laitimes

The technical team of Xu Ji Electric Company proposed a UHVDC full digital simulation modeling method

Shen Shuaihua, Li Tai, Xiao Long, Jing Xueji, and Du Shaolin of Xu Ji Electric Co., Ltd. wrote an article in Electrical Technology, introducing a RT-LAB-BASED UHV DC full digital simulation modeling method, which realizes the accurate simulation of the full digital model by creating custom components in Simulink and calling the actual engineering program code, and realizes real-time operation through the method of split-core operation and multi-step time-sharing data transmission. The comparison with the actual simulation test results verifies the accuracy of the all-digital model, which shows that the model can be run offline (non-real-time) in Simulink or in RT-LAB in real time, and has a wide range of application scenarios.

Please click here to enter a description of the image (maximum 18 words)

With the vigorous development of new energy power generation in mainland China, more and more new energy stations are connected to the AC power grid at the sending end of the high-voltage/UHVDC transmission system, and in order to study its stability, it is necessary to carry out electromagnetic transient modeling of new energy stations and DC transmission systems, among which the accuracy of the DC transmission system model is particularly important. In this paper, we introduce an ultra-high voltage DC simulation modeling method based on RT-LAB simulator, which is developed based on actual engineering programs and has high accuracy.

In DC transmission simulation modeling, the commonly used electromagnetic transient simulation software mainly includes real-time digital simulator (RTDS), advanced digital power system simulator (ADPSS), PSCAD (power system computer aided design), HYPERSIM, RT-LAB et al. Based on the above-mentioned simulation software, the actual control and protection device is configured, and the accurate simulation of the DC transmission system can be realized, but limited by a variety of unfavorable factors such as the complexity of the actual control and protection device system, the long construction period, and the large floor area, this mode is generally used for the test test of the control and protection manufacturer.

The internal control system is designed in the above-mentioned simulation software, but the logic is relatively simple, which is very different from the logic of the actual engineering program, and cannot accurately reflect the actual engineering characteristics. The literature introduces the modeling method of UHVDC actual control protection based on PSCAD, which is interconnected with the operator's operating system, but the model cannot be run in real time. Therefore, in order to fully realize all the functions of the control and protection system, accurately reflect the characteristics of the actual project, and effectively reduce the simulation cost, it is extremely important to develop an all-digital simulation model of electromagnetic transient based on the actual engineering control and protection program.

RT-LAB is a real-time simulator based on Matlab/Simulink developed by Opal-RT in Canada. Compared with simulation software such as RTDS, users can directly convert Simulink models into recognizable models in RT-LAB, and the models can be run offline in Simulink or in real time on RT-LAB lower computers, which has the advantages of convenient use and wide range of application scenarios.

In this paper, taking the Ximeng-Taizhou UHVDC transmission project as an example, a circuit model is built in Simulink, and the converted engineering program is called by using the custom module S-function in Simulink to realize the establishment of a full digital simulation model based on the actual engineering control protection program. The accuracy of the simulation model is verified by comparing it with the real waveform simulated by the actual object.

1 Introduction to Simulation Modeling

1.1 Simulation modeling scheme

The simulation modeling scheme introduced in this paper is based on the HCM3000 DC control and protection platform independently developed by Xu Ji, and transforms, cuts and converts the ViGET program supporting the control and protection platform to generate callable C code. The converted engineering program C code is encapsulated into a library file in Simulink and RT-LAB lower computers, and the interface function is created and the customized module S-function in Simulink is used to call the encapsulated library file, and finally the simulation modeling of the actual engineering control protection program in RT-LAB is realized. The RT-LAB modeling process is shown in Figure 1.

1.2 Program interface transformation

控制保护模块之间涉及的通信方式主要有时分复用(time division multiplexing, TDM)通信、高速控制总线集成闪存控制器(integrated flash controller, IFC)通信、局域网(local area network, LAN)通信、现场总线等。 在搭建全数字仿真模型之前需要对这些通信的接口模块进行修改。

1) TDM communication

The configuration information of the TDM communication function block in the engineering program does not need to be modified, but when the two energy processing unit (EPU) boards in the entire chassis generate pulses, the pins of the TDM communication function block need to be modified, and the task scheduling of the EPU becomes timer mode.

2) IFC Communication

IFC communication is mainly used for chassis-to-chassis communication, and the settings of the function blocks using this communication protocol do not need to be changed, but only the ACD/PTR pins of these function blocks need to be manually set to different values to represent different data destinations or sources.

3) LAN communication

Most of the signals of LAN communication are slow signals that interact between the control system and the background, between the control and protection systems, and between stations. In the all-digital simulation model, the signals uploaded to the background by the control and protection system are not considered, and the signals such as the operation mode, various reference value commands and switch commands are directly read into the control and protection system through the input pins.

4) Fieldbus communication

In the primary system model, there is no need for the opening and closing command, only the opening and closing state of the switch. The DCSIM device is a switch simulation device independently developed by Xu Ji Company, in the actual project, the DCSIM device receives the opening and closing command of the control protection, and sends the opening and closing status of the switch to RT-LAB. In order to reduce the modification of the program, the simulation model still uses the DCSIM device to simulate the switch. In the actual valve control system, the opening and closing commands of the bypass switch are sent to the DCSIM device by both fieldbus and hardwire. The protection system controls the switches only by means of direct inputs/outputs of the switching quantities.

2. Simulation model establishment

2.1 Simulation model structure

The Ximeng-Taizhou UHVDC transmission project starts from the Ximeng converter station in Inner Mongolia and ends at the converter station in Taizhou, Jiangsu Province, with a DC line length of 1 618km, a bipolar DC line and two 12 pulsating converters in series at each pole, with a rated voltage of ± 800kV, a DC transmission capacity of 10 000MW and a DC rated current of 6 250A. Taking this practical project as an example, a control and protection simulation model as shown in Figure 2 is built based on the RT-LAB platform.

Figure 2 shows the structure of the full digital simulation model of Xitai DC control and protection, which is divided into a primary main circuit model and a secondary control and protection system.

The primary main circuit model was built using components from the Simulink library and components from the RT-LAB library developed by Opal-RT.

The secondary control and protection system includes measurement system, AC and DC station control, extreme control, polar protection, valve control, valve protection, and the control and protection system of the rectifier station and the inverter station is completely symmetrical.

The primary main circuit model needs to be sub-kernelized, which is realized by the ARTEMIS solver in RT-LAB [19], and the exchange data between the primary main circuit model and the secondary control and protection system is realized in RT-LAB through the OpComm module.

2.2 Custom Function Blocks

By writing an interface program, you can use the custom module S-function to invoke the converted engineering program in Simulink. The custom function block is divided into the main function module and the general module, wherein the main function module calls the engineering program, and in a simulation model, the main function module of a control protection program can only appear once, and the general module is developed to realize the cross-core data communication between the main function modules, and can appear many times in a simulation model.

Figure 3 is the custom function block of the extreme control program of Ximeng station, which is a main function module, and all the functions of the extreme control can be realized by calling the library file compiled and generated by the extreme control program. The left side of the module is the input signal, and the two on the right are the output signal and the debug signal, respectively, where the debugging signal is used to monitor the values of some variables inside the function block, which is only used during debugging.

Fig. 4 shows two general-purpose function blocks, wherein XT_Adr2Arr is used to convert an address into an array (double type) after controlling the protection module, and then the data in the array is transmitted across cores; the XT_Arr2Adr is used before controlling the protection module, after receiving the data transmitted across cores, the module is used to convert the array into an address, and then is used as the input of the control protection module.

The two function blocks shown in Figure 4 are used because the input and output of the main function function blocks are the first address of the pointer, and the data is stored in the open pointer space, while the RT-LAB real-time runtime must be kernel-split, and the data communicated between the cores must be double. For this reason, data transformation is required when data is exchanged between different main functions.

As shown in Figure 5, the output signal of the main function of the pole control is first converted into a double-type array by the first address of the pointer in core 1, and then the data is transmitted across the core, and then converted into the first address of the pointer by the double-type array in the core 2 where the valve control is located, which is used as the input signal of the main function of the valve control.

2.3 Real-time operation scheme

The typical simulation step size of the electromagnetic transient simulation model is 50s, and when the model processes incomplete programs in one simulation step, the timeout problem will occur. In order to run in real time, the model needed to be split and simulated using 2 OP5600 chassis and 2 OP5607 chassis. Figure 6 shows the co-simulation model splitting scheme.

The OP5600 chassis has a total of 32 cores and a frequency frequency of 2.3 GHz, and the OP5607 chassis is mainly used to transmit data in the co-simulation of two OP5600 chassis. For co-simulation, two optical fibers are required between the two OP5607 chassis, one for synchronization and one for data exchange, and the OP5600 chassis and the OP5607 chassis need to be connected with a PCIe (peripheral component interconnect express) cable.

2.4 Time-sharing data transfers

Since RT-LAB operates on a multi-core processor parallel processing mechanism, data transfer must be completed in a simulation step. When the data length of the transfer is large, it may not be possible to complete the full data transfer in one simulation step. In the operation mechanism of RT-LAB, in order to complete the simulation, the actual simulation step size will become longer, resulting in non-real-time operation, that is, the step overflow problem.

Considering that the communication time of the control protection program is much greater than 50s in the actual engineering, in order to improve the simulation speed of the all-digital model and reduce the communication load of a single step size, a multi-step time-sharing data transmission method is proposed. Firstly, the data length (size) is determined according to different communication modes, then the number of simulation steps required to complete data transmission (nSeg) is determined according to the actual execution cycle of the main function and the simulation step size of the model, and finally the number of data that needs to be transmitted by a single simulation step is determined according to the data length and the number of simulation steps, that is, the array dimension transmitted by the OpComm module.

When nSeg is 1, the multi-step time-sharing method is not used, and the signal transmission process is shown in Figure 7.

When nSeg is greater than 1, the multi-step time-sharing method is used, and the signal transmission process is shown in Figure 8.

Both RT-LAB and HYPERSIM were developed by Opal-RT in Canada, but their operating mechanisms are different. In HYPERSIM, you can use shared memory to distribute data, which reduces the resources occupied by data transmission and eliminates the need for operations such as model splitting and time-sharing data transfer, while in RT-LAB, operations such as model splitting and time-sharing data transfer are required for real-time operation. Therefore, when building an all-digital simulation model, the RT-LAB model needs to occupy more hardware resources.

3. Simulation test verification

After the modeling of the all-digital simulation model is completed, in order to verify the integrity of its function, the steady-state parameter verification test, unlocking test, power rise and fall test, step test, reactive power control test, AC system fault test, DC line fault test, etc. are carried out with reference to the factory test report of the simulation test in the actual project.

At the same time, in order to verify the accuracy of the simulation results, the test waveform of the real platform based on the actual control and protection device based on HYPERSIM was compared, and the tests of unlocking, step response, valve group throwback, and single-phase grounding of the AC system were carried out respectively. In the test waveform, from top to bottom, DC voltage (kV), DC current (A), and trigger angle (°) are described in detail.

1) Unlock the trial

Power Forward Electrode II. Low-end converter operation, unlock test. Pole II. The low-end inverter unlocking test waveform is shown in Figure 9.

In Figure 9, the dark line is the waveform of the object imitation real test, and the light line is the waveform of the all-digital simulation model built in this paper. In the physical model test, the maximum DC current is 1 515 A, and in the full digital simulation test, the maximum DC current is 1 495 A. The DC voltage, DC current and trigger angle adjustment in the full digital simulation test waveform are basically the same as those of the physical model waveform.

2) Step response test

Power forward delivery pole I. low-end converter operation, current step test. The waveform of the current step test is shown in Figure 10.

In Figure 10, the dark line is the waveform of the object imitation real test, and the light line is the waveform of the all-digital simulation model built in this paper. The test condition is the forward power feed, the pole I current step, the step amount is 500A, and the step time is 500ms.

In the physical model test, the overshoot of the upper step response was 4.32% and the response time was 180 ms, and the recovery response was 14.13% and the response time was 127.5 ms. In the all-digital simulation experiment, the overshoot of the upper step response is 4.82% and the response time is 190 ms, and the recovery response is 17.5% and the response time is 136 ms. The DC voltage, DC current and trigger angle adjustment in the full digital simulation test waveform are basically the same as those of the physical model waveform.

3) Valve group exit test

The power forward sending pole I. double inverter operation, the low-end valve group out of the test. The waveform of the exit test of the pole I. low-end valve group is shown in Figure 11.

In Figure 11, the dark line is the waveform of the object imitation real test, and the light line is the waveform of the all-digital simulation model built in this paper. In the physical model test, the maximum DC current of Ximeng station is 1 345A, and the minimum DC voltage is 348kV. In the all-digital simulation test, the maximum DC current of Ximeng station is 1 219A, and the minimum DC voltage is 366kV. The DC voltage, DC current and trigger angle adjustment in the full digital simulation test waveform are basically the same as those of the physical model waveform.

4) Single-phase grounding test of AC system

Power reverse pole II. high-end converter operation, Ximeng converter station pole II. AC system single-phase grounding test. The single-phase grounding test waveform of the AC system is shown in Figure 12.

In Figure 12, the dark line is the waveform of the object imitation real test, and the light line is the waveform of the all-digital simulation model built in this paper. According to the waveform analysis, when the power is reversed, and a single-phase grounding fault occurs in the AC system of Ximeng station, the DC voltage, DC current and trigger angle adjustment in the all-digital simulation test waveform are in high consistency with the physical model waveform.

In summary, under the above test conditions, the all-digital simulation model and the simulation test results of the physical model system maintain a high consistency, which proves the accuracy of the all-digital simulation model.

4 Conclusion

This paper proposes a method to build an all-digital simulation model of UHVDC engineering in the RT-LAB simulator, which realizes the development of the all-digital simulation model that is highly consistent with the actual project while protecting the intellectual property rights of the control and protection manufacturers, and the simulation accuracy is greatly improved.

This paper solves the problem of real-time operation of the all-digital simulation model of UHVDC engineering in RT-LAB simulator. Models can be run offline (not in real time) in Simulink or in real time on RT-LAB, making them available for a wide range of applications.

The results of this work were published in Electrical Technology, and the title of the paper is "Full Digital Modeling and Simulation of Xitai Engineering Control and Protection System Based on RT-LAB", and the authors are Shen Shuaihua, Li Tai, Xiao Long, Jing Xueji, and Du Shaolin

Read on