Load forecast of a railway power supply network based on key performance indicators of substations

Abstract

A method for predicting network admittances (NA) and vehicle performance (P) is presented. F ) based on current and voltage values ​​(I, U) of substations (U1, U2) of a railway power supply network (10). The method involves a computer-based simulation of a plurality of labeled training railway power supply networks (TBNs) with different input parameters and a wide variation of rail traffic as input and output data, where current values ​​(I) and voltage values ​​(U) in the substations (U1, U2) are used as output data. A , U B) are output at each time point. Furthermore, an artificial neural network (ANN) is trained using the training railway power supply networks (TBN), whereby the output data of the training railway power supply networks (TBN) are used as input data for the artificial neural network (ANN) and the input data of the training railway power supply networks (TBN) are used as target data for the artificial neural network (ANN). Time-dependent current and voltage values ​​(I, U) of the substations (U1, U2) of the railway power supply network (10) are measured for a plurality of consecutive time points. Finally, the trained artificial neural network (ANN) is used to determine network admittances (NA) and vehicle power (P) based on the measured time-dependent current and voltage values ​​(U, I). F) in the railway power supply network (10). A method for operational optimization of a railway power supply network (10) is also described. Furthermore, a forecasting device (50) is described. An optimization device (60) is also described.

Description

[0001] The invention relates to a method for predicting network admittances and vehicle performance based on current and voltage values ​​from substations of a railway power supply network. The invention also relates to a method for optimizing a railway power supply network. Furthermore, the invention relates to a predictive device. In addition, the invention relates to an optimization device.

[0002] Railway power supply systems are used to supply rail vehicles with direct current or alternating current.

[0003] The operation of a railway power supply system can be improved by adjusting free parameters so that the entire system exhibits the desired behavior. Several known methods exist for determining these free parameters. What most of these methods have in common, and especially those known for determining a globally optimal solution, is that they require as input parameters both the power outputs of all trains and their positions, which in turn are needed to determine the admittances required for the calculation.

[0004] This fact presents a problem, as this information is only known to the trains and, where applicable, the train control and dispatching systems, and not to the infrastructure components of the railway power supply. Therefore, networking of systems is necessary. This networking can be complicated or undesirable for the following reasons: For cybersecurity and functional safety reasons, networking stationary train control or dispatching systems with railway power supply systems may be undesirable. Train control systems often meet very high safety requirements (SIL level) and are therefore installed in highly shielded networks. Connecting the entire railway power supply to these systems would therefore entail immense effort.

[0005] Equipping all individual rail vehicles with measurement and transmission technology and networking them with the railway power supply is very complicated because: First, all rail vehicles would need to be equipped with measuring and transmitting equipment that records information on current electrical power and position and transmits it via data transmission.

[0006] Determining the electrical power output of a rail vehicle is complex, as it would require the installation of new measuring devices in the vehicle (which is sometimes relevant for type approval) or a connection to the vehicle control system. Determining the vehicle's location also requires a connection to the existing vehicle control system or an additional GPS measuring device.

[0007] The GPS measuring device only works in areas where there are no tunnels. Therefore, this option is not available for metro vehicles in tunnels.

[0008] Information transmission must take place via a wireless transmission channel. Since the railway's own transmission channels are usually already heavily congested, the use of 5G is recommended. Here too, a sufficiently strong and reliable signal must be available in all areas where rail vehicles using this network can operate (tunnels may also pose a problem here).

[0009] The information from the rail vehicles must be transmitted simultaneously to all rail vehicles at short, regular, and, above all, time-synchronized intervals. This implementation is also complex and time-consuming.

[0010] Furthermore, it must be ensured that all vehicles operating in the network are equipped with appropriate transmission and measurement devices.

[0011] It is conceivable that such systems could already exist and be operational today, but integrating substation optimization into existing systems leads to high costs for implementing the aforementioned requirements regarding operational safety and data security. The difficulty in implementing these technical solutions is further compounded by the fact that modifications to rail vehicles are necessary to optimize the operation of the railway power supply.

[0012] Many authors have successfully attempted to optimize the energy consumption of railway systems in the literature. However, most of these optimizations are limited to load profile optimization through adjustments to train operating behavior. See Pablo Martinez Fernandez, Ignacio Villalba Sanchis, Victor Yepes, Ricardo Insa Franco, “A review of modelling and optimisation methods applied to railways energy consumption,” in Journal of Cleaner Production, Volume 222, 2019, Pages 153–162 for an overview, and S. Khayyam, N. Berr, L. Razik, M. Fleck, F. Ponci and A. Monti, “Railway System Energy Management Optimization Demonstrated at Offline and Online Case Studies,” in IEEE Transactions on Intelligent Transportation Systems, Vol. 19, no. 11, pp. 3570-3583, Nov. 2018, doi: 10.1109 / TITS.2018.2855748 can be used for a detailed application.

[0013] Alternativ werden die notwendigen Informationen von Schienenfahrzeugen während der Optimierung als verfügbar angenommen, siehe CHONGQING CRRC TIMES ELECTRIC TECHNOLOGY CO., LTD., (2022) „Urban rail intelligent traction power supply system for use in urban rail traffic system has energy operation control system used to receive bidirectional converter set power, device operation parameter, state, comprehensive vehicle position and traction / braking power“ (CN115549196-A) und ALSTOM TRANSPORT TECHNOLOGIES, (2020) „Method of dynamically adapting the operation of traction substation of an electric power supply system for vehicles, involves determining of the vehicles currently circulating in the section specifically associated with substation“ (EP3715172-A1, FR3094288-A1, AU2020202019-A1, IN202014012075-A, FR3094288-B1, EP3715172-B1, ES2886046-T3, HK40029364-A0, HK40029364-A1).

[0014] However, as previously explained, the availability of information from the rail vehicles is quite complicated in practice, making such an implementation very difficult. The difficulty in obtaining the input variables for optimization also makes optimizing the free parameters of the railway power supply itself difficult.

[0015] The task is therefore to track the dynamics of railway power supply networks during operation as precisely as possible using easily accessible parameters and to optimize the railway power supply on the basis of these parameters.

[0016] This problem is solved by a method for predicting network admittances and vehicle performance based on current and voltage values ​​of substations of a railway power supply network according to claim 1, a method for optimizing a railway network according to claim 4, a forecasting device according to claim 5 and an optimization device according to claim 6.

[0017] In the inventive method for predicting network admittances and vehicle performance based on current and voltage values ​​of substations of a railway power supply network, a computer-based simulation of a plurality of labeled training railway power supply networks with different input parameters and a wide variation of rail traffic as input and output data is carried out, wherein current and voltage values ​​in the substations at each time point are output as output data.Here, "current and voltage values ​​of substations" refers to values ​​of supply voltages and supply currents. "Wide variation of rail traffic" refers to the consideration of input parameters for different scenarios of operating conditions of rail vehicles in a rail traffic system assigned to a railway power supply network, whereby the input parameters differ significantly from one another and cover a wide range of values. "Wide" range of values ​​is defined here as a range that covers a substantial portion, preferably nearly or even completely, of the entire range of possible input parameter values. The granularity of the values ​​may depend on the available computing power, storage capacity, and other resources available for generating, storing, and processing the training data, as well as for the training itself.The labeled training railway power supply networks now have the advantage that their simulation is essentially the reverse of the actual task of predicting key parameters, especially admittances and vehicle power, of a railway power supply network. If the admittances and vehicle power, as well as the architecture of the training railway power supply network, are known, the current and voltage values ​​can be calculated relatively easily using a computer simulation. Crucially, a time series of input data is required, from which important parameters, especially network admittances and vehicle power, can be inferred in the railway power supply networks. Since these parameters cannot usually be uniquely determined from a single measurement at a single point in time, entire time series must be recorded to allow for their unambiguous determination.A time series is understood to be a plurality of input data, especially measurement data, that are assigned to different successive points in time.

[0018] Furthermore, an artificial neural network is trained using the training traction power supply networks. The output data of the training traction power supply networks, specifically the current and voltage values, are used as input data for the artificial neural network, while the input data of the training traction power supply networks, specifically the admittances and vehicle power, are used as target data for the artificial neural network. Target data refers to specifications that the artificial neural network aims to achieve through a training process. This is also referred to as labeled training data.Advantageously, the artificial neural network can now be trained with the known labeled training data in such a way that it can determine the key parameters of a railway power supply network, in particular the admittances and vehicle performance, based on the current and voltage values, which is not possible by applying a “rigid model” due to the complexity and dynamics of such a system.

[0019] As part of an application of the trained algorithm or artificial neural network, time-dependent current and voltage values ​​of the substations of the railway power supply network are measured for a multiple of consecutive time points. The trained artificial neural network is then applied to these measured values ​​to determine network admittances and vehicle power outputs within the railway power supply network. The network admittances and vehicle power outputs can be used as inputs for an optimization model, allowing for the optimization of a desired parameter of the railway power network, such as energy consumption. A dynamic network optimization, which takes into account the actual behavior of rail traffic, can be advantageously performed.Another input variable can be the vehicle weight of the rail vehicles, which correlates with the utilization of the rail vehicles. Various losses can also be used as further input variables for a given network and / or train model: With a correct approximation, a ratio between power and losses can be assumed, but this ratio can change depending on the circumstances.

[0020] In the inventive method for optimizing the operation of a railway power supply network, the inventive method for predicting network admittances and vehicle performance based on current and voltage values ​​from substations of a railway power supply network is carried out. The operation of the railway power supply network is optimized based on the network admittances and vehicle performance determined by the method. In particular, optimizable parameters based on the aforementioned determined parameters, such as energy consumption or similar, can be optimized. The inventive method for optimizing the operation of a railway power supply network shares the advantages of the inventive method for predicting network admittances and vehicle performance based on current and voltage values ​​from substations of a railway power supply network.

[0021] The forecasting device according to the invention comprises a simulation unit for computer-based simulation of a plurality of labeled training railway power supply networks with different input parameters and a wide variation of rail traffic as input and output data, wherein current and voltage values ​​in the substations are output as output data at each time point. The forecasting device according to the invention also includes a training unit for training an artificial neural network using the training railway power supply networks, wherein the output data of the training railway power supply networks are used as input data of the artificial neural network and the input data of the training railway power supply networks are used as target data for the artificial neural network.Part of the forecasting device according to the invention is a measuring unit for measuring time-dependent current and voltage values ​​of the substations of the railway power supply network for a plurality of successive points in time.

[0022] The prediction device according to the invention also has an application unit for applying the trained artificial neural network to the measured time-dependent current and voltage values ​​to determine network admittances and vehicle performance in the railway power supply network.

[0023] Part of the prediction device according to the invention can also be a transmission unit from the measuring unit to the application unit, with which the measured current and voltage values ​​are transmitted to the application unit.

[0024] The forecasting device according to the invention shares the advantages of the inventive method for forecasting network admittances and vehicle performance based on current and voltage values ​​of substations of a railway power supply network.

[0025] The optimization device according to the invention comprises the forecasting device according to the invention and also an optimization unit for optimizing a railway power supply network based on the network admittances and vehicle performance determined by the forecasting device. The optimization device according to the invention shares the advantages of the method according to the invention for the operational optimization of a railway power supply network.

[0026] A large proportion of the aforementioned components of the forecasting or optimization device can be implemented wholly or partially as software modules within a processor of a corresponding computer system. A largely software-based implementation has the advantage that even existing computer systems can be easily retrofitted via a software update to operate according to the invention.The problem is therefore also solved by a corresponding computer program product comprising a computer program that can be directly loaded into a computer system, with program sections to execute the steps of the inventive method for predicting network admittances and vehicle performance based on current and voltage values ​​of substations in a railway power supply network, or the steps of the inventive method for optimizing the operation of a railway power supply network, when the program is executed in the computer system. In addition to the computer program, such a computer program product may optionally include additional components, such as documentation, and / or additional components, including hardware components, such as hardware keys (dongles, etc.) for using the software.

[0027] For transport to and / or storage on or in the computer system, a computer-readable medium, such as a memory stick, a hard drive, or other portable or permanently installed data storage device, can be used. This medium must contain the program sections of the computer program that can be read and executed by the computer system. The computer system may, for example, have one or more cooperating microprocessors or similar components. Furthermore, the computer system requires a connection to the data acquisition systems of the substations.

[0028] However, many processes are already standardized. Nevertheless, the connection still needs to be planned.

[0029] The dependent claims and the subsequent description each contain particularly advantageous embodiments and further developments of the invention. In particular, the claims of one claim category may also be further developed analogously to the dependent claims of another claim category and their descriptive parts. Furthermore, within the scope of the invention, the various features of different embodiments and claims may also be combined to form new embodiments.

[0030] In the first variant of the inventive method for predicting network admittances and vehicle performance based on current and voltage values ​​of substations of a railway power supply network, the input data or input parameters of the labeled training railway power supply networks comprise the following types of quantities: - Vehicle positions, - electrical power output of the vehicles, - Admittances resulting from the vehicle positions and vehicle performance.

[0031] It is possible to define further input parameters for improved accuracy. However, one of the strengths of the method according to the invention lies in its ability to determine the optimal state from as few arguments as possible. Therefore, the aforementioned types of input parameters are sufficient for the first variant of the outlined optimization problem.

[0032] In the second variant of the inventive method for predicting network admittances and vehicle performance based on current and voltage values ​​of substations of a railway power supply network, the computer-based simulation of a plurality of labeled training railway power supply networks is carried out on the basis of at least one of the following types of basic data or input parameters: - Infrastructure data, - Topographic data, - Vehicle type data, - Types and number of installed electrical components.

[0033] The aforementioned information can be used to accurately characterize the overall system in order to determine network admittances and vehicle performance.

[0034] The invention is explained in more detail below with reference to the accompanying figures and exemplary embodiments. The figures show: Fig. 1 a schematic representation of a section of a railway power supply network with three rail vehicles at one time, Fig. 2 a schematic representation of a section of a railway power supply network with two rail vehicles in a time series recording, Fig. 3 a flowchart illustrating a method for predicting network admittances and vehicle performance based on current and voltage values ​​of substations of a railway power supply network according to an embodiment of the invention, Fig. 4 a flowchart illustrating a method for optimizing the operation of a railway power supply network according to an embodiment of the invention, Fig. 5 a schematic representation illustrating a forecasting device according to an embodiment of the invention, Fig. 6 a schematic representation illustrating an optimization device according to an alternative embodiment of the invention.

[0035] In Fig. Figure 1 shows a schematic representation of a section of a railway power supply network 10 with three rail vehicles F1, F2, F3 at one time. The in Fig. The section shown comprises two substations U1 and U2, each with a busbar 1 and four track feeder cables 4. Also part of the section is a double-track overhead contact line system 2 with two contact lines 2a and 2b and contact line isolation points 5, which electrically isolate feeder sections from one another. On a first feeder section (in Fig. 1. On the upper feed section of a first overhead line 2a, a first rail vehicle F1 moves. On a second feed section (in Fig. 1. Two rail vehicles F2 and F3 are moving along the lower feed-in section of a second overhead line 2b. Deriving specific rules for predicting vehicle performance and network admittances based on known current and voltage values ​​is neither trivial nor intuitive in actual networks, as many different situations can prevail within the networks. The proposed invention provides that rules for estimating and determining the state of trains operating in the network could be defined. This can be implemented, for example, using trained neural networks as artificial intelligence.

[0036] In Fig. Figure 2 is a schematic representation of a section of a railway power supply network 10 with two rail vehicles F1 and F2 illustrated in a time series recording. The rail vehicles F1 and F2 move in opposite directions, which are shown in Fig. 2 are symbolized by arrows. Vehicles F1 and F2 cross overhead line sections 5 and switch between different feeder sections. Such a switch causes a significant change in the currents flowing at substations U1 and U2, which can be determined based on current-voltage ratios. Furthermore, the admittances of the power supply paths, which must be overcome to supply vehicles F1 and F2 with electrical energy, change due to the movement of the rail vehicles F1 and F2.

[0037] In Fig. Figure 3 is a flowchart 300 illustrating a method for predicting network admittances and vehicle performance based on current and voltage values ​​of substations of a railway power supply network according to an embodiment of the invention.

[0038] In step 3.I, a computer simulation of a plurality of labeled training railway power networks (TBNs) with different input parameters and a wide variation of rail traffic as input and output data is performed, whereby the output data are values ​​of currents I and voltages U, which are measured in the substations at each point in time.

[0039] In step 3.II, an artificial neural network KNN is trained by the training railway current networks TBN, whereby the output data of the training railway current networks TBN are used as input data of the artificial neural network and the input data of the training railway current networks TBN are used as target data for the artificial neural network KNN.

[0040] In step 3.III, time-dependent current and voltage values ​​of the substations of the railway power supply network are measured for a multiple of successive time points.

[0041] Finally, in step 3.IV, the trained artificial neural network TKNN is applied to the measured time-dependent current and voltage values ​​I, U to determine network admittances NA and vehicle power P. F applied in the railway power supply network.

[0042] In Fig. Figure 4 is a flowchart 400, which illustrates a method for optimizing the operation of a railway power supply network according to an embodiment of the invention.

[0043] In step 4.I, this is done in Fig. Figure 3 illustrates a method for predicting network admittances and vehicle performance based on current and voltage values ​​of substations in a railway power supply network, according to an embodiment of the invention. Based on the prediction generated in step 4.I, the railway power supply network is optimized in step 4.II.

[0044] In Fig. Figure 5 is a schematic representation illustrating a forecasting device 50 according to an embodiment of the invention.

[0045] The forecasting device 50 has a simulation unit 51 for computer-based simulation of a plurality of labeled training railway power networks TBN with different input parameters EP with a wide variation of rail traffic as input data and output data, wherein current I and voltage U in the substations U1, U2 are output as output data at each time point in time.

[0046] Furthermore, the forecasting device 50 includes a training unit 52 for training an artificial neural network (ANN) using the training railway power networks (TBNs), wherein the output data of the training railway power networks (TBNs) are used as input data for the artificial neural network (ANN) and the input data of the training railway power networks (TBNs) are used as target data for the artificial neural network (ANN). Training the artificial neural network (ANN) results in the creation of a trained artificial neural network (TKNN).

[0047] Part of the forecasting device 50 is also a measuring unit 53 for measuring time-dependent current and voltage values ​​I, U of the substations U1, U2 of the railway power supply network 10 for a plurality of successive time points.

[0048] The forecasting device 50 has an application unit 54 for applying the trained artificial neural network TKNN to the measured time-dependent current and voltage values ​​I, U to determine network admittances NA and vehicle power P F in the railway power supply network 10.

[0049] In Fig. Figure 6 is a schematic representation illustrating an optimization device 60 according to an alternative embodiment of the invention.

[0050] The optimization device 60 has a forecasting device 50, as described in Fig. Figure 5 illustrates this. Furthermore, the optimization device 60 includes an optimization unit 61 for optimizing a railway power supply network 10 based on the network admittances NA and vehicle power P determined by the forecasting device 50. F. As a result of the optimization process of the optimization device 60, an optimized railway power supply network 10 is output.

[0051] Finally, it should be noted once again that the methods and devices described above are merely preferred embodiments of the invention and that the invention can be varied by a person skilled in the art without departing from the scope of the invention, insofar as it is defined by the claims. For the sake of completeness, it should also be noted that the use of the indefinite articles "a" or "an" does not preclude the possibility that the features in question may be present multiple times. Likewise, the term "unit" does not preclude the possibility that it consists of several components, which may also be spatially distributed. Regardless of the grammatical gender of a particular term, persons of male, female, or other gender identities are included. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] CN 115549196-A

[0013] EP 3715172

[0013] FR 3094288-A1

[0013] AU 2020202019-A1

[0013] IN 202014012075-A

[0013] FR 3094288-B1

[0013] EP 3715172-B1

[0013] ES 2886046-T3

[0013] Cited non-patent literature

[0000] Pablo Martinez Fernandez, Ignacio Villalba Sanchis, Victor Yepes, Ricardo Insa Franco, “A review of modeling and optimization methods applied to railways energy consumption”, in Journal of Cleaner Production, Volume 222, 2019, Pages 153-162

[0012] S. Khayyam, N. Berr, L. Razik, M. Fleck, F. Ponci and A. Monti, „Railway System Energy Management Optimization Demonstrated at Offline and Online Case Studies,“ in IEEE Transactions on Intelligent Transportation Systems, vol. 19, no. 11, pp. 3570-3583, Nov. 2018, doi: 10.1109 / TITS.2018.2855748

[0012] CHONGQING CRRC TIMES ELECTRIC TECHNOLOGY CO., LTD., (2022) „Urban rail intelligent traction power supply system for use in urban rail traffic system has energy operation control system used to receive bidirectional converter set power, device operation parameter, state, comprehensive vehicle position and traction / braking power

[0013] ALSTOM TRANSPORT TECHNOLOGIES, (2020) „Method of dynamically adapting the operation of traction substation of an electric power supply system for vehicles, involves determining of the vehicles currently circulating in the section specifically associated with substation

[0013]

Claims

[1] Methods for predicting network admittances (NA) and vehicle performance (P) F ) based on current and voltage values ​​(I, U) of substations (U1, U2) of a railway power supply network (10), comprising the steps: - Computer-based simulation of a plurality of labeled training railway power supply networks (TBN) with different input parameters (EP) with a wide variation of rail traffic as input and output data, wherein current values ​​(I) and voltage values ​​(U) of substations (U1, U2) are output as output data per time point, - Training an artificial neural network (ANN) using the training railway power supply networks (TBN), where the output data of the training railway power supply networks (TBN) are used as input data of the artificial neural network (ANN) and the input data of the training railway power supply networks (TBN) are used as target data for the artificial neural network (ANN), - Measuring time-dependent current and voltage values ​​(I, U) of the substations (U1, U2) of the railway power supply network (10) for a plurality of successive time points, - Applying the trained artificial neural network (TKNN) to the measured time-dependent current and voltage values ​​(U, I) to determine values ​​of network admittances (NA) and vehicle power (P). F ) in the railway power supply network (10). [2] Method according to claim 1, wherein the input data of the labeled training railway power supply networks (TBN) comprise the following quantities: - Rail vehicle positions, - electrical power (P F ) of the rail vehicles (F1, F2, F3). [3] Method according to claim 1 or 2, wherein the computer-based simulation of a plurality of labeled training railway power supply networks (TBNs) is carried out on the basis of input data of at least one type of the following basic data: - Infrastructure data, - Topographic data, - Vehicle type data, - Timetable data, - installed electrical components. [4] Method for optimizing the operation of a railway power supply network (10), comprising the steps: - Carrying out the procedure according to one of the foregoing claims, - Optimizing the operation of the railway power supply network (10) based on the network admittances (NA) and vehicle performance (P) determined in the procedure F ). [5] Forecasting device (50), comprising: - a simulation unit (51) for computer-based simulation of a plurality of labeled training railway power supply networks (TBN) with different input parameters (EP) with a wide variation of rail traffic as input data and output data, wherein current values ​​(I) and voltage values ​​(U) of substations (U1, U2) are output as output data per time point, - a training unit (52) for training an artificial neural network (ANN) through the training railway power supply networks (TBN), wherein the output data of the training railway power supply networks (TBN) are used as input data of the artificial neural network (ANN) and the input data of the training railway power supply networks (TBN) are used as target data for the artificial neural network (ANN), - a measuring unit (53) for measuring time-dependent current and voltage values ​​(I, U) of the substations (U1, U2) of the railway power supply network (10) for a plurality of successive time points, - an application unit (54) for applying the trained artificial neural network (TKNN) to the measured time-dependent current and voltage values ​​(I, U) to determine network admittances (NA) and vehicle performance (P) F ) in the railway power supply network (10). [6] Optimization device (60), comprising: - a forecasting device (50) according to claim 5, - an optimization unit (61) for optimizing a railway power supply network (10) on the basis of the network admittances (NA) and vehicle power (P) determined by the forecasting device (50). F ). [7] Computer program product comprising instructions which, when the program is executed by a computer, cause it to perform the steps of a method according to any one of claims 1 to 4. [8] Computer-readable storage medium comprising instructions which, when executed by a computer, cause it to perform the steps of a method according to claims 1 to 4.

Images