Research on vibration and interface simulation technology of aero-engine

2.1. Power module

The different modules of the interface simulator need different power supplies, so different power supplies are designed according to different circuits. The power supply voltage is 24 V, and a WRB2412S-3WR2 chip is used to convert 24 V into –12 V, while using a WRB2412MD-10W chip to convert 24 V into 12 V, they supply power for the analog input and output channels. LM2576-5V chip is used to convert 24 V voltage into 5 V, which can supply power for the DAC chip. An AMS1117-3.3 chip is used to convert 5 V into 3.3 V, supplying power for digital input and output channels and the TM4C1294NCPDT chip.

2.2. CPU module

The Interface simulator uses a TM4C1294NCPDT as CPU, which is made by Texas Instruments. This ARM chip has powerful functions, and it has 256 KB flash and 1 MB SRAM, eight 32-bit counter, four PWM generator and two 12- bit ADC, etc. In addition, it also integrates MAC/PHY Ethernet, which supports 10 MB or 100 MB communication rate, which is better than the chip with only serial communication module and has better real-time performance. It has better real-time performance than the serial communication.

The peripheral circuit uses a 25 MHz crystal to provide the external clock for the TM4C1294NCPDT chip, and the internal system clock can reach 120 MHz by frequency doubling. It can meet requirements of different Peripherals by dividing or doubling the frequency of the system clock.

2.3. Input module

According to the frequency of vibration signals, different kinds of sensors are used for measurement, and the acceleration sensor is usually used to measure the vibration of aero-engines. Some acceleration sensors cannot obtain the voltage signal directly, but they can obtain the charge signal, such as the piezoelectric accelerometer sensor. Therefore, it is essential to preprocess before the AD sampling. Generally, the perception of vibration signal includes sensor measurement, converter pretreatment, and A/D sampling, the principle of which is shown in Fig. 2.

The input module includes two parts: analog output channel and digital input channel. The analog input channel is designed based on the common voltage signal, current signal and the voltage signal after pretreatment. Its input has ranges of 0-20 mA, 0-5 V and 0-10 V, and the signal will be converted to 0-2.5 V standard voltage signals after conditioning. Positive and negative power supplies are used in amplifier circuits to ensure that the output is 0 when the input is 0. In order to meet the requirements of a wide range of signal input, this paper uses a combination of resistor1 ($R_1$) and resistor2 ($R_2$) to make it, as shown in Fig. 3.

When $R_1$ is zero, current signals can be obtained: $V_{out}=R_2\times I_{in}$.

When $R_1$ is not zero, voltage signals can be obtained: $V_{out}=\frac{R_2}{R_2+R_1}\times V_{in}$.

Where, $V_{out}$ is the obtained voltage, $I_{in}$ is the input current and $V_{in}$ is the input voltage.

Considering the engine vibration monitoring not only deal with the vibration signal, but also deal with operating parameters and flight parameters related to engine, so this module is designed to have 16 input channels, which can meet the requirements of engine monitoring. Therefore, the voltage obtained by each channel can be written as follow:

where, $A{I}_{AR{M}_x}$ is the obtained analog signal, $A{I}_x$ is analog input signal, subscript $x$ is the channel number, it can be 1 to16.

In addition to vibration signals, operating and flight parameters. There are some signs indicating the working state of the switching signal, such as “super vibration warning lights” and so on. There are some switching signals indicating state of the engine, such as “Excessive vibration warning lights” and so on.

The high or low voltage level represents two different states. The interface simulator collects these signals, converts them into digital “0” or “1”, and sends to the computer with model. In order to be compatible with different electrical switches, the input voltage range should be as large as possible. In this paper, the maximum voltage of digital input channel is design to 24 V, while the ARM chip pin input is only 0-3.3 V, so the optical isolation technology is used to convert the 24 V drive voltage into 3.3 V. Besides, switching between high or low voltage can be achieved with a jumper.

Fig. 2The procedure of Vibration signals perception

Fig. 3Schematic diagram of the analog input channel

2.4. Communication module

TM4C1294NCPDT chip integrated MAC/PHY Ethernet, support 10 MB and 100 MB communication rate. It has better real-time performance compared with the common serial communication. A RJ45 network port with transformer is connected with the related pins of the chip to realize the Ethernet communication. As shown in Fig. 4.

The interface simulator uses LwIP protocol stack to realize Ethernet communication. It is an independent and simple implementation of the TCP/IP protocol. This protocol stack has a complete TCP/IP function, but the consumption of the microprocessor resources is very low.

2.5. Output module

The output module is composed of the analog signal output channel, the digital signal output channel and the frequency signal output channel. The vibration information is calculated by computer and be sent back to the chip. There is no integrated DAC chip in TM4C1294NCPDT chip, so the ARM chip needs two DAC7568 chips, which can be connected with the 16 analog output channels. In order to ensure the simulation quality of the vibration signal, SSI (Synchronous serial interface) is used to provide the real time communication. DAC output voltage is in a range of 0-2.5 V, this paper uses a dedicated V/I converter AD694 chip to make voltage signal into 0-20 mA current signal, which can improve the stability of the signal. Amplifier is used to double the voltage to 0-5 V voltage output. Then a resistor divider can be used to provide the 0-50 mV voltage output. The practical implementation is shown in Fig. 5.

Fig. 4Schematic diagram of the Ethernet communication

Fig. 5Schematic diagram of the analog output channel

The digital output channel is mainly used to output some high or low voltage level, which indicates the state of the engine. Some working states can be obtained by the model, and the electronic controller will take action according to these states. The digital signal output channel is divided into two parts: one part is outputting directly by ARM through a 74LS244 chip, the other part is the contact output module. When the ARM output is high, the transistor is switched on and the relay is closed at this time. When the ARM output is low, the transistor cut off, and the relay is disconnected.

In addition, the output module also contains two Frequency outputs. These two outputs are connected to the PWM pin of the ARM chip through a buffer. By programming the duty and period, a simulation of speed signal can be realized. They can also simulate a torsional vibration signal with a PWM comparator.

The product of the interface simulator designed in this paper is shown in Fig. 6(a), and the whole condition monitoring of engine vibration and hardware in the loop simulation platform is shown in Fig. 6(b).

Fig. 6Interface simulator and experimental platform

a)

b)

3.1. Modeling of vibration signals

A high pressure rotor Dynamic model of aero engine can be expressed as follows:

where, $n_h$ is rotational speed of high pressure rotor, $J_h$ is rotary inertia of high pressure rotor, $\mathrm{\Delta }{M}_h$ is residual moment.

Here, only considering the vibration of high pressure rotor of the engine, the model can be simplified into:

In the Eq. (3), $G$ is the overload of the engine vibration, $K_g$ is the vibration coefficient, $G_d$ is the DC component, $G_{ub}$ is the unbalance amount, $\theta$ is the phase of the vibration frequency band, $G_m$ is the amplitude of vibration interference; ${\theta }_m$ is interference phase.

Simplified mathematical model of vibration sensor is described as below:

where, $U_{sensor}$ is voltage output of the sensor; $K_{sensitivity}$ is sensitivity.

A charge signal of vibration can produce current signals by conditioning, and the charge signal is:

In the Eq. (5), ${\mathrm{C}}_{\mathrm{r}\mathrm{e}\mathrm{f}}$ is the reference capacitor, $Q_1$ is the charge signal by conditioning. In the Eq. (6), ${\mathrm{I}}_2$ is the generated current signal, $\mathrm{\Delta }{I}_2$, $\mathrm{\Delta }{U}_{sensor}$, $\mathrm{\Delta }{C}_{ref}$ and $\mathrm{\Delta }{Q}_1\left(t\right)$ are uncertain signals generated by external disturbance.

The charge signal of vibration collected by the controller is:

The voltage of the sampling capacitor of the electronic controller is:

where, $C_{sample}$ is the sampling capacitor, $\mathrm{\Delta }{U}_{sample}\left(t\right)$ and $\mathrm{\Delta }{C}_{sample}\left(t\right)$ are uncertain signals generated by the external disturbance.

With the interface simulator, the intermediate pretreatment can be omitted, and the vibration signals can be output to the electronic controller directly in the form of the voltage. A voltage value can be converted to an amplitude value by a certain scale:

where, $U_{out}$ is the voltage which the interface simulator output to the controller directly, $k_c$ is the conversion coefficient.

3.2. Simulation and analysis

The interface simulator is used to generate a vibration signal, which is transmitted to the engine electronic controller. Then the controller collects the signal and analyzes it. The accuracy and reliability of the interface simulator can be verified when comparing with the mathematical model. In order to facilitate the verification experiment, the expression of overload signal in the Eq. (3) is simplified and assigned:

where, $A_1$ and $A_m$ are amplitude after simplification, $\mathrm{\Delta }G\text{'}$ is the external disturbance, and can be taken as a random Gauss white noise signal:

According to the actual vibration of an aero engine, the parameters in the Eq. (10) can be assigned as follow: $n_h=$ 150 Hz, ${\theta }_1={\theta }_m=$ 0, $A_1=$ 4.5, $A_2=$ 2, $A_3=$ 1.2, $A_4=$ 0.6, $A_5=$ 0.3, $A_6=$ 0.1.

In addition, the interface simulator output is selected the range of 0 to 5 V, which means the maximum output amplitude is 5 V.

Fig. 7Chart of the simulation result

Fig. 8Signals obtained after blind source separation

The result of simulation has been shown in Fig. 7(a). These are the voltage values collected by the electronic controller in 5 seconds. The result shown in Fig. 7(b) can be achieved by dealing with the voltage values with fast Fourier Transform. The frequency of the vibration signal analyzed in frequency domain is consistent with the assumed values in the model, so the accuracy of the interface simulator gets proved. High order signal does not appear in Fig. 7, because their amplitude is too small and the sampling frequency is limited. From the perspective of external interference, the interference in this frequency band is very small, and the reliability of the interface simulator is very good.

For the second harmonic and the non-integral harmonic, FFT is difficult to determine the vibration caused by the high pressure rotor, or the vibration caused by the low pressure rotor. A blind source separation method is also needed, and the analog signal can be obtained as a result of ICA processing, which is shown in Fig. 8. Fundamental frequency and frequency doubling can be seen in the Fig. 8 as well as some other harmonics, which also proves the stability of the interface simulator.