In common jet aircraft, engine bleed air is the primary air source for the ECS, as shown in
Figure 1. Bleed air is compressed air taken from the 5th stage (intermediate pressure stage) or the 9th stage (high-pressure stage) of an engine’s compressor stage. The fan air is usually used to cool the compressor stage’s bleed air. The characteristics of a specific model engine’s rapid rise in the bleed air temperature are shown in
Figure 2 [
20]. In the gas turbines, the HP bleed air temperature is usually within 100~650 °C (212~120 °F), and IP bleed air temperature is usually within 100~50 °C (150~932 °F). Furthermore, based on the maximum heating rate and the most severe heating curve facing the aircraft’s control system, the HP and IP bleed air’s maximum inlet temperature rise rate is about 100 °C/s. HP bleed air pressure is typically within 150~3500 kPa (21.75~507.6 psi), and IP bleed air pressure is usually within 150~2500 kPa (21.75~362.6 psi). The maximum inlet pressure’s rise rate of the HP and IP bleed air is about 1000 kPa/s. The temperature of the fan air is typically 30~150 °C (85~300 °F), and the pressure is typically 100~700 kPa (14.5~102 psi).
Therefore, to truly simulate the working state of the engine, the primary problem is that the HP and IP bleed air simulation equipment must have a sufficiently high temperature and pressure. Moreover, it must meet the requirements of rapid response and good steady-state performance of temperature and pressure control. However, the heating equipment has the characteristics of a wide temperature range, nonlinear, and large inertia. Moreover, temperature and pressure control in the facility are strongly coupled. These characteristics usually lead to temperature and pressure dynamic control failure. By investigating the fast-heating method, we adopted the hot and cold blending method to realize the rapid response of the temperature. Then, this paper used a particular system structure to reduce pressure and temperature control coupling. In addition, this paper adopted the lookup-table-based PID (LPID) controller to acquire the rapid response and good steady-state performance of temperature and pressure control.
2.1. Fast Air Heating and Boost
To meet the test requirements of most ECS and its subsystem components, we built a facility capable of quickly adjusting the temperature and pressure of three air sources. We expected the facility to control bleed air at three states: the LP, the IP, and the HP. The three stage bleed air rigs’ system flow and control principle are all the same, except the pressure stage of the air source, equipment, and pipe diameter used. A 3.8 MPa air source supplies the HP bleed air simulation rig (HPR). A 2.5 MPa air source provides the IP bleed air simulation rig (IPR), and a 0.8 MPa air source provides the fan air simulation rig (FAR). As shown in
Figure 3, compressed air is provided by a compressed air station composed of air compressors, air storage tanks, air treatment and purification systems, and refrigeration dryers. The inlet pressure of the ECS test piece is adjusted by
and
. The valves could be fully opened or fully closed within one second. The valve
is installed on the drain branch with a smaller pipe diameter, which can fine-tune the pressure and improve the accuracy of pressure steady-state control.
At present, the standard methods of heating air are used: direct heating of an electric heater and indirect heat transfer of high-temperature gas. As the temperature control technology of gas combustion is immature, and the temperature control precision is low, this paper uses electric heating to obtain the high-temperature heat source. However, as the electric heaters’ dynamic performance is usually not the focus of designers and users, there is no credible data to support the dynamic performance design of the facility.
To achieve the rapid temperature adjustment of the facility, we draw on the system configuration method of aircraft cockpit temperature control. The hot and cold blending method can quickly respond to the aircraft cockpit’s temperature. The specific method is as follows: the facility adds a bypass in the front of the heater branch by adjusting the hot branch’s control valve and the bypass’s control valve. One part of the supply air flows through the heater, and another flows through the bypass. The two parts were mixed downstream of the heater. By changing the ratio of the hot and cold air mixture, we can achieve the purpose of rapid temperature control. However, the use of the hot and cold blending means that during a specific time of the test, a dry fir of electric heat may occur when the flow of the hot branch is too small. This paper used the counter-flow plate-fin heat exchanger for indirect heating. The heat exchanger is more conducive to realizing the actual engineering, and there is no danger of dry fir. Moreover, the heater is constantly working under atmospheric conditions, which helps reduce the R&D and production costs of the heater.
Figure 4 shows the test facility’s structure to achieve rapid temperature and pressure adjustment. In order to achieve a rapid response to temperature, the fast pneumatic valves
and
were used to control the temperature. One of the two valves was opened, and the other was closed. The two valves could be fully opened or fully closed within one second.
In addition, to achieve the dynamic requirements, the key actors are the thermal inertia and volume inertia of the pipeline downstream of the heat exchanger. The inertia causes a significant delay in temperature change and makes the temperature control system produce some overshoot and fluctuation.
The temperature transmission delay from the heat exchanger’s outlet to the test piece’s inlet is the same as the airflow flowing through the pipe, which is determined using the pipe length and the velocity airflow .
Due to the difference between the airflow’s temperature and the pipe wall’s temperature, heat loss occurs during the airflow transmission. The heat loss can be expressed according to the Sukhov Formula [
24]:
where
is the conductivity coefficient of the pipe,
.
is the external environment temperature,
.
is the specific heat ratio of air,
is the mass flow rate of air,
.
is the inlet temperature of the test piece,
.
is the supply air temperature at
. The subscript
represents the moment when the system is in steady-state thermal equilibrium.
where
is the convective heat transfer coefficient;
is the inside diameter of the pipe;
is the temperature of the pipe’s inner wall;
is the average temperature of the duct airflow. According to the above formula, to solve the problem of the thermal inertia of the pipeline connecting the heat exchanger’s outlet and the test piece’s inlet, the following measures were taken in the system.
Firstly, to ensure the facility’s safety, the distance between the outlet of the heater exchanger and the interface of the test piece was shortened as much as possible to reduce the inertia and retardation of the system.
Secondly, the should be increased to satisfy the test requirements. Specifically, during the test, ensure that the fan flow rate is sufficient, and control the heater outlet temperature ≥ , is the maximum temperature required for the ECS test. In this system, both the heat exchanger and the high-temperature pipeline are adequately insulated to reduce heat loss. The difference between the hot side inlet temperature and the cold side outlet temperature of the heat exchanger is generally less than 50 , and the temperature difference between and is generally less than 20 .
Thirdly, when considering the maximum flow demand of the system, the inside diameter of the test piece shall be reduced. This can reduce the heat loss of the pipeline and increase the airflow velocity to reduce the thermal delay.
Fourthly, before starting the test, the inlet pipe of the test piece shall be pre-heated to increase and speed up the dynamic response of temperature. To reduce the preheating time, the wall thickness of the inlet pipe should be as small as possible while ensuring safety.
The above points can effectively ensure that based on reducing the power consumption, the test piece’s inlet temperature rapid response meets the experimental requirements during the ECS dynamic test.
2.2. Decoupling
As shown in
Figure 4, the hot branch has an extra resistance device (the heat ex-changer) compared to the cold branch. The temperature control valve opening changes in the dynamic test will change the temperature simulation unit’s flow resistance and affect the pressure control. Similarly, the heat exchanger resistance will cause serious nonlinear flow distribution in the hot and cold branches, which may lead to the failure of temperature control.
For these reasons, this paper used a particular system structure to reduce pressure and temperature control coupling.
Firstly, valves and shall be set on the hot and cold branches, respectively. When valves , , and are in the fully open position, the opening of valve shall be adjusted to make the resistance of the cold and hot branches similar.
Secondly, considering that and pneumatic valves have the same size, once one of them is in the open position, the other one shall be closed. In this way, the temperature control valves’ action will not affect the flow resistance of the system.
Thirdly, the principle of selecting the pressure control valve is to make the outlet flow and the opening of the valve have a linear relationship as much as possible to reduce the influence of pressure changes on temperature control. In engineering, due to the existence of equipment and pipeline resistance, the valve’s flow rate will be distorted. In this system, the resistance of the series pipeline behind the valve is much greater than the resistance of the regulating valve, so chooses an equal percentage valve.
2.3. The LPID Controller
To take into account the rapid response and good steady-state performance of the pressure and temperature control, this paper used the lookup-table-based PID (LPID) controller to control the temperature and pressure of the bleed air simulation test [
25].
2.3.1. The LPID Control
The LPID controller takes the measurement value as input and outputs a 4~20 mA.DC signal after comparing the measurement value with the set value every 5 ms. After the two pressure control valves receive the same control signal, one will open up, and the other will close down. When receiving a 4 mA signal, the pressure control valve on the main pipe is fully closed, and the control valve on the linked branch is fully open. The temperature controller takes the temperature measurement value as input. After comparing the input value with the set temperature value, it outputs a 4~20 mA.DC signal and controls the temperature control valves on two temperature control branches simultaneously. Like the pressure control valves, the temperature control valve on the high-temperature branch is fully opened after receiving a 20 mA signal. The other control valve on the room temperature branch is fully closed.
The LPID control is developed from traditional PID. The distinguishing feature of the proposed controller is pre-calculated to make a PID parameter scheduling table. We established the relationship between three PID parameters , , , and the error based on this specific system. According to different temperature and pressure errors, the PID gains can be self-adjusted online to make the controlled object have a good dynamic and static performance, meeting different control requirements. The specific gain scheduling rules are as follows:
When the , we use the strong PID to minimize error quickly;
When the , we adopt the weak PID;
When the , which indicates the absolute error value, tends to be very small, we can use the PI to decrease the static error;
When the deviation is slight to a certain extent, , the concept of the dead zone can even be introduced, at this moment, the output of the controller remains invariable, namely, .
Where, , , and are limit values. The limit values and the different PID gain parameters need to be debugged in actual engineering.
2.3.2. Simulation Example
The sampling time is 5 ms, the ideal position signal is . In the simulation program, due to the discretization, there are two delays in the control input.
The unit step response simulation results of a traditional PID controller and LPID controller are shown in
Figure 5. The simulation results showed that the LPID controller has advantages, high control precision, a low overshoot amount, a fast response, and high stability compared to the traditional PID controller.