文献笔记 - Ground/Wall Effects on a Tilting Ducted Fan

2.2. Potential Flow Model

The flow above a fan is similar to a sink but modeling the fan as a sink [27] results in trends contradictory to the experimental data; i.e., in GE thrust decreases for a constant power. This is due to the fact that the fan performance in GE is mostly dependent on the flow pattern beneath the fan, which is more similar to a source rather than a sink. The fan has been represented by a source in [28] and the predicted trends agree well with experiments. The source model, however, is directionally uniform and cannot represent the directionality of the flow of a tilted fan. In an effort to represent the directionality of the flow field of a fan in forward flight, a directional source is proposed in [28]. This model, however, does not satisfy the Laplace equation and is not consistent with potential flow theory. In order to propose a potential flow model which can represent the directional nature of the fan behavior, herein, the combination of a sink (representing the suction of the fan) and a source (representing the flow below the fan), i.e., a potential dipole, is used. The dipole model can represent some of the fan flow characteristics which were not modeled accurately in the sink and source model. For example, the net flow rate of a sink/source is not zero while it is zero for a dipole (and a fan). Moreover, a dipole, similar to a fan, is a directional element. Therefore, one can expect that the dipole can predict the fan behavior more accurately compared to the sink/source model. In Figure 13, the potential functions (j) and the streamlines of a potential sink, source and dipole in GE are presented.

风扇上方的流动类似于水槽，但将风扇建模为水槽[27]会导致与实验数据相矛盾的趋势；即在地面效应推力减小的恒定功率。这是因为地面效应的风机性能主要取决于风机下方的流动模式，而风机下方更类似于源而不是汇。在[28]中，风扇由一个来源表示，预测的趋势与实验结果一致。然而，源模型在方向上是均匀的，不能表示倾斜风扇的流动方向。为了表示向前飞行中风扇流场的方向性，[28]中提出了一种方向性源。然而，该模型不满足拉普拉斯方程，也不符合势流理论。为了提出一种可以表示风扇行为方向性的势流模型，本文使用了汇（表示风扇的吸力）和源（表示风扇下方的流动）的组合，即势偶极子。偶极子模型可以表示一些在汇源模型中没有准确建模的风扇流动特性。例如，汇/源的净流量不为零，而偶极子（和风扇）的净流量为零。此外，偶极子类似于扇形，是一种定向元件。因此，可以预期，与汇/源模型相比，偶极子可以更准确地预测风扇行为。在图13中，给出了地面效应中势阱、源和偶极的势函数（j）和流线。

Since the dipole is a solution of the ideal flow equations, it is expected that the model fails to accurately predict the fan performance when the fan is located at very close distances to the walls and boundary layer fluid is recirculated through the fan. Therefore, to find the range of applicability and the limitations of the model, CFD simulations were performed for a wide range of distances from the ground/wall.

由于偶极子是理想流动方程的解，因此预计当风扇距离墙壁非常近并且边界层流体通过风扇再循环时，该模型无法准确预测风扇性能。因此，为了找到模型的适用范围和局限性，对离地面/墙壁的各种距离进行了CFD模拟。

The forces, similar to the CFD simulations, can be found from the integral form of the momentum equations over a control volume around the dipole. For the thrust, however, the control volume approach does not yield acceptable trends since the dipole, similar to other potential flow solutions, cannot represent the irreversible effects of the fan (the jump in the total pressure across the fan). Instead, another approach which yields acceptable values for the thrust has been used in the literature [27,28]. The ground effect alters the mass flux of the fan and thus changes the fan performance. Hence, the changes in the fan performance can be related to the induced velocity on the fan by the ground. This can be achieved by considering a constant power for the flights in and out of GE:

与CFD模拟类似，这些力可以从偶极子周围控制体积上的动量方程的积分形式中找到。然而，对于推力，控制体积方法不能产生可接受的趋势，因为偶极子与其他潜在的流动解决方案类似，不能代表风扇的不可逆效应（风扇总压力的跳跃）。相反，文献[27,28]中使用了另一种产生可接受推力值的方法。地面效应改变了风扇的质量流量，从而改变了风扇性能。因此，风扇性能的变化可能与地面对风扇的诱导速度有关。这可以通过考虑进出地面效应的飞行中功率恒定来实现：

where V is the axial velocity on the fan, T is the thrust, the subscript ∞ denotes out of GE and G denotes in GE. Thus, the thrust ratio is inversely proportional to the fan axial velocities:

其中，V为风机上的轴向速度，T为推力，下标∞表示在地面效应之外，G表示在地面效应之内。因此，推力比与风机轴向速度成反比：

The change in the fan axial velocity in GE is equal to the induced velocity on the fan by the image in the direction of the fan axis, Vi . Thus eqn (15) yields the following equation:

地面效应（GE）下风扇轴向速度的变化等于​​镜像涡​​沿风扇轴线方向对风扇产生的诱导速度 Vi​。因此，方程（15）可导出如下方程：

For a fan in GWE, the procedure is the same and the only difference is that three images must be considered instead of one.

对于处于GWE（地面壁效应）中的风扇，其分析步骤相同，唯一区别在于需考虑三个镜像涡而非一个。

The value of Vi is dependent on the strength of the dipole, which must be defined based on the fan characteristics. In this work, by assuming at height h0 the thrust ratio Th0 is given, the dipole strength is calculated. From eqn (16):

Vi的值取决于偶极子的强度，偶极子的强度必须根据涵道特性来定义。在这项工作中，通过假设在高度h0处给出推力比Th0，计算偶极子强度。根据方程式（16）：

At this height the induced velocity by the image on the dipole is:

在此高度下，​​镜像涡​​在偶极子上产生的诱导速度为

where h is the dipole strength. By combining eqns (17) and (18) the following equation for the dipole strength is obtained:

其中h是偶极强度。通过结合方程式（17）和（18），得到以下偶极强度方程式：

In this work, the strength is set based on the thrust ratio obtained for a fan at 1R above the ground. In all the thrust calculations only the ratio hV∞ appears and the explicit value of the dipole strength is not required. However, for the other component of the force calculated from the integral form of the momentum equations, the explicit value of the dipole strength is required. The same procedure can be used, the force at one point is calculated from the dipole model as a function of the dipole strength and based on the corresponding CFD simulations the explicit value of the dipole is calculated. Here, the fan tilted by 10˚ (0.1745 rad) at height of 1R above the ground was used to calibrate the dipole strength.

在这项工作中，强度是基于风扇在离地面1R处获得的推力比来设定的。在所有推力计算中，只出现hV∞比值，不需要偶极强度的显式值。然而，对于根据动量方程的积分形式计算的力的另一个分量，需要偶极强度的显式值。可以使用相同的程序，根据偶极子模型计算一点的力，作为偶极子强度的函数，并根据相应的CFD模拟计算偶极子的显式值。在这里，使用在离地面1R的高度倾斜10˚（0.1745 rad）的风扇来校准偶极子强度。

2.2.1. Evaluation Test

In order to evaluate the capability of the dipole model, the predicted thrust ratios for an isolated fan are compared to experimental data for two fans with different characteristics [14,15]. The experimental data and the dipole prediction for the two fans are shown in Figure 14, where the predicted values from the source model [28] are also presented. From the plot it can be observed that, except at very close distances to the ground, the dipole model predicts the trends for the different fans accurately. The source model, however, predicts one curve for different fans since the strength is defined only based on the average axial velocity on the fan.

为了评估偶极子模型的能力，将独立风扇的预测推力比与具有不同特性的两个风扇的实验数据进行了比较[14,15]。图14显示了两个风扇的实验数据和偶极子预测，其中还显示了源模型[28]的预测值。从图中可以看出，除了离地面非常近的距离外，偶极子模型准确地预测了不同风扇的趋势。然而，源模型预测了不同风扇的一条曲线，因为强度仅基于风扇上的平均轴向速度来定义。

4. CONCLUSION

The objective of this work was to characterize the effects of ground/wall on the performance of a ducted fan and develop a simple model that can provide the control system with real time predictions. For this, detailed CFD simulations were performed in Fluent for different configurations of the fan-ground-wall. The methodology was chosen based on the previous approaches in the literature for similar flow types and the results were validated based on the available experimental and theoretical information. While the CFD simulations are accurate, they require large computational resources and long time to converge. For example, using 4 nodes each containing two dual-core 2.4GHz and 4GB RAM, a typical simulation of the ducted fan in GWE require 1-2 weeks to converge. In mini-UAVs there is not enough space to install the required hardware; and even if it was, the simulations are not fast enough for online flight control of small UAVs (where less than a second delay can reduce the control mechanism performance significantly).

这项工作的目的是表征地面/墙壁对管道风机性能的影响，并开发一个简单的模型，为控制系统提供实时预测。为此，在Fluent中对风机地墙的不同配置进行了详细的CFD模拟。根据文献中类似流动类型的先前方法选择了该方法，并根据可用的实验和理论信息对结果进行了验证。虽然CFD模拟是准确的，但它们需要大量的计算资源和较长的时间来收敛。例如，使用4个节点，每个节点包含两个双核2.4GHz和4GB RAM，GWE中涵道风扇的典型模拟需要1-2周的时间才能收敛。在小型无人机中，没有足够的空间安装所需的硬件；即使是这样，模拟也不够快，无法对小型无人机进行在线飞行控制（在这种情况下，小于一秒的延迟会显著降低控制机制的性能）。

Several studies have shown that potential flow models can accurately simulate the fan flow in the proximity to walls. Therefore, in order to develop an online model, a dipole flow model was proposed. The proposed model is able to predict the fan performance changes at configurations that the fans can get in recent mini-UAVs designed for indoor flights. These configurations (e.g., fan tilting) were not possible to study with the previously proposed models with the same level of computational work.

几项研究表明，势流模型可以准确地模拟靠近墙壁的风扇流动。因此，为了开发在线模型，提出了偶极流模型。所提出的模型能够预测风扇在最近为室内飞行设计的迷你无人机中可以获得的配置下的风扇性能变化。这些配置（例如风扇倾斜）无法用之前提出的具有相同计算工作量的模型进行研究。

By comparing the dipole predictions with the CFD simulation results, it was shown that the dipole model can predict the fan performance accurately. From both methods the following trends were observed for a ducted fan hovering in a constant power condition:

通过将偶极子预测与CFD模拟结果进行比较，表明偶极子模型可以准确地预测风扇性能。从这两种方法中，观察到在恒定功率条件下悬停的管道风扇的以下趋势：

1. The GE are negligible at heights above 2R. In GE as the ground is approached the thrust increases. For typical ducted fans this change is seen to be more severe. For example at 1R above the ground, while the change in the thrust for an unducted fan is about 10% of the thrust out of GE; for a ducted fan it is about 20%.

1.地面效应在2R以上的高度可忽略不计。在地面效应，随着地面接近，推力增加。对于典型的管道风机，这种变化更为严重。例如，在地面以上1R处，而无导流风扇的推力变化约为地面效应推力的10%；对于管道风扇，这大约是20%。

2. The presence of the side wall does not affect the thrust significantly. However, it introduces asymmetry into the flow and thus the normal force is non-zero. The value of this force, for the height of 1R above the ground, is negligible at distances larger than 6R from the side wall. This limit will probably be larger at lower heights since the blockage effects due to the side wall will be stronger at lower heights.

2.侧壁的存在不会显著影响推力。然而，它在流动中引入了不对称性，因此法向力为非零。对于离地面1R的高度，该力的值在距离侧壁大于6R时可以忽略不计。在较低的高度，这个限制可能会更大，因为侧壁的堵塞效应在较低高度会更强。

3. In out of GWE region, fan tilting does not change the fan performance significantly. When the fan tilts in GWE, the thrust almost remains constant however the normal force changes. The trend of this change is almost linear even at the largest angle considered in this study.

在GWE以外的地区，风扇倾斜不会显著改变风扇性能。当风扇在GWE中倾斜时，推力几乎保持不变，但法向力会发生变化。即使在本研究中考虑的最大角度下，这种变化的趋势也几乎是线性的。

The dipole model has two parameters that must be defined based on the fan characteristics; the induced velocity out of GE and the strength. Herein, these parameters are defined based on the fan thrust at one height above the ground and normal force on a tilted fan at one height in GE. During the flight, the information at these two points must be estimated first (using sensor information and an iterative procedure for example) and the dipole parameters must be calibrated accordingly. Once these parameters are set, the dipole model can provide online predictions for other configurations. It may be more desirable to define the dipole parameters based on the fan characteristics and without any further information about the fan performance in GWE. In this paper it was not investigated if there is a way to define the dipole parameters independently. The focus was to show a simple potential flow model with coefficients defined empirically is in good consistency with the simulations which solve higher complexity physics.

偶极子模型有两个参数，必须根据风扇特性定义；地面效应的诱导速度和强度。在此，这些参数基于地面某一高度处的风扇推力及地面效应中倾斜风扇在某一高度处的法向力进行定义。在飞行过程中，必须首先估计这两个点的信息（例如使用传感器信息和迭代程序），并且必须相应地校准偶极参数。一旦设置了这些参数，偶极子模型就可以为其他配置提供在线预测。可能更希望基于风扇特性来定义偶极子参数，而不需要关于GWE中风扇性能的任何进一步信息。本文没有研究是否有独立定义偶极子参数的方法。重点是表明一个简单的势流模型，其系数由经验定义，与解决更高复杂性物理的模拟具有良好的一致性。

A real time, predictive model for GWE can improve the control system effectiveness in different ways. For example, the flight path can be easily predicted while it is extremely difficult with iterative methods. Predicting the path is important in view of optimizing the energy and increasing the safety of the flight. Moreover, compared to iterative mechanisms, less time is required to compensate for GWE and thus faster maneuvers are possible. For very fast maneuvers, however, the results of this work may not be applicable since only the hover condition is considered and dynamic GWE are not studied.

GWE的实时预测模型可以通过不同的方式提高控制系统的有效性。例如，飞行路径可以很容易地预测，而迭代方法则非常困难。从优化能量和提高飞行安全性的角度来看，预测路径非常重要。此外，与迭代机制相比，补偿GWE所需的时间更少，因此可以实现更快的机动。然而，对于非常快速的机动，这项工作的结果可能不适用，因为只考虑了悬停条件，而没有研究动态GWE。

While the predictions are greatly improved over the previous real time models, similar to any potential flow, the dipole model cannot capture the irreversible effects and fails at regions where these effects are not negligible. These limits were determined based on the comparisons with the CFD simulations. In real flights achieving these limits may not be possible due to other limitations such as fuselage shape and sensors, and thus the dipole limitations may not be practically important.

虽然与之前的实时模型相比，预测得到了极大的改进，但与任何势流类似，偶极模型无法捕捉到不可逆效应，并且在这些效应不可忽视的区域失败。这些限制是根据与CFD模拟的比较确定的。在实际飞行中，由于机身形状和传感器等其他限制，实现这些限制可能是不可能的，因此偶极子限制可能并不重要。

Another limitation of the dipole model is in predicting the normal force values in GWE. Since only a single point dipole is used to represent the fan, it is observed that the dipole must be located at closer distances to the wall to predict the same force values as the fan. However, the trends are accurately predicted. Hence, as the fan approaches the wall, the normal force at two points can be estimated to calibrate the dipole model.

偶极子模型的另一个局限性在于预测GWE中的法向力值。由于仅使用单点偶极子来表示风扇，因此观察到偶极子必须位于离墙壁更近的位置，才能预测与风扇相同的力值。然而，这些趋势是准确预测的。因此，当风扇接近墙壁时，可以估计两点处的法向力来校准偶极子模型。