22][23], and the effect of heat-retaining and diversion facilities has been ignored. The reasonable layout of the diversion dike and heat-retaining wall is an effective way to control heat transport and decrease excess temperature at intake, but there is a lack of detailed study on the effect of heat-retaining and diversion facilities. Therefore, in this paper, a distorted physical model is established and validated with field measurements. The objective is to study the influence of the construction and height of heat-retaining and diversion facilities near the intake/outlet on thermal discharge characteristics, including the temperature rise distribution and excess temperature at intake by scenario comparison. This paper is the first to present, in detail, the effects of heat-retaining and diversion facilities. The results include important reference values for similar engineering applications and scientific evaluation.

Project Overview
Luoyuan Bay is situated on the northeast coast of Fujian Province, China, surrounded by mountains and facing the sea in the northeast. The bay is wide and shallow with an area of 188.6 km 2 . The length from the baymouth to the bayhead is about 25 km, and the maximum water depth is 74 m. The mean width is 7 km and the mouth width is only 1.3 km. Luoyuan Bay is a typical semienclosed harbor that exchanges water with the open sea only through the Kemen Channel ( Figure 1). The wind and waves are weak, and there is no obvious impact of rivers or tributaries. According to the observed data, the tidal current featuring reciprocating flow is dominated by a regular semidiurnal tide. The mean tidal range is about 4.94 m and the maximum is 6.91 m. The flow velocity at the baymouth is larger than that in other areas. The maximum flood rate is 1.09 m s −1 and the maximum ebb rate is 1.17 m s −1 . The maximum velocity near the plant is about 0.70 m s −1 . During the spring tide, the maximum rate of the residual current is 0.13 m s −1 , and during the neap tide, the maximum is 0.06 m s −1 . The residual current near the plant comes along the channel to the inner bay, which is consistent with the flood current direction. maximum is 0.06 m s −1 . The residual current near the plant comes along the channel to the inner bay, which is consistent with the flood current direction. Huadian Kemen Power Plant is located on the south bank of Luoyuan Bay (Figure 1). The capacity of Phase I and II is 2400 MW, with the cooling water circulating at 81.68 m 3 s −1 . The capacity of Phase III is 2000 MW, and the units require a cooling flow of 65.66 m 3 s −1 . The phases of the power plant use seawater as the cooling water through their intakes and discharge the heated water into the bay at a temperature of 8.5 °C above the ambient temperature through their outfalls. Figure 2 shows the plane layout of intakes and outfalls. The intake heads of Phase I and Phase II are set next to each other, and the joint outlet is built on the northeast side of the plant. The layout of the deep intake and surface discharge is adopted in the prophase projects. The bottom elevation of the intake is −15.6 m and 1.9 m away from the bottom of the harbor basin to prevent sedimentation and sediment from entering. The top elevation of the intake is −6.6 m, which is 4.2 m away from the annual average low tide level. The diversion ditch is used to connect the intake and the pump house. An open channel for intake is used in Phase III, and the outlet is arranged on the east side of the prophase outlet. The bottom elevation of the open channel is about −10.3 m. In addition, the outlets of Phase I, II and III are equipped with diffusion facilities to dissipate energy and reduce the exit velocity. The baffle sill with a top elevation of −2.5 m is set at the tail of the outlet, and the stone apron with an elevation of −3.0 m is set behind the baffle sill.  Huadian Kemen Power Plant is located on the south bank of Luoyuan Bay (Figure 1). The capacity of Phase I and II is 2400 MW, with the cooling water circulating at 81.68 m 3 s −1 . The capacity of Phase III is 2000 MW, and the units require a cooling flow of 65.66 m 3 s −1 . The phases of the power plant use seawater as the cooling water through their intakes and discharge the heated water into the bay at a temperature of 8.5 • C above the ambient temperature through their outfalls. Figure 2 shows the plane layout of intakes and outfalls. The intake heads of Phase I and Phase II are set next to each other, and the joint outlet is built on the northeast side of the plant. The layout of the deep intake and surface discharge is adopted in the prophase projects. The bottom elevation of the intake is −15.6 m and 1.9 m away from the bottom of the harbor basin to prevent sedimentation and sediment from entering. The top elevation of the intake is −6.6 m, which is 4.2 m away from the annual average low tide level. The diversion ditch is used to connect the intake and the pump house. An open channel for intake is used in Phase III, and the outlet is arranged on the east side of the prophase outlet. The bottom elevation of the open channel is about −10.3 m. In addition, the outlets of Phase I, II and III are equipped with diffusion facilities to dissipate energy and reduce the exit velocity. The baffle sill with a top elevation of −2.5 m is set at the tail of the outlet, and the stone apron with an elevation of −3.0 m is set behind the baffle sill.

Materials and Methods
Water 2020, 12, x FOR PEER REVIEW 3 of 15 maximum is 0.06 m s −1 . The residual current near the plant comes along the channel to the inner bay, which is consistent with the flood current direction. Huadian Kemen Power Plant is located on the south bank of Luoyuan Bay ( Figure 1). The capacity of Phase I and II is 2400 MW, with the cooling water circulating at 81.68 m 3 s −1 . The capacity of Phase III is 2000 MW, and the units require a cooling flow of 65.66 m 3 s −1 . The phases of the power plant use seawater as the cooling water through their intakes and discharge the heated water into the bay at a temperature of 8.5 °C above the ambient temperature through their outfalls. Figure 2 shows the plane layout of intakes and outfalls. The intake heads of Phase I and Phase II are set next to each other, and the joint outlet is built on the northeast side of the plant. The layout of the deep intake and surface discharge is adopted in the prophase projects. The bottom elevation of the intake is −15.6 m and 1.9 m away from the bottom of the harbor basin to prevent sedimentation and sediment from entering. The top elevation of the intake is −6.6 m, which is 4.2 m away from the annual average low tide level. The diversion ditch is used to connect the intake and the pump house. An open channel for intake is used in Phase III, and the outlet is arranged on the east side of the prophase outlet. The bottom elevation of the open channel is about −10.3 m. In addition, the outlets of Phase I, II and III are equipped with diffusion facilities to dissipate energy and reduce the exit velocity. The baffle sill with a top elevation of −2.5 m is set at the tail of the outlet, and the stone apron with an elevation of −3.0 m is set behind the baffle sill.

•
For wide and shallow waters, it is necessary to adjust the model scale and use the distorted physical model [24,25]. Factors such as the simulation range, terrain condition, laboratory size and water supply capacity should be considered in determining the distorted scale. Due to the great amount of heated water discharged from the Huadian Kemen Power Plant and strong heat accumulation in Luoyuan Bay, the simulation range is relatively large. Generally, the average isotherm of 0.3 • C should be included in the model area. Based on the early numerical simulation results [26], the simulation range of 20 km × 8 km was determined (see Figure 1). Because of the great topographic variation and the large shoal waters in the range, the distortion ratio cannot be too small. However, considering the mixing properties in the near zone, a large distortion should also be avoided [27]. Finally, the distorted scale of 3.33 was selected with a horizontal scale of 400 and a vertical scale of 120 by comprehensive consideration.
where F r is Froude number, F d is densimetric Froude number, V is velocity, g is gravity acceleration, H is water depth, ρ is fluid density, ∆ρ is the density difference between heated water and ambient water and the subscript "r" denotes the ratio between prototype and model.
Based on the Froude number similarity, the densimetric Froude number similarity means that the relative density difference in both the prototype and model is equal. Assuming that the density difference linearly changes with the temperature difference, a temperature rise scale of 1 is required to meet the heat similarity.
According to the Froude number similarity, the velocity, time, discharge and roughness scales can be determined as follows: Velocity scale : Time scale : Discharge scale : Roughness scale : where V r is velocity scale between prototype and model, H r is vertical scale, t r is time scale, L r is horizontal scale, Q r is discharge scale, n r is roughness scale. The important model scales are shown in Table 1. Considering the similarities of turbulence and heat diffusion, the Reynolds number of the model exceeds the critical value. The flow is turbulent, and the flow pattern is similar.

Model Set-Up
According to the bathymetric and topographic survey, the bed level and horizontal outline of the model were determined. Afterwards, the physical model was carefully shaped with a horizontal scale of 400 and a vertical scale of 120. The bed surface of the model was coated with cement mortar, and the concrete structures, such as outlets and wharfs, were made of plexiglass to meet the roughness similarity. Figure 3 is the original photograph of the physical model, and the model panorama can be seen in Figure 3a. The length of the model was 50 m and the width was 20 m. The maximum water depth was about 0.62 m. The flow rate of the cooling water was 0.00028 m 3 s −1 to simulate the 147.34 m 3 s −1 rate of Phase I, II and III in the prototype.
Water 2020, 12, x FOR PEER REVIEW 5 of 15 According to the bathymetric and topographic survey, the bed level and horizontal outline of the model were determined. Afterwards, the physical model was carefully shaped with a horizontal scale of 400 and a vertical scale of 120. The bed surface of the model was coated with cement mortar, and the concrete structures, such as outlets and wharfs, were made of plexiglass to meet the roughness similarity. Figure 3 is the original photograph of the physical model, and the model panorama can be seen in Figure 3a. The length of the model was 50 m and the width was 20 m. The maximum water depth was about 0.62 m. The flow rate of the cooling water was 0.00028 m 3 s −1 to simulate the 147.34 m 3 s −1 rate of Phase I, II and III in the prototype. The physical model was constructed to study both the tidal currents and the temperature distribution. The tidal current was generated by the tide supply system, which was composed of a reservoir, pumps and valves. The tidal current was simulated by 64 submerged pumps, which were set on the upstream and downstream open boundaries to control inflow or outflow, and the flow rate was provided by numerical prediction. The closed circulation system for cooling water ( Figure  3b) consisted of the micropump, electric heater and float flowmeter.
Automatic tracking level meters with a precision of 0.02 mm were selected to measure the water level. The VECTRINO velocimeters ( Figure 3c) were used for tidal flow measurement. Figure 4a shows the layouts of the T1-T3 water level sites and the A1-A4 velocity sites. T1 and A1-A4, which are the same as the observation sites, can be used for model validation. T2 and T3 were added to get the water level at different positions in the whole simulation range. The LMT multipoint digital scanning temperature recorders were used for water temperature measurement. According to the different influences of thermal discharge, 256 temperature measuring probes ( Figure 3d) with a The physical model was constructed to study both the tidal currents and the temperature distribution. The tidal current was generated by the tide supply system, which was composed of a reservoir, pumps and valves. The tidal current was simulated by 64 submerged pumps, which were set on the upstream and downstream open boundaries to control inflow or outflow, and the flow rate was provided by numerical prediction. The closed circulation system for cooling water (Figure 3b) consisted of the micropump, electric heater and float flowmeter.
Automatic tracking level meters with a precision of 0.02 mm were selected to measure the water level. The VECTRINO velocimeters ( Figure 3c) were used for tidal flow measurement. Figure 4a shows the layouts of the T1-T3 water level sites and the A1-A4 velocity sites. T1 and A1-A4, which are the same as the observation sites, can be used for model validation. T2 and T3 were added to get the water level at different positions in the whole simulation range. The LMT multipoint digital scanning temperature recorders were used for water temperature measurement. According to the different influences of thermal discharge, 256 temperature measuring probes ( Figure 3d) with a precision of ±0.06 • C were set and most of them were set near the surface, as shown in Figure 4b. Due to the large tidal range in the study area, the probes were installed on the automatic tracking elevator with a precision of ±0.1-0.2 mm to fit the water depth variation. In addition, five vertical measuring lines, C1-C5, were arranged to measure the water temperature at different depths near the intake/outfall. C1-C4 were the same as the observation sites, and C5 near the intake of Phase III was added to get the intake temperature.
Water 2020, 12, x FOR PEER REVIEW 6 of 15 precision of ±0.06 °C were set and most of them were set near the surface, as shown in Figure 4b. Due to the large tidal range in the study area, the probes were installed on the automatic tracking elevator with a precision of ±0.1-0.2 mm to fit the water depth variation. In addition, five vertical measuring lines, C1-C5, were arranged to measure the water temperature at different depths near the intake/outfall. C1-C4 were the same as the observation sites, and C5 near the intake of Phase III was added to get the intake temperature.

Model Validation
The observed tidal current and water temperature data were obtained from a field survey that was performed by the Third Institute of Oceanography, Ministry of Natural Resources from November to December, 2012 with the operation of Phase I and II. The heated water discharge at the outlet of Phase I and II was 0.00016 m 3 s −1 to simulate the discharge of 81.68 m 3 s −1 in the prototype.
It was found by experiment that the flood tide current was mainly divided into two streams after passing through the Kemen Channel. One stream flowed along the northwest and the other reached the southwest bayhead. During the ebb tide, the currents dominated by the two streams converged in front of the channel and flowed to the open sea. The flood and ebb currents flowed in opposite directions and the property of reciprocating flow was obvious, which was consistent with previous studies [31][32][33][34]. Due to space limitations, this paper only shows the validation results of tide level, velocity and direction at station T1/A2 near the power plant during the spring and neap tides, as shown in Figure 5. The model simulation is in good agreement with the field measurement. It was concluded that the model can be used for further studies on heat transport and diffusion. precision of ±0.06 °C were set and most of them were set near the surface, as shown in Figure 4b. Due to the large tidal range in the study area, the probes were installed on the automatic tracking elevator with a precision of ±0.1-0.2 mm to fit the water depth variation. In addition, five vertical measuring lines, C1-C5, were arranged to measure the water temperature at different depths near the intake/outfall. C1-C4 were the same as the observation sites, and C5 near the intake of Phase III was added to get the intake temperature.

Model Validation
The observed tidal current and water temperature data were obtained from a field survey that was performed by the Third Institute of Oceanography, Ministry of Natural Resources from November to December, 2012 with the operation of Phase I and II. The heated water discharge at the outlet of Phase I and II was 0.00016 m 3 s −1 to simulate the discharge of 81.68 m 3 s −1 in the prototype.
It was found by experiment that the flood tide current was mainly divided into two streams after passing through the Kemen Channel. One stream flowed along the northwest and the other reached the southwest bayhead. During the ebb tide, the currents dominated by the two streams converged in front of the channel and flowed to the open sea. The flood and ebb currents flowed in opposite directions and the property of reciprocating flow was obvious, which was consistent with previous studies [31][32][33][34]. Due to space limitations, this paper only shows the validation results of tide level, velocity and direction at station T1/A2 near the power plant during the spring and neap tides, as shown in Figure 5. The model simulation is in good agreement with the field measurement. It was concluded that the model can be used for further studies on heat transport and diffusion. precision of ±0.06 °C were set and most of them were set near the surface, as shown in Figure 4b. Due to the large tidal range in the study area, the probes were installed on the automatic tracking elevator with a precision of ±0.1-0.2 mm to fit the water depth variation. In addition, five vertical measuring lines, C1-C5, were arranged to measure the water temperature at different depths near the intake/outfall. C1-C4 were the same as the observation sites, and C5 near the intake of Phase III was added to get the intake temperature.

Model Validation
The observed tidal current and water temperature data were obtained from a field survey that was performed by the Third Institute of Oceanography, Ministry of Natural Resources from November to December, 2012 with the operation of Phase I and II. The heated water discharge at the outlet of Phase I and II was 0.00016 m 3 s −1 to simulate the discharge of 81.68 m 3 s −1 in the prototype.
It was found by experiment that the flood tide current was mainly divided into two streams after passing through the Kemen Channel. One stream flowed along the northwest and the other reached the southwest bayhead. During the ebb tide, the currents dominated by the two streams converged in front of the channel and flowed to the open sea. The flood and ebb currents flowed in opposite directions and the property of reciprocating flow was obvious, which was consistent with previous studies [31][32][33][34]. Due to space limitations, this paper only shows the validation results of tide level, velocity and direction at station T1/A2 near the power plant during the spring and neap tides, as shown in Figure 5. The model simulation is in good agreement with the field measurement. It was concluded that the model can be used for further studies on heat transport and diffusion. precision of ±0.06 °C were set and most of them were set near the surface, as shown in Figure 4b. Due to the large tidal range in the study area, the probes were installed on the automatic tracking elevator with a precision of ±0.1-0.2 mm to fit the water depth variation. In addition, five vertical measuring lines, C1-C5, were arranged to measure the water temperature at different depths near the intake/outfall. C1-C4 were the same as the observation sites, and C5 near the intake of Phase III was added to get the intake temperature.

Model Validation
The observed tidal current and water temperature data were obtained from a field survey that was performed by the Third Institute of Oceanography, Ministry of Natural Resources from November to December, 2012 with the operation of Phase I and II. The heated water discharge at the outlet of Phase I and II was 0.00016 m 3 s −1 to simulate the discharge of 81.68 m 3 s −1 in the prototype.
It was found by experiment that the flood tide current was mainly divided into two streams after passing through the Kemen Channel. One stream flowed along the northwest and the other reached the southwest bayhead. During the ebb tide, the currents dominated by the two streams converged in front of the channel and flowed to the open sea. The flood and ebb currents flowed in opposite directions and the property of reciprocating flow was obvious, which was consistent with previous studies [31][32][33][34]. Due to space limitations, this paper only shows the validation results of tide level, velocity and direction at station T1/A2 near the power plant during the spring and neap tides, as shown in Figure 5. The model simulation is in good agreement with the field measurement. It was concluded that the model can be used for further studies on heat transport and diffusion. precision of ±0.06 °C were set and most of them were set near the surface, as shown in Figure 4b. Due to the large tidal range in the study area, the probes were installed on the automatic tracking elevator with a precision of ±0.1-0.2 mm to fit the water depth variation. In addition, five vertical measuring lines, C1-C5, were arranged to measure the water temperature at different depths near the intake/outfall. C1-C4 were the same as the observation sites, and C5 near the intake of Phase III was added to get the intake temperature.

Model Validation
The observed tidal current and water temperature data were obtained from a field survey that was performed by the Third Institute of Oceanography, Ministry of Natural Resources from November to December, 2012 with the operation of Phase I and II. The heated water discharge at the outlet of Phase I and II was 0.00016 m 3 s −1 to simulate the discharge of 81.68 m 3 s −1 in the prototype.
It was found by experiment that the flood tide current was mainly divided into two streams after passing through the Kemen Channel. One stream flowed along the northwest and the other reached the southwest bayhead. During the ebb tide, the currents dominated by the two streams converged in front of the channel and flowed to the open sea. The flood and ebb currents flowed in opposite directions and the property of reciprocating flow was obvious, which was consistent with previous studies [31][32][33][34]. Due to space limitations, this paper only shows the validation results of tide level, velocity and direction at station T1/A2 near the power plant during the spring and neap tides, as shown in Figure 5. The model simulation is in good agreement with the field measurement. It was concluded that the model can be used for further studies on heat transport and diffusion.
: the open boundary with the pumps).

Model Validation
The observed tidal current and water temperature data were obtained from a field survey that was performed by the Third Institute of Oceanography, Ministry of Natural Resources from November to December, 2012 with the operation of Phase I and II. The heated water discharge at the outlet of Phase I and II was 0.00016 m 3 s −1 to simulate the discharge of 81.68 m 3 s −1 in the prototype.
It was found by experiment that the flood tide current was mainly divided into two streams after passing through the Kemen Channel. One stream flowed along the northwest and the other reached the southwest bayhead. During the ebb tide, the currents dominated by the two streams converged in front of the channel and flowed to the open sea. The flood and ebb currents flowed in opposite directions and the property of reciprocating flow was obvious, which was consistent with previous studies [31][32][33][34]. Due to space limitations, this paper only shows the validation results of tide level, velocity and direction at station T1/A2 near the power plant during the spring and neap tides, as shown in Figure 5. The model simulation is in good agreement with the field measurement. It was concluded that the model can be used for further studies on heat transport and diffusion. The temperature analysis showed that during the observation period (in winter), the water temperature in the bay was lower than the external temperature, and the surface water temperature was lower than that of the bottom. The water temperature was greatly affected by the tide. The temperature variation with time, presented as two peaks and two lows in a tidal day, was synchronous with flux and reflux. The water temperature was higher at a high tide level and lower at a low tide level.
Because some units did not operate in some observation periods of the spring tide and middle tide, the average output of the power plant only accounted for 42% of the installed capacity. The temperature rise area caused by thermal discharge was relatively small. It was approximately considered that thermal discharge had little impact on the temperature rise at the far sites of C1 and C4, and the natural climate was the main reason for temperature change at C1 and C4 for the period. The natural temperature variation of C2 near the intake was obtained by data interpolation of C1 and C4. Then, the measured values of temperature rise at the intake were acquired using the natural temperature and observed data of C2 at adverse the neap tide (output of 1464 MW). Table 2 shows the comparison of the intake temperature rise of Phase I and II between the experimental simulation and field measurement. The depth-averaged values of the model were in good agreement with field measurements. There was a deviation in the surface temperature rise, but the relative error was less than 20%. Due to the complex influence of natural climate change and actual plant operation conditions, etc., the deviation was within a reasonable range. The model is thus available for predicting thermal discharge characteristics.

Layout Principle of Heat-Retaining and Diversion Facilities
In addition to the direct removal by environmental water and heat exchange with air, part of the waste heat returns to the intake, resulting in an increase in intake water temperature and a decrease in plant efficiency. Therefore, in the engineering design stage, building a heat-retaining The temperature analysis showed that during the observation period (in winter), the water temperature in the bay was lower than the external temperature, and the surface water temperature was lower than that of the bottom. The water temperature was greatly affected by the tide. The temperature variation with time, presented as two peaks and two lows in a tidal day, was synchronous with flux and reflux. The water temperature was higher at a high tide level and lower at a low tide level.
Because some units did not operate in some observation periods of the spring tide and middle tide, the average output of the power plant only accounted for 42% of the installed capacity. The temperature rise area caused by thermal discharge was relatively small. It was approximately considered that thermal discharge had little impact on the temperature rise at the far sites of C1 and C4, and the natural climate was the main reason for temperature change at C1 and C4 for the period. The natural temperature variation of C2 near the intake was obtained by data interpolation of C1 and C4. Then, the measured values of temperature rise at the intake were acquired using the natural temperature and observed data of C2 at adverse the neap tide (output of 1464 MW). Table 2 shows the comparison of the intake temperature rise of Phase I and II between the experimental simulation and field measurement. The depth-averaged values of the model were in good agreement with field measurements. There was a deviation in the surface temperature rise, but the relative error was less than 20%. Due to the complex influence of natural climate change and actual plant operation conditions, etc., the deviation was within a reasonable range. The model is thus available for predicting thermal discharge characteristics.

Layout Principle of Heat-Retaining and Diversion Facilities
In addition to the direct removal by environmental water and heat exchange with air, part of the waste heat returns to the intake, resulting in an increase in intake water temperature and a decrease in plant efficiency. Therefore, in the engineering design stage, building a heat-retaining wall near the intake to reduce the heated water recycling is often considered. The diversion dike is usually arranged near the outlet to control the flow direction. It can be made into a curved type and a linear type, and the crest elevation is generally not lower than the design water level. In the study of thermal discharge, the diversion dike can not only be used to guide the flow and affect the heated water zone, but also can function as a heat-retaining wall.

Scenario Description
The construction of heat-retaining and diversion facilities can promote heat diffusion downstream and reduce the direct heat return, but it may also cause heat accumulation in the near-field. In this paper, combined with the Huadian Kemen Power Plant, a physical model was conducted to study the influences of heat-retaining wall and diversion dike when Phase I, II and III were all in operation. As shown in Figure 2, the heat-retaining wall was set on the side of the Phase III intake open channel close to the outlet, and the diversion dike was arranged near the outlet. Because the measured maximum water level was 3.04 m, the constructions could fully operate when the crest elevation was 3 m. However, the engineering cost should also be considered in practice. In order to explore the effect of the constructions and their heights on the thermal discharge characteristics, 6 scenarios were designed, and 12 tests, including the spring tide and neap tide, were carried out, as listed in Table 3.

Thermal Properties without the Constructions
Before the construction of heat-retaining and diversion facilities (Scenario 1), the heated water could easily transport and diffuse upstream and downstream under the influence of reciprocating current and residual flow. Figure 6 shows the distribution of the maximum and average temperature rises at the surface during the spring and neap tides. Due to the high-velocity flow of the baymouth and north deep-water area, the heated water was mainly accumulated in the south shore with a zonal distribution. Compared with the neap tide, the tidal dynamic of the spring tide was stronger, which led to greater heat transport with the tide. The heat accumulation was obvious in the bay during the neap tide. The area of temperature rise was affected by the above two conditions. Under the present installed capacity, the temperature rise area of 0.5 • C in the spring tide was larger than that in the neap tide, but the heat still did not reach the north bank of the baymouth. The area of 1 • C had little difference and the 4 • C area in the neap tide was larger than that in the spring tide (Figure 7).
The excess temperature at the intake varied periodically with the tide, and the variation trends of the spring and neap tides were similar. The maximum occurred at the early stage of the flood tide. At this time, the high temperature rise zone flowed towards the intake with the flood current. About two hours after the low tide level, the intake temperature rise of Phase I and II reached the highest point, about 1.34-1.53 • C in the spring and neap tides, and the mean was 0.75-0.90 • C. Because of the distance of the intake of Phase III from the outfalls, the maximum lagged by a half hour. The intake temperature rises of the spring tide and neap tide were the same, and the maximum was 2.15 • C with a mean of 0.98 • C (Table 4).
In general, for the semienclosed bay, heat accumulation was more obvious with an increase of thermal discharge, and neap tide was the adverse tide.
According to the measured data, the highest natural water temperature in summer was 31.2 • C, so the highest intake temperature of Phase I and II was 32.7 • C, and the maximum of Phase III was 33.4 • C which did not meet the cooling water intake requirement of the steam turbine unit (≤33 • C) [35]. In order to ensure normal unit operation, the intake temperature rise should not exceed 1.8 • C. Therefore, it is necessary to build heat-retaining and diversion facilities to reduce the intake temperature. Because of the distance of the intake of Phase III from the outfalls, the maximum lagged by a half hour. The intake temperature rises of the spring tide and neap tide were the same, and the maximum was 2.15 °C with a mean of 0.98 °C (Table 4). In general, for the semienclosed bay, heat accumulation was more obvious with an increase of thermal discharge, and neap tide was the adverse tide.
According to the measured data, the highest natural water temperature in summer was 31.2 °C, so the highest intake temperature of Phase I and II was 32.7 °C, and the maximum of Phase III was 33.4 °C which did not meet the cooling water intake requirement of the steam turbine unit (≤33 °C) [35]. In order to ensure normal unit operation, the intake temperature rise should not exceed 1.8 °C Therefore, it is necessary to build heat-retaining and diversion facilities to reduce the intake temperature.

Effect of Heat-Retaining and Diversion Facilities on Temperature Rise Distribution
The influences of heat-retaining and diversion facilities were studied by comparing Scenarios 1-4. Figure 7 shows the scenario comparison of maximum and average temperature rise areas at the surface during the spring and neap tides. In order to explore the influence on the vertical temperature rise near the power plant, temperature rise variations of different depths at C2 are presented in Figure 8.
The relative position between the facilities and outlets was the key factor affecting the high temperature rise area. The diversion dike near the outlet (Scenario 2) blocks the heat diffusion upstream and leads to the heat distribution close to the shoreline. The high temperature water is concentrated in the shallow waters behind the wharfs, and the area has an obvious increase. Compared with Scenario 1, the maximum area of 4 °C in the spring and neap tides increased by 0.4-0.6 km 2 with an increase of 59%, and the average area of 4 °C increased by 0.2-0.3 km 2 , about three times larger. Because the heat-retaining wall (Scenario 3) was a little far away from the outlet, the effect on the high temperature rise area was small. The maximum area of 4 °C had a slight increase of 0.1-0.2 km 2 with the change range not exceeding 20% of Scenario 1, and the average area variation was inconspicuous.
For the area of low temperature rise, there was a general decreasing trend when the heat-retaining and diversion facilities were constructed. The diversion dike accelerated the heat transport to the baymouth where the heat exchange was stronger, so the low temperature rise area

Effect of Heat-Retaining and Diversion Facilities on Temperature Rise Distribution
The influences of heat-retaining and diversion facilities were studied by comparing Scenarios 1-4. Figure 7 shows the scenario comparison of maximum and average temperature rise areas at the surface during the spring and neap tides. In order to explore the influence on the vertical temperature rise near the power plant, temperature rise variations of different depths at C2 are presented in Figure 8.
The relative position between the facilities and outlets was the key factor affecting the high temperature rise area. The diversion dike near the outlet (Scenario 2) blocks the heat diffusion upstream and leads to the heat distribution close to the shoreline. The high temperature water is concentrated in the shallow waters behind the wharfs, and the area has an obvious increase. Compared with Scenario 1, the maximum area of 4 • C in the spring and neap tides increased by 0.4-0.6 km 2 with an increase of 59%, and the average area of 4 • C increased by 0.2-0.3 km 2 , about three times larger. Because the heat-retaining wall (Scenario 3) was a little far away from the outlet, the effect on the high temperature rise area was small. The maximum area of 4 • C had a slight increase of 0.1-0.2 km 2 with the change range not exceeding 20% of Scenario 1, and the average area variation was inconspicuous.
For the area of low temperature rise, there was a general decreasing trend when the heat-retaining and diversion facilities were constructed. The diversion dike accelerated the heat transport to the baymouth where the heat exchange was stronger, so the low temperature rise area was reduced, especially in the spring tide. Compared with Scenario 1, the maximum area of 1 • C had a 44% reduction, about 6.9 km 2 . The average area of 1 • C during the spring and neap tides had a 30% reduction, about 0.7-1.5 km 2 . However, because the heat retention between the intake and outlet was caused by the heat-retaining wall, the area of low temperature rise in Scenario 3 had an increase during the unfavorable neap tide. The increase was relatively small and the change of maximum and mean temperature rise areas of 1 • C was about 16%-23% of Scenario 1.
When the diversion dike and the heat-retaining wall were constructed (Scenario 4), the two effects were superimposed. Compared with Scenario 1, the low temperature rise area was reduced but the high temperature rise area had an increase. The maximum area of 1 • C during the spring and neap tides had a 39% decrease, about 3.6-8.5 km 2 . The mean area of 1 • C was reduced by 0.5-1.3 km 2 with the reduction of 24%. The maximum area of 4 • C increased by about 0.5 km 2 with an increase of 53%, and the average temperature rise area of 4 • C increased by 0.2 km 2 .
In the vicinity of the power plant (C2), the water depth changes greatly along the transverse direction. This belongs to the heat transition zone with great buoyancy effects. As shown in Figure 8, the thermal stratification was very obvious in Scenario 1. After building the heat-retaining and diversion facilities, vertical mixing was stronger and the thermal stratification was weakened. In addition, the temperature rise changes greatly with time in a tidal day. Because C2 is located on the west side of the outlets, the heated water discharged flows to C2 under the influence of the flood current, leading to the significant fluctuation of water temperature at C2 during the flood tide.
Water 2020, 12, x FOR PEER REVIEW 10 of 15 a 30% reduction, about 0.7-1.5 km 2 . However, because the heat retention between the intake and outlet was caused by the heat-retaining wall, the area of low temperature rise in Scenario 3 had an increase during the unfavorable neap tide. The increase was relatively small and the change of maximum and mean temperature rise areas of 1 °C was about 16%-23% of Scenario 1. When the diversion dike and the heat-retaining wall were constructed (Scenario 4), the two effects were superimposed. Compared with Scenario 1, the low temperature rise area was reduced but the high temperature rise area had an increase. The maximum area of 1 °C during the spring and neap tides had a 39% decrease, about 3.6-8.5 km 2 . The mean area of 1 °C was reduced by 0.5-1.3 km 2 with the reduction of 24%. The maximum area of 4 °C increased by about 0.5 km 2 with an increase of 53%, and the average temperature rise area of 4 °C increased by 0.2 km 2 .
In the vicinity of the power plant (C2), the water depth changes greatly along the transverse direction. This belongs to the heat transition zone with great buoyancy effects. As shown in Figure 8, the thermal stratification was very obvious in Scenario 1. After building the heat-retaining and diversion facilities, vertical mixing was stronger and the thermal stratification was weakened. In addition, the temperature rise changes greatly with time in a tidal day. Because C2 is located on the west side of the outlets, the heated water discharged flows to C2 under the influence of the flood current, leading to the significant fluctuation of water temperature at C2 during the flood tide.     Table 4 shows the scenario comparison of the intake temperature rise. Except for Scenario 1, the results all met the limit of 1.8 • C. Figure 9 shows the scenario comparison of temperature rise variations at the intake in the adverse neap tide. Table 4 and Figure 9 show that the diversion dike built near the outfall (Scenario 2) effectively reduced heat return to the intake of Phase I and II by discharging heated water downstream. Compared with no diversion dike, the maximum decreased by about 0.3 • C, and the average value was reduced by about 0.1 • C. Because the diversion dike is far away from the intake open channel, it has little impact on the intake temperature rise of Phase III during the neap tide. Figure 9 indicates that except for the difference of maximum, the intake temperature rise variations of Phase III were mostly the same before and after the construction of the diversion dike.

Effect of Heat-Retaining and Diversion Facilities on Excess Temperature at Intake
When the heat-retaining wall was constructed on the side of the intake open channel (Scenario 3), the heated water was blocked outside the wall during the flood tide, which directly prevented the heat return to the intake channel. Compared with Scenario 1, the intake temperature rise of Phase III had an obvious decline, and the maximum decreased by 1 • C with an average reduction of 0.2 • C. As shown in Figure 9, the peak decreased distinctly after the construction of the heat-retaining wall and the variation tended to be gentle. The temperature rise fluctuation was small, and the difference between maximum and minimum was less than 0.5 • C. The rule of temperature rise varying with tide was unapparent. For the intake temperature rise of Phase I and II, due to the barrier of the wall, the retention time of heated water between the intake and outlet was prolonged, and the heat accumulation during the spring tide was enhanced slightly. The time increase in high temperature at intake led to the average temperature rise of 0.1 • C.
The diversion dike near the outlet mainly reduced the intake temperature rise of Phase I and II, while the heat-retaining wall near the open channel had the main influence on Phase III. When the heat-retaining wall and diversion dike were built (Scenario 4), the two effects were superimposed and the combined effect was the best. Compared with Scenario 1, the intake temperature rises of Phase I, II, and III were improved and the influence on Phase III was more prominent. The blocking effect of the heat-retaining wall and the leading of the diversion dike effectively decreased the direct heat return to the intake of Phase III. The peak of the intake temperature rise was clearly decreased and the temperature rise varied slowly with the tide. The maximum of Phase III was reduced by 1.0-1.3 • C and the mean reduction was 0.2 • C. For the intake temperature rise of Phase I and II, the temperature rise variation with the tide was basically the same as Scenario 1. The two effects of the heat-retaining wall and the diversion dike were partially offset, resulting in the maximum being reduced by about 0.3 • C and an average effect that was not obvious.

Effect of Construction Height on Excess Temperature at Intake
In addition to ensuring that the constructions were fully effective, the engineering investment also should be taken into account in actual engineering. Considering the crest elevations of 3 m, 2 m and 1 m, the impact on the intake temperature rise was examined. Through contrastive analysis of Scenarios 4, 4-1 and 4-2, the effect of the construction height was studied.
With the reduction of the crest elevation, overtopping occurs during the high tide. Then, the heat-retaining effect is reduced, and the thermal stratification is weaker with stronger vertical mixing. Figure 10 shows excess temperature variations at the intake during the spring and neap tides.
When the crest elevation had reduced to 2 m (two-thirds of the high tide level), the overtopping time was about 4-6 h during the spring and neap tides. Due to the short duration, the unfavorable impact was relatively small. For Phase I and II, the intake temperature rise variation of 2 m crest elevation was generally consistent with that of 3 m. The difference of Phase III between 2 m and 3 m crest elevations had a slight increase and it was more obvious during the rising tide. The maximum increased by 0.1-0.3 °C with a mean increase of 0.1 °C. The small difference indicates that the constructions with 2 m crest elevations still effectively reduce heat return.
When the crest elevation had reduced to 1 m (one-third of the high tide level), the effects of heat-retaining and diversion facilities were greatly reduced. The overtopping time was about 10 h, accounting for about two-fifths of a tidal day. Compared with the 3 m crest elevation, the period of high temperature rise at the intake increased significantly and the vertical mixing was stronger, resulting in the apparent increase of intake temperature rise. The average intake temperature rise of Phase I and II increased by 0.15 °C, and the mean of Phase III increased by 0.3 °C. The minimum and maximum also increased greatly. The results show that the crest elevation of 1 m could not

Effect of Construction Height on Excess Temperature at Intake
In addition to ensuring that the constructions were fully effective, the engineering investment also should be taken into account in actual engineering. Considering the crest elevations of 3 m, 2 m and 1 m, the impact on the intake temperature rise was examined. Through contrastive analysis of Scenarios 4, 4-1 and 4-2, the effect of the construction height was studied.
With the reduction of the crest elevation, overtopping occurs during the high tide. Then, the heat-retaining effect is reduced, and the thermal stratification is weaker with stronger vertical mixing. Figure 10 shows excess temperature variations at the intake during the spring and neap tides.
When the crest elevation had reduced to 2 m (two-thirds of the high tide level), the overtopping time was about 4-6 h during the spring and neap tides. Due to the short duration, the unfavorable impact was relatively small. For Phase I and II, the intake temperature rise variation of 2 m crest elevation was generally consistent with that of 3 m. The difference of Phase III between 2 m and 3 m crest elevations had a slight increase and it was more obvious during the rising tide. The maximum increased by 0.1-0.3 • C with a mean increase of 0.1 • C. The small difference indicates that the constructions with 2 m crest elevations still effectively reduce heat return.

Conclusions
A distorted physical model was conducted and applied to the Huadian Kemen Power Plant in Luoyuan Bay to study the hydraulic and thermal characteristics of thermal discharge. Based on the validation of the model with measured data, the influences of heat-retaining and diversion facilities and their height on thermal discharge were studied. The results obtained are summarized as follows.
1. Heat-retaining and diversion facilities could prevent the heated water from spreading upstream during flood tide, and effectively reduce the heat direct return. The diversion dike also promotes heat transport to outside the bay. The heat-retaining and diversion facilities strengthen the vertical mixing near the plant and weaken the thermal stratification. Therefore, they have certain impacts on the heat distribution and intake temperature rise. 2. The heat-retaining and diversion facilities successfully reduced the intake temperature rise and complied with the intake requirement. Due to the different construction positions and functions, the diversion dike and the heat-retaining wall mainly reduced the intake temperature rise of prophase and Phase III, respectively. When the diversion dike and heat-retaining wall were both constructed, the greatest effects of superposition were obtained. The maximum of Phase III decreased significantly by 1.0-1.3 °C with an average reduction of 0.2 °C, and the maximum of Phase I and II decreased by 0.3 °C with little mean change. 3. From an economic perspective, the crest elevations of 2 m and 1 m were considered and compared with 3 m to analyze the impact of height. The results show that when the crest elevation was reduced to 2 m, due to the short overtopping time of 4-6 h the intake temperature rise of Phase I and II had no obvious change and the average temperature rise of Phase III increased by 0.1 °C. The constructions with 2 m crest elevations still played an effective role in reducing heat return. 4. The construction of heat-retaining and diversion facilities could effectively reduce the area of low temperature rise. Compared with no facilities, the maximum area of 1 °C was reduced by 3.6-8.5 km 2 with a decrease of 39% and the average area of 1 °C was reduced by 0.5-1.3 km 2 with a decrease of 24%. However, the area of high temperature rise became larger, and the influence of the diversion dike near the outfall was more obvious. Because the heat-retaining wall was far away from the outlet, it had little effect. The change range of the maximum area of 4 °C was less than 20%, and the average area variation of 4 °C was unapparent. When the crest elevation had reduced to 1 m (one-third of the high tide level), the effects of heat-retaining and diversion facilities were greatly reduced. The overtopping time was about 10 h, accounting for about two-fifths of a tidal day. Compared with the 3 m crest elevation, the period of high temperature rise at the intake increased significantly and the vertical mixing was stronger, resulting in the apparent increase of intake temperature rise. The average intake temperature rise of Phase I and II increased by 0.15 • C, and the mean of Phase III increased by 0.3 • C. The minimum and maximum also increased greatly. The results show that the crest elevation of 1 m could not effectively reduce heat return.

Conclusions
A distorted physical model was conducted and applied to the Huadian Kemen Power Plant in Luoyuan Bay to study the hydraulic and thermal characteristics of thermal discharge. Based on the validation of the model with measured data, the influences of heat-retaining and diversion facilities and their height on thermal discharge were studied. The results obtained are summarized as follows.

1.
Heat-retaining and diversion facilities could prevent the heated water from spreading upstream during flood tide, and effectively reduce the heat direct return. The diversion dike also promotes heat transport to outside the bay. The heat-retaining and diversion facilities strengthen the vertical mixing near the plant and weaken the thermal stratification. Therefore, they have certain impacts on the heat distribution and intake temperature rise.

2.
The heat-retaining and diversion facilities successfully reduced the intake temperature rise and complied with the intake requirement. Due to the different construction positions and functions, the diversion dike and the heat-retaining wall mainly reduced the intake temperature rise of prophase and Phase III, respectively. When the diversion dike and heat-retaining wall were both constructed, the greatest effects of superposition were obtained. The maximum of Phase III decreased significantly by 1.0-1.3 • C with an average reduction of 0.2 • C, and the maximum of Phase I and II decreased by 0.3 • C with little mean change.

3.
From an economic perspective, the crest elevations of 2 m and 1 m were considered and compared with 3 m to analyze the impact of height. The results show that when the crest elevation was reduced to 2 m, due to the short overtopping time of 4-6 h the intake temperature rise of Phase I and II had no obvious change and the average temperature rise of Phase III increased by 0.1 • C.
The constructions with 2 m crest elevations still played an effective role in reducing heat return.

4.
The construction of heat-retaining and diversion facilities could effectively reduce the area of low temperature rise. Compared with no facilities, the maximum area of 1 • C was reduced by 3.6-8.5 km 2 with a decrease of 39% and the average area of 1 • C was reduced by 0.5-1.3 km 2 with a decrease of 24%. However, the area of high temperature rise became larger, and the influence of the diversion dike near the outfall was more obvious. Because the heat-retaining wall was far away from the outlet, it had little effect. The change range of the maximum area of 4 • C was less than 20%, and the average area variation of 4 • C was unapparent. 5.
The experimental results are valuable for the study of thermal discharge and the construction of heat-retaining and diversion facilities in similar waters. However, this paper has focused on heat transport, and other factors involved in the engineering layout should be considered overall. Funding: This research received no external funding.