WO2021130298A1

WO2021130298A1 - System, apparatus and method for managing energy consumption at a technical installation

Info

Publication number: WO2021130298A1
Application number: PCT/EP2020/087744
Authority: WO
Inventors: Poonam JYOTI; Akash MITTAL; Vinay Ramanath; Arko CHATTERJEE
Original assignee: Siemens Mobility GmbH
Priority date: 2019-12-26
Filing date: 2020-12-23
Publication date: 2021-07-01

Abstract

A system (105), an apparatus (110) and a method for managing energy consumption at a technical installation (107) is disclosed. The method comprises computing, by a processing unit (130), energy units likely to be consumed by the technical installation (107) based on data corresponding to one or more parameters using an artificial intelligence model (300). The one or more parameters are associated with factors impacting an operation of the technical installation (107). Further, a load forecast model for a predefined time period is generated based on the computed energy units likely to be consumed by the technical installation (107). Further, an overall impact on the technical installation (107) is determined by analysing the load forecast model. Further, one or more actions are performed at the technical installation (107) in such a manner that the overall impact on the technical installation (107) is kept optimal.

Description

SYSTEM, APPARATUS AND METHOD FOR MANAGING ENERGY CONSUMPTION AT A TECHNICAL INSTALLATION

Description

The present invention generally relates to managing energy consumption at a technical installation, and more particularly relates to a system, apparatus and method for managing energy consumption at a technical installation.

Typically, utility companies require technical installations such as substations, railways and manufacturing units to provide in advance, an electrical load forecast for a predefined time period. The electrical load forecast, hereinafter referred to as a load forecast, includes a prediction of demand for electrical energy by the technical installation. The demand may be one of a short-term demand, medium-term demand and long-term demand. The load forecast helps utility companies in managing power generation based on requirements elicited from the technical installations. However, if the load forecast for the predefined time period is different from the actual load consumed during the predefined time period, the utility company may impose a penalty on the technical installation. The penalty may be proportional to the difference in the energy units provided in the load forecasted and the actual energy units consumed by the technical installation.

In light of the above, it is necessary to forecast energy consumption at a technical installation in an accurate manner and manage the energy consumption at the technical installation based on the forecasted energy consumption. Therefore, it is an object of the present invention to provide a system, an apparatus and a method for forecasting energy consumption at a technical installation and managing energy consumption at the technical installation based on the forecasted energy consumption.

The object of the present invention is achieved by a computer-implemented method for managing energy consumption at a technical installation. The term 'technical installation' as used herein may refer to any one of a mobility infrastructure, a manufacturing facility, a process plant, a commercial establishment, or any other complex set-up where it may be necessary to forecast electrical load at any given point of time.

The method comprises computing, by a processing unit, energy units likely to be consumed by the technical installation based on data corresponding to one or more parameters using an artificial intelligence model. The term 'data' as used herein may refer to parameter values associated with the one or more parameters. The data may be associated with real-time values of the one or more parameters, historic values of the one or more parameters, near real-time values associated with one or more parameters or a combination thereof. The one or more parameters may be associated with technical factors, environmental factors, design factors, engineering factors, geopolitical factors and time-related factors impacting the operation of the technical installation. For example, the technical factors may include factors related to maintenance, capacity, power supply requirements and so on, associated with electrical loads in the technical installation. The environmental factors may include weather conditions such as temperature or precipitation levels that may impact the operation of the technical installation. The design factors may include design specifications associated with one or more machinery operating within the technical installation. The geopolitical factors may include factors such as protests, national emergencies or calamities that may impact the operation of the technical installation. The time-related factors may include a time of the day, a season and so on that may impact the operation of the technical installation.

In a preferred embodiment, the method of computing the energy units likely to be consumed by the technical installation comprises obtaining the data corresponding to the one or more parameters associated with the energy consumption by the technical installation. The data may be obtained from one or more devices communicatively coupled to the processing unit. Further, the method comprises computing the impact of each of the one or more parameters on the operation of the technical installation. In one embodiment, the impact of each of the one or more parameters on the operation of the technical installation is computed by determining a correlation between each of the one or more parameters and the energy units consumed by the technical installation, and computing the impact associated with each of the one or more parameters based on the correlation determined. The impact may be computed as an impact level associated with the one or more parameters. The impact level may be calculated as a value, say between -1 and 1. For example, if the parameter is associated with a geopolitical factor such as a political protest or calamity, then the impact level may be computed as 0.8 if a political protest is present. In one example, the impact level is the same as a correlation coefficient between a parameter and the energy units consumed. In another example, the impact may be calibrated based on the correlation coefficient. In one embodiment, the correlation is determined by analysing historic data associated with the one or more parameters and corresponding energy units consumed by the technical installation. Further, at least one parameter is dynamically selected from the one or more parameters based on the impact computed for each of the one or more parameters. In one example, each of the impact levels may be assigned a rank in increasing or decreasing order of values. Further, the parameters having selected ranks, say first, second and third ranks may be selected.

In another example, the at least one parameter may be selected by comparing the impact levels for each of the parameters against a predefined threshold, e.g., all parameters having impact levels greater than 0.5 may be selected. In yet another example, all parameters having impact levels in a predefined range, say between 0.5 and 1, may be selected. Further, the at least one parameter selected is provided as input to the artificial intelligence model for computing the energy units likely to be consumed by the technical installation.

Advantageously, dynamic selection of the at least one parameter helps in filtering out irrelevant parameters, thus enabling faster computation of the energy units likely to be consumed by the technical installation, compared to existing art.

The artificial intelligence model may be any model that may be trained using machine learning techniques, such as decision trees, support vector machines, random forest networks, deep neural networks, recurrent neural networks and so on. The artificial intelligence model may be trained using a training dataset to compute the energy units likely to be consumed by the technical installation for a predefined time period by using machine learning techniques. The machine learning techniques may include supervised learning, unsupervised learning or reinforcement learning. The artificial intelligence model may be trained using the historic data associated with the energy consumption by the technical installation. For example, the historic data from the past five years may be used for training the artificial intelligence model. The artificial intelligence model may be retrained regularly, say every week, based on updated historic data. In one embodiment, the artificial intelligence model is selected from a plurality of artificial intelligence models based on a predefined criteria. The predefined criteria may be associated with one or more of the number of parameters selected, type of parameters selected and the values of the parameters selected.

In a preferred embodiment, the artificial intelligence model is trained based on at least one penalty formula associated with the technical installation. More specifically, the error function that defines an error between an input to the artificial intelligence model and an output from the artificial intelligence model, is based on the penalty formula associated with the technical installation. The penalty formula may provide penalty tariffs imposed for the different time intervals within the predefined time period. The penalty tariffs indicate a penalty amount imposed by the utility company on the technical installation when the computed energy units likely to be consumed by the technical installation is different from the actual energy units consumed. The penalty tariffs may be defined for different time intervals within the day. The error function may be associated with at least one of a mean absolute error and a mean square error. In a preferred embodiment, the error function is a parameterized error function that varies throughout the pre-defined time period based on the variations in the penalty tariffs. In other words, the error function is a function of time and is defined for different intervals of time within the predefined time period. For example, a first error function corresponding to a first time interval may be associated with 2%, a second error function corresponding to to a second time interval may be associated with 3.5% and so on.

Advantageously, the artificial intelligence model trained using the parameterized error function helps in preventing overfitting of the artificial intelligence model to abnormalities in the energy consumption by the technical installation. Further, the artificial intelligence model is trained based on the penalty formula in order to reduce the difference between the energy units likely to be consumed by the technical installation and the actual energy units consumed by the technical installation. In other words, the error function is defined such that when the artificial intelligence model is trained based on the error function so that the accuracy of forecasting is improved, thus reducing the penalty amount that may be imposed on the technical installation.

Furthermore, the method comprises generating a load forecast model for a predefined time period based on the computed energy units likely to be consumed by the technical installation. In one embodiment, generating the load forecast model for the predefined time period based on the computed energy units likely to be consumed by the technical installation comprises determining at least one mathematical relation between the energy units likely to be consumed by the technical installation and one or more specific points in time within the predefined time period. The mathematical relation may be determined based on the computed energy units likely to be consumed by the technical installation, using a curve-fitting technique such as regression analysis. The mathematical relation may be a time-based function. Further, the load forecast model is generated based on the at least one mathematical relation determined. The load forecast model may be at least one of an analytical model, a numerical model and an empirical model indicative of the energy units likely to be consumed by one or more components in the technical installation at the one or more specific points in time within the predefined time period. In one example, the predefined time period may be associated with one day and the one or more specific points in time may be defined for each hour in the day. In another example, the predefined time period may be associated with a month and the one or more specific points in time may be defined for every day within the month. In yet another example, the predefined time period may be associated with a year and the one or more specific points in time may be defined for every week or month in the year.

Advantageously, the present invention may be adapted to generate the load forecast models for different demand periods using the data corresponding to the one or more parameters.

The method further comprises determining an overall impact on the technical installation by analysing the load forecast model. The overall impact may be associated with at least one of a financial impact and an operational impact. In one embodiment, determining the overall impact on the technical installation by analysing the load forecast model comprises determining a trend associated with the load forecast model within the predefined time period. In one example, the trend may be a time-series plot associated with the energy consumption at the specific points in time over the predefined time period. Further, the overall impact on the technical installation is computed by analysing the trend associated with the load forecast model. The trend may be analysed based on different factors such as time of the day corresponding to a certain number of energy units likely to be consumed, total number of energy units likely to be consumed over the predefined time period, rate of change of energy consumption over the predefined time period, outages and so on. Based on the analysis, the operational impact may be determined as an operational efficiency of the technical installation or a reliability of the technical installation during the predefined time period. For example, the operational efficiency may be determined based on the energy units likely to be consumed at a specific point in time within the predefined time period and energy efficiency associated with one or more components in the technical installation. In another example, the financial impact may be computed as total cost of energy based on the total energy units likely to be consumed by the technical installation and a predefined cost per energy unit. In one example, the total energy units likely to be consumed may be determined by integrating the energy units likely to consumed by the technical installation, over a time-range in the trend. Further, the total energy units likely to be consumed may be multiplied with the predefined cost per energy unit to get the total cost of energy.

Moreover, the method comprises performing one or more actions at the technical installation in such a manner that the overall impact on the technical installation is kept optimal. In a preferred embodiment, the method for performing the one or more actions at the technical installation comprises scheduling an operation of the technical installation in order to keep the overall impact on the technical installation optimal. Alternatively, recommendations for scheduling the operation may be generated for the determined overall impact on the technical installation based on one or more rules stored in a database. Further, a notification indicating the recommendations for scheduling the operation may be generated on a Graphical User Interface (GUI). In another embodiment, the method for performing the one or more actions at the technical installation comprises sending instructions to an energy management system installed at the technical installation. The energy management system may include computer-aided tools that may be used for monitoring, controlling and/or optimizing energy consumed by the technical installation. For example, the instructions may be provided to the energy management system for utilization of low tariff periods within the predefined time period for operation of large loads, prevention of overloading of certain components of the technical installation, improving power quality and so on.

Advantageously, the present invention helps in automatically managing the energy consumption through the one or more actions, thus reducing human errors by requiring less human intervention.

The method may comprise computing a deviation between the energy units likely to be consumed and a predetermined load at a specific point in time within the predefined time period. The deviation may be computed as a difference between the energy units likely to be consumed and the predetermined load for the specific point in time. The deviation may result in a surge or spike in the energy consumption by the technical installation. The predetermined load may be a maximum load, a minimum load or an optimal load defined for the technical installation. In one example, the predetermined load may be a maximum load defined for the technical installation and the energy units likely to be consumed by the technical installation may be greater than the maximum load defined for the technical installation. Based on the computed deviation, one or more actions may be performed for reducing the computed deviation. In the present example, the one or more actions may be associated with reducing the energy units consumed by the technical installation to a value less than or equal to the maximum load by say, disconnecting one or more predetermined non-crucial loads before the specific point in time. In other words, the operation of the technical installation may be scheduled such that the deviations are reduced.

Advantageously, the method for performing the one or more actions for reducing the deviations help in reducing the penalty amount that may be imposed by a utility company on the technical installation for usage above contracted energy units.

Additionally, the method may comprise generating at least one simulation instance based on the load forecast model.

In one embodiment, the load forecast model may be used to generate an analytical model in machine-executable form for performing the simulation. The simulation instance may refer to a thread of simulation independent of all other threads during execution. The method may comprise executing the simulation instance in a simulation environment for different values of the one or more parameters. The simulation instance may be executed in the simulation environment as stochastic simulations, deterministic simulations, dynamic simulations, continuous simulations, discrete simulations, local simulations, distributed simulations and so on. The method further comprises analysing simulation results generated from the simulation for determining an influence of the one or more parameters on the load forecast model. The analysis may be performed based on descriptive techniques, exploratory techniques, inferential techniques, predictive techniques, causal techniques, qualititative analysis techniques, quantitative analysis techniques and so on.

Advantageously, the present invention analyses the effects of different parameters on the load forecast model, thereby aiding an operating personnel in decision-making process.

Also, the method may comprise outputting the load forecast model on the GUI. In one example, the load forecast model is outputted in the form of a graphical plot between the energy units likely to be consumed and the specific points in time. In another example, the load forecast model may be shown in tabular format on the GUI. The GUI may be associated with electronic devices, including but not limited to, personal computers, workstations, mobile phones and personal digital assistants. The notification may be generated on two-dimensional displays, Augmented Reality based interfaces, Virtual Reality based interfaces or Mixed Reality based interfaces. In one example, the notification may be generated in the form of pop-up display on the GUI. In addition to the load forecast model, the one or more actions performed to optimize the overall impact on the technical installation may also be shown on the GUI. Further, analysis of the simulation results may also be displayed on the GUI. In a preferred embodiment, the GUI is an interactive display that allows the operating personnel to visualise variations in the load forecast model for different values of the one or more parameters. In one embodiment of the present invention, the technical installation is at least one railway line in a rail infrastructure. The railway infrastructure may include one or more railway stations, overhead electric lines that power electric trains, signalling units and other electrically operated subsystems that are used for managing operation of rail vehicles. The railway line may comprise a plurality of railway stations and overhead electric lines that supply power to the electric trains. The energy consumption in the railway line may vary based on number of electric trains operated, number of passengers boarding the electric trains, number of passenger cars on the electric train, technical specifications associated with the locomotive engines used in the electric train, running delays associated with the electric trains and so on. The number of passengers boarding the electric train may in turn depend on several other factors such as time of the day, presence of holidays or protests, natural calamaties, weather, delays, road traffic situations and so on. In other words, the above-mentioned parameters affect the energy consumption along the railway line.

The object of the present invention is achieved by an apparatus comprising one or more processing units, and a memory unit communicatively coupled to the one or more processing units. The memory unit comprises one or modules stored in the form of machine-readable instructions executable by the one or more processing units. The one or more modules are configured to perform method steps described above. The execution of the one or more modules may also be performed using co-processors such as Graphical Processing Unit (GPU), Field Programmable Gate Array (FPGA) or Neural Processing/Compute Engines. In addition, the memory unit may also include a database. According to an embodiment of the present invention, the apparatus can be an edge computing device. As used herein "edge computing" refers to computing environment that is capable of being performed on an edge device (e.g., connected to the sensors unit in an industrial setup on one end and to a remote server(s) such as for computing server(s) or cloud computing server(s) on other end), which may be a compact computing device that has a small form factor and resource constraints in terms of computing power. A network of the edge computing devices can also be used to implement the apparatus. Such a network of edge computing devices is referred to as a fog network.

In another embodiment, the apparatus is a cloud computing system having a cloud computing based platform configured to provide a cloud service for managing energy consumption by the technical installation. As used herein, "cloud computing" refers to a processing environment comprising configurable computing physical and logical resources, for example, networks, servers, storage, applications, services, etc., and data distributed over the network, for example, the internet. The cloud computing platform may be implemented as a service for analyzing data. In other words, the cloud computing system provides on-demand network access to a shared pool of the configurable computing physical and logical resources. The network is, for example, a wired network, a wireless network, a communication network, or a network formed from any combination of these networks.

Additionally, the object of the present invention is also achieved by a system for managing energy consumption by the technical installation. The system comprises one or more devices capable of providing the data corresponding to the one or more parameters and an apparatus as described above, communicatively coupled to the one or more sources. The one or more devices may include sensing units, electronic devices, edge devices, servers, databases and so on. The electronic device may include any device capable of sending data across a network. Non-limiting examples of electronic devices include personal computers, laptop computers, personal digital assistant (PDA), tablets and mobile phones.

The object of the present invention is also achieved by a computer-program product having machine-readable instructions stored therein, which when executed by a processor, cause the processor to perform a method as describe above.

The above-mentioned attributes, features, and advantages of this invention and the manner of achieving them, will become more apparent and understandable (clear) with the following description of embodiments of the invention in conjunction with the corresponding drawings. The illustrated embodiments are intended to illustrate, but not limit the invention.

The present invention is further described hereinafter with reference to illustrated embodiments shown in the accompanying drawings, in which:

FIG 1A illustrates a block diagram of a system for managing energy consumption by a rail infrastructure, in accordance with one embodiment of the present invention; FIG IB illustrates functional components of a forecasting module, in accordance with one embodiment of the present invention;

FIG 2 illustrates a flowchart of a method for computing the energy units likely to be consumed along the railway line, in accordance with one exemplary embodiment of the present invention;

FIG 3 illustrates a deep neural network, in accordance with one exemplary embodiment of the present invention, in accordance with one embodiment of the present invention;

FIG 4 illustrates a flowchart of a method for managing energy consumption along the railway line, in accordance with one exemplary embodiment of the present invention; and

FIG 5 illustrates a flowchart of a generalised method for managing energy consumption at a technical installation is shown, in accordance with one embodiment of the present invention.

Hereinafter, embodiments for carrying out the present invention are described in detail. The various embodiments are described with reference to the drawings, wherein like reference numerals are used to refer to like elements throughout. In the following description, for purpose of explanation, numerous specific details are set forth in order to provide a thorough understanding of one or more embodiments. It may be evident that such embodiments may be practiced without these specific details. Referring to FIGS 1A and IB, a system 105 for managing energy consumption along a railway line within a rail infrastructure 107 is shown, in accordance with one embodiment of the present invention. For ease of explanation and for the sake of brevity, the present invention is explained by considering only the energy consumed by electric trains operating within the railway line.

The system 105 comprises an apparatus 110. In the present embodiment, the energy consumption at the railway infrastructure is managed based by considering factors such as holidays, political events, natural calamaties, time of the day, time of the year, presence of rains and temperature levels. Each of the above-mentioned factors may be quantified using one or more parameters. The one or more parameters may have same or different values across specific points in time within a predefined time period during which the energy units are likely to be consumed. In the present example, the parameters may have same or different values for each hour within a day during which the energy units are likely to be consumed. For example, a first parameter may indicate name of a holiday or a political event, for example, 'Christmas', 'Hartal', 'national emergency'; a second parameter may indicate a time-stamp or a time-range; a third parameter may indicate a precipitation level, for example, '20%'; and a fourth parameter may indicate temperature, for example, '25°C', '104°F'. The data corresponding to the parameters may be obtained from a plurality of sources. In the present embodiment, the data may be extracted based on information gathered from an electronic device 115 and a workstation 120. The information gathered may be, for example, in the form of structured text or unstructured text.

The apparatus 110 is communicatively coupled to the electronic device 115 and the workstation 120 over a network 122. The electronic device 115 may be associated with a weather station that provides regular weather updates and weather forecasts. The workstation 120 may be associated with railway personnel and may be located within the rail infrastructure 107. Further, the workstation 120 may be configured to receive inputs associated with holidays, political events, natural calamities and so on. The workstation 120 may also receive inputs related to unexpected holidays or events. The inputs may be in the form of text, speech or gestures. Speech or gesture-based inputs may be further converted to text for enabling further processing. The text may be structured text or unstructured text.

The apparatus 110 may be a (personal) computer, a workstation, a virtual machine running on host hardware, a microcontroller, or an integrated circuit. As an alternative, the apparatus 110 may be a real or a virtual group of computers (the technical term for a real group of computers is "cluster", the technical term for a virtual group of computers is "cloud").

The apparatus 110 includes a communication unit 125, one or more processing units 130, a display 135, a Graphical User Interface (GUI) 140 and a memory 145. communicatively coupled to each other. In one embodiment, the communication unit 125 includes a transmitter (not shown), a receiver (not shown) and Gigabit Ethernet port (not shown). The memory 145 may include 2 Giga byte Random Access Memory (RAM) Package on Package (PoP) stacked and Flash Storage. The one or more processing units 130 are configured to execute the defined computer program instructions in the modules. Further, the one or more processing units 130 are also configured to execute the instructions in the memory 145 simultaneously. The display 135 includes a High- Definition Multimedia Interface (HDMI) display and a cooling fan (not shown). Additionally, operating personnel may access the apparatus 110 through the GUI 140. The GUI 140 may include a web-based interface, a web-based downloadable application interface, and so on.

The processing unit 130, as used herein, means any type of computational circuit, such as, but not limited to, a microprocessor, microcontroller, complex instruction set computing microprocessor, reduced instruction set computing microprocessor, very long instruction word microprocessor, explicitly parallel instruction computing microprocessor, graphics processor, digital signal processor, or any other type of processing circuit. The processing unit 130 may also include embedded controllers, such as generic or programmable logic devices or arrays, application specific integrated circuits, single-chip computers, and the like.

In general, a processing unit 130 may comprise hardware elements and software elements. The processing unit 130 can be configured for multithreading, i.e. the processing unit 130 may host different calculation processes at the same time, executing them either in parallel or switching between active and passive calculation processes.

The memory 145 may be volatile memory and non-volatile memory. The memory 145 may be coupled for communication with the processing unit 130. The processing unit 130 may execute instructions and/or code stored in the memory 145. A variety of computer-readable storage media may be stored in and accessed from the memory 145. The memory 145 may include any suitable elements for storing data and machine- readable instructions, such as read only memory, random access memory, erasable programmable read only memory, electrically erasable programmable read only memory, a hard drive, a removable media drive for handling compact disks, digital video disks, diskettes, magnetic tape cartridges, memory cards, and the like.

In the present embodiment, the memory 145 includes a forecasting module 150, a training module 155, an analysis module 160, a simulation module 165, a report-generation module 170 and a database 175 in the form of machine- readable instructions on any of the above-mentioned storage media and may be in communication to and executed by processing unit 130. The following description explains functions of the modules when executed by the one or more processing units 130.

The database 175 may store historic data associated with energy consumption along the railway line, for say the past five years. In the present example, the historic data may include energy units consumed along the railway line each hour for every day in the past five years. The historic data may further comprise parameter values or data associated with each of the parameters corresponding to the energy units consumed in each hour. In addition, the database 175 may also store impacts computed for each of the parameter values. In addition, the database 175 may also store list of holidays, political events and so on for each calendar year. In one implementation, the list of holidays, political events and so on may be updated based on inputs received from the electronic device 115 or the workstation 120. In another implementation, the apparatus 110 may automatically detect presence of an event based on a trend in the energy units consumption. The historic data may also include number of passengers travelling on each of the electric trains, running delays associated with the electric trains, penalties paid for each day due to the difference between actual energy units consumed and the energy units forecasted for the day and so on. Further, the database 175 may also store penalty formulae used for calculating the penalties. The penalty formulae may be stored for different geographical locations associated with different railway lines.

In one embodiment, the forecasting module 150 further comprises a data extraction module 180, a selection module 185, a computation module 190 and a generation module 195. The data extraction module 180 is configured for extracting the data associated with the parameters from the information obtained from the electronic device 115 and the workstation 120. More specifically, the artificial intelligence model may extract the data from structured or unstructured information. The selection module 185 is configured for computing impacts associated with each of the parameters. Further, based on the impacts computed, at least one parameter may be dynamically selected for computing the energy units likely to be consumed by the technical installation. The computation module 190 is configured for computing the energy units likely to be consumed along the railway line using the at least one parameter selected by the selection module. The generation module 195 is configured for generating the load forecast model for the railway line based on the energy units likely to be consumed along the railway line during the predefined time period. For example, the generation module 195 may build the load forecast model from the computed energy units likely to be consumed by the rail infrastructure 107 at specific points in time.

The training module 155 is configured for training the plurality of artificial intelligence models based on the historic data, using two parametrised error functions. The parametrised error functions are associated with mean absolute error and mean square error. The parametrised error functions may be a function of time and may have different values of error for different points of time. The analysis module 160 is configured for computing deviations between the energy units likely to be consumed during the predefined time period and a predetermined load. The analysis module 160 may further generate one or more recommendations for reducing the deviation based on rules stored in the database 175. The simulation module 165 is configured for simulating the load forecast model to determine effect of variations in one or more parameters on the load forecast model.

The report-generation module 170 is configured to generate reports related to the load forecast model. For example, the report-generation module 170 may generate a report indicating the load forecast model for November 20, 2019 from 00:00 am to 23:59 pm with a granularity of 15 minutes. The reports may generated based on predefined templates stored in the database 175. The reports may include plots, formulae, costs associated with the energy units likely to be consumed, deviations, recommendations for reducing the deviations, overall impact on the rail infrastructure 107, actions performed for keeping the overall impact on the rail infrastructure 107 optimal and so on. The report generated by the report-generation module 170 may be displayed on the GUI 140.

Those of ordinary skilled in the art will appreciate that the hardware depicted in FIG 1A may vary for different implementations. For example, other peripheral devices such as an optical disk drive and the like, Local Area Network (LAN)/ Wide Area Network (WAN)/ Wireless (e.g., Wi-Fi) adapter, graphics adapter, disk controller, input/output (I/O) adapter, network connectivity devices also may be used in addition or in place of the hardware depicted. The depicted example is provided for the purpose of explanation only and is not meant to imply architectural limitations with respect to the present disclosure.

A system in accordance with an embodiment of the present disclosure includes an operating system employing a graphical user interface. The operating system permits multiple display windows to be presented in the graphical user interface simultaneously with each display window providing an interface to a different application or to a different instance of the same application. A cursor in the graphical user interface may be manipulated by a user through the pointing device. The position of the cursor may be changed and/or an event such as clicking a mouse button, generated to actuate a desired response.

One of various commercial operating systems, such as a version of Microsoft Windows™ may be employed if suitably modified. The operating system is modified or created in accordance with the present disclosure as described.

The present invention is not limited to a particular computer system platform, processing unit, operating system, or network. One or more aspects of the present invention may be distributed among one or more computer systems, for example, servers configured to provide one or more services to one or more client computers, or to perform a complete task in a distributed system. For example, one or more aspects of the present invention may be performed on a client-server system that comprises components distributed among one or more server systems that perform multiple functions according to various embodiments. These components comprise, for example, executable, intermediate, or interpreted code, which communicate over a network using a communication protocol. The present invention is not limited to be executable on any particular system or group of system, and is not limited to any particular distributed architecture, network, or communication protocol.

Disclosed embodiments provide systems and methods for managing energy consumption at a technical installation, specifically a rail infrastructure.

Referring to FIG 2, in conjunction with FIGS 1A and IB, a flowchart of a method 200 for computing the energy units likely to be consumed along the railway line is shown, in accordance with one exemplary embodiment of the present invention. The method comprises steps 205-220.

At step 205, the data extraction module 180 obtains data corresponding to the parameters associated with the energy consumption along the railway line based on information received from the electronic device 115 and the workstation 120. In case of structured information, the data may be extracted based on labels present in the information. Similarly, in case of unstructured information, the data may be extracted based on patterns present in the information. More specifically, data corresponding to a specific point in time is extracted. For example, for computing the energy units likely to be consumed at 10:30 pm on 20th November, 2019, the data corresponding to values of the parameters at the specific point in time are extracted.

At step 210, the selection module 185 computes impact of each of the parameters on the operation of electric trains along the railway line. The impact associated with each of the parameters may be computed by performing correlation analysis on the historic data in order to obtain a correlation between the energy units consumed and the parameters. The correlation analysis technique may be for example, Spearman's correlation technique or any other correlation techniques known in the art. Further, correlation coefficients derived from the correlation analysis for each of the parameters may be assigned as the computed impact for the parameter. For example, the impact associated with 'Christmas' may be determined as is 0.9, the impact associated with temperature of '18°C' may be determined as 0.20, the impact associated with precipitation of '35%' may be determined as 0.6, the impact associated with time stamp of 11:45:00 pm may be 0.7 and so on.

At step 215, the selection module 185 dynamically selects at least one parameter from the parameters based on the impact computed for each of the parameters. In the present embodiment, the parameters having impact greater than, say 0.3 may be selected.

At step 220, the at least one parameter is provided as input to an artificial intelligence model for computing the energy units likely to consumed along the railway line. The computation module 190 uses a plurality of deep neural network based artificial intelligence models. Further, the artificial intelligence model for computing the energy units likely to be computed may be selected from the plurality of artificial intelligence models may be selected based on the type of the selected parameters. FIG 3 shows a deep neural network based artificial intelligence model, in accordance with one exemplary embodiment of the present invention.

Referring to FIG 3, in conjunction with FIGS 1A, IB and 2, a deep neural network 300 is shown, in accordance with one exemplary embodiment of the present invention.

The deep neural network is associated with computing energy units likely to be consumed along the railway line for two selected parameters. The deep neural network may comprise an input layer, a first hidden layer, a second hidden layer and an output layer. The input layer comprises three input nodes 302, 304 and 306 that provides the data corresponding to the parameters to the first hidden layer. The first hidden layer comprises four neurons (or computation units) 308, 310, 312 and 314, each of which receives the data corresponding to the selected parameters from the input nodes 302, 304 and 306. The data may be weighted in accordance with the activation function used in the neurons 308, 310, 312 and 314. The neurons 308, 310, 312 and 314 process the data received from the input layer in accordance with the activation function associated with the respective neuron. Similarly, the second hidden layer comprises four neurons 316, 318, 320 and 322, each of are associated with an activation function. The outputs from the first hidden layer may be combined with weights or coefficients before being fed into the second hidden layer. Each of the four neurons 316, 318, 320 and 322 processes outputs from each of the neurons 308, 310, 312 and 314 using the respective activation function. The outputs from the hidden layer are further provided to the output layer. In the present example, the output layer comprises one output node 324. The output node 324 provides the computed value of the energy units likely to be consumed along the railway line based on the data associated with the selected parameters.

The deep neural network is trained by the training module 155 based on the parametrised error functions. For example, assume that the parametrised error function at say, time tl is e(tl)<0.02. The actual output of the deep neural network at tl is compared to an expected output. In the present example, the actual output is the energy units likely to consumed as computed by the deep neural network. The expected output may be obtained from the historic data associated with the energy consumption. If the difference between the actual output and the expected output is greater than 2%, then the weights in the deep neural network are modified such that the error is brought down to a value below 2% at time tl. The weights are modified using a suitable optimization algorithm.In one example, the optimization algorithm is gradient descent method.

Referring to FIG 4, in conjunction with FIGS 1, 2 and 3, a flowchart of a method 400 for managing energy consumption along the railway line is shown, in accordance with one exemplary embodiment of the present invention. The method comprises steps 405 to 420. At step 405, energy units likely to be consumed along the railway line is computed, as explained using FIG 3, based on the data corresponding to the parameters associated using the artificial intelligence model.

At step 410, a load forecast model for the predefined time period is generated based on the computed energy units likely to be consumed along the railway line. The load forecast model is generated as a numerical model of the energy units likely to be consumed at specific points in time within the predefined time period on the GUI 140.

At step 415, an overall impact on the rail infrastructure 107 is determined by analysing the load forecast model. In the present embodiment, the overall impact may be associated with operational efficiency of the rail infrastructure 107.

At step 420, one or more actions are performed in such a manner that the operational efficiency of the rail infrastructure 107 is kept optimal. For example, the one or more actions may be associated with scheduling of an electric train or a fleet of electric trains operating within the railway line. The electric trains may be scheduled based on the technical specifications associated with the electric trains such as number of passengers cars, load and so on, in such a way that the operational efficiency of the rail infrastructure 107 is kept optimal. For example, the energy units likely to be consumed in a line, may be calibrated against number of passengers that are likely to board an electric train that runs through the line. Further, the electric trains may be scheduled to have less number of passenger cars on November 20, 2019, if the load forecast model for November 20, 2019 indicates a decreased energy consumption compared to the energy consumed on November 19, 2019.

In one embodiment, deviations between the energy units likely to be consumed and a predetermined load value, at each point of time within the predefined time period may be computed. Further, one or more actions for reducing the deviations may be performed. The one or more actions may be identified based on rules defined in a database 175. The one or more actions may include generating suggestions for scheduling of the electric trains. The suggestions may be generated based on technical specifications associated with the electric train and also based on the data associated with the parameters.

In one embodiment, the load forecast model may be used for visualising the influence of different parameters, such as number of passengers boarding the train, time of the day, running delays associated with the electric trains and so on, on the load forecast model. The influence of the different parameters may be determined by simulating a simulation instance corresponding to the load forecast model. The simulation instance may be an analytical model in machine-readable format. The simulation instance may be executed using the historic data associated with the energy consumption along the railway line for generating the simulation results. Further, the effect of the different parameters on the load forecast model may be graphically respresented on the GUI 140. The GUI 140 may further enable the operating personnel to visualise changes in the load forecast model for different values of the parameters.

Similarly, for each of the railway lines, a load forecast model may be generated for the predefined time period. In another embodiment, a cumulative load forecast model may be generated for all the railway lines within the rail infrastructure 107 similar to generation of the load forecast model for a single railway line.

Referring to FIG 5, a flowchart of a generalised method 500 for managing energy consumption at a technical installation is shown, in accordance with one embodiment of the present invention. The method 500 includes steps 505-520.

At step 505, energy units likely to be consumed by the technical installation is computed based on data corresponding to one or more parameters using an artificial intelligence model.

At step 510, a load forecast model for a predefined time period is generated based on the computed energy units likely to be consumed by the technical installation.

At step 515, an overall impact on the technical installation is determined by analysing the load forecast model.

At step 520, one or more actions are performed at the technical installation in such a manner that the overall impact on the technical installation is kept optimal.

The present invention may take the form of a computer program product comprising program modules accessible from computer-usable or computer-readable medium storing program code for use by or in connection with one or more computers, processors, or instruction execution system. For the purpose of this description, a computer-usable or computer-readable medium is any apparatus that may contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The medium may be electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (or apparatus or device) or a propagation mediums in and of themselves as signal carriers are not included in the definition of physical computer- readable medium include a semiconductor or solid state memory, magnetic tape, a removable computer diskette, random access memory (RAM), a read only memory (ROM), a rigid magnetic disk and optical disk such as compact disk read-only memory (CD-ROM), compact disk read/write, and DVD. Both processors and program code for implementing each aspect of the technology may be centralized or distributed (or a combination thereof) as known to those skilled in the art.

While the invention has been illustrated and described in detail with the help of a preferred embodiment, the invention is not limited to the disclosed examples. Other variations can be deducted by those skilled in the art without leaving the scope of protection of the claimed invention.

Claims

Patent claims

1.A computer-implemented method for managing energy consumption at a technical installation (107), the method comprising: computing, by a processing unit (130), energy units likely to be consumed by the technical installation (107) based on data corresponding to one or more parameters using an artificial intelligence model (300); generating a load forecast model for a predefined time period based on the computed energy units likely to be consumed by the technical installation (107); determining an overall impact on the technical installation (107) by analysing the load forecast model; and performing one or more actions at the technical installation (107) in such a manner that the overall impact on the technical installation (107) is kept optimal.

2. The method according to claim 1, wherein computing the energy units likely to be consumed by the technical installation (107) based on the data corresponding to one or more parameters using the artificial intelligence model (300) comprises: obtaining the data corresponding to the one or more parameters associated with the energy consumption by the technical installation (107); computing an impact of each of the one or more parameters on the operation of the technical installation (107); dynamically selecting at least one parameter from the one or more parameters based on the impact computed for each of the one or more parameters; and providing the at least one parameter as input to the artificial intelligence model (300) for computing the energy units likely to consumed by the technical installation (107).

3.The method according to claim 1, wherein the artificial intelligence model (300) is selected from a plurality of artificial intelligence models based on a predefined criteria.

4.The method according to claim 1, wherein the artificial intelligence model (300) is trained based on at least one penalty formula associated with the technical installation (107).

5.The method according to claim 1, wherein the one or more parameters are associated with at least one of technical factors, environmental factors, design factors, engineering factors, geopolitical factors and time-related factors impacting the operation of the technical installation (107).

6.The method according to claim 1, wherein generating the load forecast model for the predefined time period based on the computed energy units likely to be consumed by the technical installation (107) comprises: determining at least one mathematical relation between the energy units likely to be consumed by the technical installation (107) and one or more specific points in time within the predefined time period; and generating the load forecast model based on the at least one mathematical relation determined.

7.The method according to claim 6, wherein the load forecast model is at least one of an analytical model, a numerical model and an empirical model indicative of the energy units likely to be consumed by one or more components in the technical installation (107) at the one or more specific points in time within the predefined time period.

8.The method according to claim 1, wherein determining the overall impact on the technical installation (107) by analysing the load forecast model comprises: determining a trend associated with the load forecast model within the predefined time period; and computing the overall impact on the technical installation (107) by analysing the trend associated with the load forecast model.

9.The method according to claim 1, wherein the overall impact on the technical installation (107) is associated with at least one of a financial impact and an operational impact.

10.The method according to claim 1, wherein performing the one or more actions at the technical installation (107) in such a manner that the overall impact on the technical installation (107) is kept optimal comprises: scheduling an operation of the technical installation (107) in order to keep the overall impact on the technical installation (107) optimal.

11.The method according to claim 1, further comprising: computing a deviation between the energy units likely to be consumed and a predetermined load at a specific point in time within the predefined time period.

12.The method according to claim 1, further comprising: generating at least one simulation instance based on the load forecast model; executing the simulation instance in a simulation environment for different values of the one or more parameters; and analysing simulation results generated from the simulation for determining an influence of the one or more parameters on the load forecast model.

13.The method according to claim 1, further comprising: outputting the load forecast model on a Graphical User Interface (140).

14.The method according to claim 1, wherein the technical installation (107) is at least one railway line in a rail infrastructure.

15. An apparatus (110) for managing energy consumption at a technical installation (107), the apparatus (110) comprising: one or more processing units (130); and a memory unit (145) communicatively coupled to the one or more processing units (130), wherein the memory unit (145) comprises one or more modules stored in the form of machine-readable instructions executable by the one or more processing units (130), wherein the one or more modules are configured to perform method steps according to any of the claims 1 to 14.

16. A system (105) for managing energy consumption at a technical installation, the system (105) comprising: one or more devices (115, 120) capable of providing data corresponding to one or more parameters associated with the energy consumption by the technical installation (107); and an apparatus (110) according to claim 15, communicatively coupled to the one or more devices (115, 120), wherein the apparatus is configured for managing energy consumption by the technical installation (107) based on the one or more parameters, according to any of the method claims 1 to

14.

17.A computer-program product having machine-readable instructions stored therein, which when executed by one or more processing units (130), cause the processing units (130) to perform a method according to any of the method claims 1 to 14.