Computer-implemented method for charging vehicle in electric vehicle charging system on the basis of lexicographic optimization in first step and objective function with weighted object in second step, and system therefor

Abstract

To provide a system and a computer-impelemented method for charging an electric vehicle.SOLUTION: In a first step, a first charging schedule is generated by performing lexicographic optimization on the basis of a hierarchy of charging, a charging object that should be considered when generating the first charging schedule for one or more charging stations of an electric vehicle charging system is acquired, and the hierarchy of those charging objects is determined. The charging of the electric vehicles is controlled in accordance with the first charging schedule. In a second step, a second charging schedule is generated by performing an optimization based on an objective function having combined weighted objects. A weight is learned in an offline optimization that is based on recorded data from the first step. In the second step, the charging of the electric vehicles is controlled in accordance with the second charging schedule.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The description herein relates generally to charging electric vehicles, and more particularly to a computer-implemented method, computer program product, and electric vehicle charging system in which a charging schedule for one or more charging stations is generated. [Background technology]

[0002] Unmanaged charging of the growing number of electric vehicles poses significant challenges for the stable operation of the power grid. Smart charging allows for the coordinated control of the charging process and is seen as an important step towards the successful integration of electric vehicles into the power grid. Moreover, it could bring benefits to charging station operators and end users (electric vehicle drivers) compared to unmanaged charging.

[0003] A smart charging system controls at least the charging power at one or more charging stations that charge the batteries of electric vehicles to achieve a particular objective, such as minimizing energy costs and / or reducing peaks in electric load. The component of a smart charging system that utilizes an optimization-based approach to calculate the charging power is commonly referred to as a charging scheduler.

[0004] It is often desirable for a charging scheduler to consider multiple (at least partially) conflicting objectives, such as, for example, minimizing peak load, maximizing the number of vehicles charged, and minimizing battery degradation. Because these objectives conflict, there is no solution (i.e., a charging schedule) that jointly optimizes all of the objectives.

[0005] The most common approach to address this problem is to combine multiple objectives in the form of a weighted sum into one objective function of the optimization problem to be solved, as disclosed in US Pat. No. 8,725,306 (B2).

[0006] However, it is usually difficult to determine the appropriate weights that should reflect the priorities of these objectives. Moreover, in the context of charging management, it is often desirable to define the priorities in the form of a hierarchy of the considered objectives, where the objectives at higher levels of the hierarchy have a clearly higher priority than those at lower levels of the hierarchy. The weighted sum approach requires the adjustment of the weights of the objective function, which makes it difficult to dynamically adapt the hierarchy to changing priorities. US Patent Application Publication No. 2022 / 0327443(A1) proposes to overcome this problem by generating a hierarchy of objectives from multiple basic objectives, and then solving the single-objective problems one by one according to the objectives in the hierarchy. Although such an approach is capable of dynamically adapting the objectives or their priorities by using a hierarchy of multiple objectives, which may be selected from the entire set of available objectives, it is difficult to apply this approach in all situations. The reason is that due to the increasing number of involved charging stations and vehicles to be charged and the growing complexity of the entire charging system, it is time-consuming to generate all charging schedules using lexicographical optimization, which creates latency for updating the applied charging system, which is undesirable. Summary of the Invention [Problem to be solved by the invention]

[0007] It is desirable to overcome the above-mentioned drawbacks and provide an improved method for scheduling the charging of electric vehicles. More specifically, it is desirable to provide a computer-implemented method, a computer program product, and an electric vehicle charging system that allows the generation of a charging schedule with less effort and cost, while at the same time ensuring that the charging schedule can be generated in a short time during normal use of the system. This is achieved by the method, the program, and the electric vehicle charging system according to the attached independent claims.

[0008] The present disclosure provides a computer-implemented method, a computer program product, a scheduling apparatus, and an electric vehicle charging system. [Means for solving the problem]

[0009] In one broad aspect, a computer-implemented method for scheduling charging of an electric vehicle and charging the electric vehicle accordingly with an electric vehicle charging system is provided. The method comprises two stages in which a charging schedule is generated in different ways. In a first stage, the method uses lexicographic optimization of a hierarchically ordered objective. The first stage comprises: - obtaining charging objectives to be considered in generating a first charging schedule for one or more charging stations and a stationary battery of a vehicle charging system; - determining a hierarchy of charging purposes; - generating a first charging schedule by performing lexicographical optimization based on a hierarchy of charging objectives; - charging the electric vehicle by applying a first charging schedule.

[0010] In the second stage, the system does not calculate the first charging schedule, but switches to another principle for determining the charging schedule, which is hereafter referred to as the second charging schedule. In the first stage, a relatively slow but flexible lexicographic approach is used, whereas the second stage applies an objective function with weighted objectives. The weights for such a weighted sum approach are determined in an offline learning process, where the weights are obtained from an optimization of the weights of the weighted sum objective function with respect to the schedule obtained in the lexicographic optimization. This is realized based on the recorded data from the first stage. Then, once the weights are determined, the objective function with the weighted objectives is used to determine the second charging schedule. The learning of the weights can be performed in any known manner suitable for determining an optimized set of weights that results in a second charging schedule that is as close as possible to the first charging schedule obtained in the lexicographic optimization.

[0011] In another broad aspect, a program is provided which, when executed or loaded into a computer, causes the computer to carry out the steps of the above method.

[0012] In another broad aspect, a vehicle charging system for scheduling charging of an electric vehicle and charging the vehicle accordingly is provided. The system comprises a processing means configured to obtain charging objectives to be considered in generating a first charging schedule for one or more charging stations of the charging system, determine a hierarchy of the charging objectives, and generate the first charging schedule by performing a lexicographical optimization based on the hierarchy of the charging objectives. The processing means is further configured to control the one or more charging stations according to the first charging schedule in a first stage. The system further comprises a memory configured to record data in the first stage. The processing means is further configured to learn weights of an objective function with the combined weighted objectives in an offline optimization based on the recorded data from the first stage. Furthermore, the processing means creates a second charging schedule by performing an optimization of the objective function, and then applies the obtained second charging schedule to charge the electric vehicle accordingly.

[0013] In another broad aspect, an electric vehicle charging system is provided that includes a scheduling device, a charging station for charging an electric vehicle, a stationary battery, and a controller for controlling the charging station and the stationary battery charger based on a charging schedule.

[0014] Any of the systems and / or functions described herein may be implemented using individual hardware circuits, using software working with at least one of a programmed microprocessor, a general-purpose computer, using an application specific integrated circuit (ASIC), or using one or more digital signal processors (DSPs). It should be noted that the processing means used to determine the schedule and learn the weights of the objective function preferably comprises a plurality of distributed processors connected to exchange information, as will be explained in more detail in the following description. In the following description, in order to achieve concise terminology, it is mainly referred to as a charging process. However, this does not limit the present invention, and the term charging is used as a synonym for charging and / or discharging.

[0015] Further features and aspects are defined in the dependent claims.

[0016] In the method for scheduling charging of an electric vehicle, charging objectives to be considered when generating a charging schedule for one or more charging stations of a charging system are obtained from the entire set of basic objectives. To adjust the considered objectives to the actual situation, relevant objectives can be selected. For example, some charging objectives are specified from the entire set of basic objectives based on the input of the operator of the charging system. Alternatively, the charging objectives can be automatically determined / updated by analyzing or classifying user settings and user behaviors, and / or by detecting problems in the charging system, such as unbalanced utilization of charging stations, low utilization rate (possibly too expensive for users), high peak load, bad charging stations, high electricity prices, excess self-generated electricity, renewable energy share, etc., and assigning a charging objective to each detected problem based on a table. Then, a hierarchy of charging objectives is determined, and a first charging schedule is generated by performing lexicographical optimization based on the hierarchy of charging objectives.

[0017] In lexicographic optimization, a priority is imposed on which charging objectives to consider by ordering the objective functions according to the importance or significance of the objectives, rather than assigning weights. After the objective functions are ordered by importance, the most important objective is solved first as a single-objective problem, which is defined as: minf 1 (x) (1) x∈X (2)

[0018] y * 1 is the first objective function f 1 The optimal solution for (x), where X is the set of feasible solutions defined by various constraints:

[0019] y * 1 :=min{f 1 (x)Ix∈X} (3)

[0020] The second objective is then re-optimized as a single-objective problem with the added constraint, which is defined as: f 1 (x)≦y * 1 (4)

[0021] Thus, according to the hierarchy of charging objectives, the results of higher priority optimizations form additional constraints for the lower priority single-objective problems that are subsequently solved. This process is repeated for the remaining objectives / subproblems, and in the step of solving the subproblem M, the optimal solution y obtained in the previous step is used as explained above. * M-1 is added as a new constraint. The algorithm terminates after solving the problem according to each charging objective defined in the hierarchy.

[0022] In this way, the operator can easily configure the hierarchy of objectives considered by the charging scheduler. The hierarchy can even be dynamically changed at run-time. This is difficult to achieve with the traditional weighted sum approach, because a set of weights must be predefined for each different configuration, and it is already difficult to determine the appropriate weights even for only one configuration. Furthermore, if the scheduler is to be extended with additional potential objectives, a new set of weights must be determined. As a solution to this problem, the method of the present invention uses a lexicographic optimization that does not rely on weights in the first phase of the system's operation. In this way, a flexible and easily scalable charging scheduler can be achieved. However, it turns out that the adaptation of the hierarchy, or even the consideration of completely new objectives, is not required very often. However, a system that uses only a hierarchical approach suffers from a lack of speed of the optimization process, since multiple optimization problems need to be solved to calculate the charging schedule. Most of the time, the system operates with an unchanging hierarchy of objectives, and therefore flexibility is not needed. Thus, the present invention allows the entire system to operate using a weighted sum approach, where the second charging schedule is determined based on the objective function using a weighted sum of objectives. The optimization routine using such a weighted sum of objectives is much faster, and recalculation of the charging schedule is required at short time intervals, which is a major performance advantage, especially for systems of increasing complexity that are on the rise.

[0023] The invention has the advantage that the weights for determining the objective function used in the second stage are determined offline using information recorded during the first stage, during which the system generates the first charging schedule using a lexicographical optimization based on a hierarchy of objectives. Data is collected (recorded) while operation using such optimization is possible. These data provide information on the inputs for calculating the first schedule using lexicographical optimization. For each first charging schedule generated, data is recorded, defining the charging schedule in relation to the inputs taken into account when generating the schedule (e.g. boundary conditions describing the situation for which the respective schedule is generated). This may be any kind of environmental conditions, such as charging demand from the user, the actual charging state of the connected vehicle, date, time, temperature, etc.

[0024] If the hierarchy does not change over a period of time, an equivalent first charging schedule is generated taking into account a number of different boundary conditions. As described above, these boundary conditions and the respective generated charging schedules are recorded by storing the collected data in a memory.

[0025] Starting from an objective function with weighted objectives, the system then learns the weights of the objective function in an offline process, for example by an optimization process that adjusts the individual weights until a criterion is met. Such optimization routines are known in the art and may be chosen accordingly, for example according to the complexity of the function or the amount of available data.

[0026] Once the offline optimization is completed, the weights are used to define an objective function, on the basis of which a further optimization of the second charging schedule is then performed in a second stage. The optimization in the second stage (and the charging according to the second charging schedule) can be performed using the scheduler that is also used to generate the first charging schedule, but the optimization of the weights is preferably performed by a separate entity. Such distribution of work improves the overall performance, since the scheduler's processor does not have to "waste" processing power by optimizing the weights at the same time as generating the first charging schedule (which is slow anyway). However, in general, the optimization of the weights can be performed at times of low computational load, for example at night. In that case, the system can calculate the weights internally using a single processor without the need to transfer the recorded data. In general, the system comprises a processing means, which may be realized by a single processor or by multiple processors that jointly calculate the weights of the charging schedule and / or the objective function.

[0027] The charging scheduler may specify, for each charging station, at least one of a charging or discharging power, a charging current, a charging curve, and a charging amount in a first phase or a second phase. Preferably, the system is not limited to charging vehicles, and the charging schedule includes charging and discharging a stationary battery, for which the system comprises a stationary battery charger.

[0028] Typically, the energy available for charging is taken from the power grid or from a photovoltaic system that may be present at or on the premises of the charging station. However, the photovoltaic system may be enhanced by incorporating a stationary battery as an energy storage mechanism. Conventionally, the charging of a dedicated photovoltaic energy buffer is only controlled locally by the photovoltaic system itself, without taking into account the requirements of the entire system. Advantageously, according to a preferred embodiment of the present invention, an energy storage mechanism in the form of a stationary battery can be used by the entire charging management system and does not require a separate control independent of the charging system for the vehicle.

[0029] Furthermore, for the charging process, the stationary battery can be treated as one additional battery to be charged / discharged in addition to the charging of the vehicle, so the scheduler can easily incorporate the stationary battery. The necessary adaptation of the optimization process is done by adapting the hierarchy in the first stage and generating new basic objectives if necessary. Specific objectives considering the stationary battery are the target state of charge, the low threshold for discharge, the overall CO2 reduction for the charging / discharging of the connected vehicles and the stationary battery. 2 There may be a low threshold for the increase of emissions. A change in the battery used as a buffer may also be quickly implemented and reflected by the adapted basic objectives and / or hierarchies. This is particularly important when the stationary battery is a regenerative battery that has already accumulated a large number of charge and discharge cycles. According to the actual state of such a stationary battery, the hierarchy of the respective objectives, e.g. protection of the battery by defining upper / lower limits for the state of charge and / or the charging current and / or the discharging current, may be adapted. After adaptation, the system operates in the first stage, and the system may automatically switch to the first stage when a conformity of the objectives and / or hierarchies is detected.

[0030] Stationary batteries can be used to buffer energy and thereby mitigate peaks in energy requirements for the public power grid. Such replenishment can be used to reduce the amount of energy consumed from the power grid. The batteries can then use the stored electrical energy when the cost of energy is low. In contrast, the main source for charging the stationary batteries can be selected, for example only renewable energy is stored in the buffer (stationary batteries). If there is a demand for green (renewable) energy, this can be provided at least partially in situations where the energy produced by photovoltaic systems, etc. is not sufficient. However, the source of energy available for charging the stationary batteries can change dynamically. In winter, it may be reasonable to mainly use the stationary batteries to reduce costs, whereas in summer, the buffering aspect of photovoltaic energy prevails. Such dynamic aspects require adaptation of the optimization process and, according to the invention, can be easily adjusted by the operator by rearranging the objectives in the hierarchy. Finally, the stationary batteries can be used as an emergency power source to ensure the operation of the system.

[0031] Alternatively or additionally, the charging schedule may be generated or regenerated upon request, the request including at least one of information regarding the charging objective and information regarding a hierarchy of charging objectives.

[0032] In the determination step in the first phase, the hierarchy of charging objectives may be set according to a pre-set hierarchy or may be determined by modifying the pre-set hierarchy based on information included in the request. This information may be provided by an operator of the charging system. The objectives included in the hierarchy may be communicated to a processor that determines weights of the objective function, so that the objective of the objective function always corresponds to the objective of the hierarchy. Alternatively, the objective function always includes all basic objectives.

[0033] The information regarding the charging purpose may indicate the aspect most frequently desired by users, and the information regarding the hierarchy may indicate the operator's individual priorities to be taken into account in the operator's charging system.

[0034] Typically, the lexicographical optimization in the first stage takes longer than the weighted sum approach, and the method can determine whether the constructed objective hierarchy can be adapted to speed up the optimization without affecting the optimization result, and automatically adapt the hierarchy if applicable. To this end, the method includes: - reducing the number of charging objectives by determining and deleting at least one of a duplicate charging objective, an inapplicable charging objective, and an automatically satisfied charging objective from the obtained charging objectives, where the number of charging objectives after reduction corresponds to the number of optimization problems to be solved in the lexicographic optimization; and - reducing the number of optimization problems by combining at least two non-conflicting charging objectives into one common optimization problem; It may further include at least one of:

[0035] An electric vehicle charging system for charging an electric vehicle comprises a processing means for obtaining charging objectives to be considered in generating a first charging schedule for one or more charging stations and possibly a stationary battery charger of the charging system, a means for determining a hierarchy of the charging objectives, and a means for generating the first charging schedule by performing lexicographical optimization based on the hierarchy of the charging objectives. The processing means further generates a second charging schedule in a second stage. The processing means provides the first charging schedule in the first stage and provides the second charging schedule in the second stage to enable the one or more charging stations and possibly the stationary battery charger to be controlled according to the first or second charging schedule.

[0036] The electric vehicle charging system may further comprise a scheduling device having a processor implementing at least a portion of the processing means, a charging station for charging the electric vehicle, a stationary battery and its stationary battery charger, and a controller for controlling the charging station and possibly the stationary battery charger based on the first or second charging schedule provided by the scheduling device.

[0037] In addition, the control device may be configured to send a request to the scheduling device to generate a charging schedule, the request including at least one of information regarding the charging objective to be taken into account and information regarding the hierarchy, and the scheduling device configured to determine the hierarchy of the charging objective based on the information included in the request.

[0038] Additionally, the control device may be configured to determine at least one charging purpose selected by an operator of the system in the first phase and to send a request including information on the at least one selected charging purpose to the scheduling device, which is configured to modify the pre-set hierarchy of charging purposes based on the at least one selected charging purpose to determine the hierarchy. Preferably, the system reacts to any change in the hierarchy, whether a reordering or addition or removal of purposes or a restart of the first phase. The transition to the second phase may be automatically triggered when sufficient data has been recorded, e.g. when a predefined number of first charging schedules have been generated.

[0039] Alternatively or additionally, the control device may be configured to determine, in a first stage, a hierarchical level for charging purposes selected by the operator and to send a request including information regarding the selected hierarchical level to the scheduling device, and the scheduling device determines the hierarchical level according to the selected hierarchical level.

[0040] The scheduling device and the control device according to the present disclosure each include a processing unit configured to perform the steps described above. The processing unit may be a controller, a microcontroller, a processor, a microprocessor, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or any combination thereof.

[0041] Embodiments of the method and system of the present invention will now be described in conjunction with the accompanying drawings. [Brief description of the drawings]

[0042] [Figure 1] FIG. 1 is a schematic diagram of a system according to one embodiment of the present disclosure. [Diagram 2] FIG. 2 illustrates the mapping of base objectives to an objective hierarchical structure. [Diagram 3] FIG. 2 is a block diagram of a scheduling device. [Figure 4] 1 is a flowchart of a method according to one embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0043] In the figures, the same reference numbers refer to the same or equivalent structures. Descriptions of structures with the same reference numbers that are in different figures have been avoided for brevity where possible.

[0044] FIG. 1 shows a block diagram of an electric vehicle charging system. The electric vehicle charging system is configured to charge an electric vehicle EV 1 , E.V. 2 ...EV N Multiple charging stations for charging batteries CS 1 , C.S. 2 , ..., C.S. N a charging scheduler CS for generating a first and a second charging schedule; and a charging station CS for controlling the charging station CS based on the first or second charging schedule. 1 , C.S. 2 , ..., C.S. NThe charging station CS is provided with a charging management system CMS, which is a control unit that controls the charging scheduler CS (scheduling device). The charging scheduler CS (scheduling device) can be incorporated into the charging management system CMS (control device). 1 , C.S. 2 , ..., C.S. N can be, for example, a public charging station or a charging station on the premises of a company for charging employees' EVs. Furthermore, the charging management system CMS is connected to a stationary battery BAT, which may be equipped to buffer the electric energy, or to a stationary battery charger BC. The stationary battery charger BC is controlled by the charging management system CMS, which controls the charging of the battery via the stationary battery charger BC or the charging station CS. 1 , C.S. 2 ,....C.S. N controls energy recovery to discharge a stationary battery so that the energy is used to charge the vehicle.

[0045] The charging management system CMS is responsible for at least the charging station CS 1 , C.S. 2 , ..., C.S. N and a stationary battery charger BC. 1 , C.S. 2 , ..., C.S. N and repeatedly transmits control signals to the stationary battery charger BC. 1 , C.S. 2 , ..., C.S. N The stationary battery BAT is connected to the charging station CS 1 , C.S. 2 , ..., C.S. N (only), connected electric vehicle EV 2 ...EV N Maximum and minimum charging power for connected electric vehicles (EVs) 2 ...EV N, and the battery level of the stationary battery BAT to the charging management system CMS. In addition, the charging management system CMS may obtain information such as the departure time or the desired charging state from the driver of the connected EV (EV driver information). 1 , C.S. 2 , ..., C.S. N All or part of may further be connected to locations where energy consumers and / or generators are located, for example, corporate buildings with a certain basic consumption and a photovoltaic system. Additionally or alternatively, a photovoltaic system may be provided that is dedicated to the charging management system CMS. In order to suitably adapt the charging schedule taking into account the energy provided by the photovoltaic system, the charging management system CMS receives information on the estimated energy provided by the photovoltaic system (PV forecast). The charging management system CMS may also take into account information from local locations for charging management. Furthermore, the charging management system CMS receives external information such as electricity prices, and driver information such as driver / vehicle ID, desired charging state, expected arrival time, desired charging time / power and battery condition (charge state, temperature). Furthermore, the charging management system CMS receives information on the energy mix, which defines the main sources (e.g., wind energy, nuclear energy, ...) of electrical energy that can be supplied to the charging station and that can be used to charge the vehicle and the stationary battery BAT. The system can also use locally available photovoltaic energy for individual charging stations at locations with individual photovoltaic systems or photovoltaic systems that generate electrical energy exclusively for the charging management system. For the photovoltaic systems, the charging management system CMS receives a forecast of the expected amount of energy over time.

[0046] Each charging station CS is assigned to provide an adequate charging power for which a specific objective is achieved, such as minimizing energy costs, reducing peaks in the electric load, minimizing waiting time for vehicle users, or maximizing the share of regenerative energy used to charge the vehicle. 1 , C.S.2 , ..., C.S. N The charging power provided by is pre-planned by the charging scheduler CS using the optimization-based approach of the present invention.

[0047] In addition, the charging power for charging the stationary battery BAT or the discharging power for discharging the stationary battery to contribute to the charging of the vehicle is planned by the charging scheduler CS. The inclusion of schedules for charging and discharging the stationary battery also increases the flexibility of the system and therefore allows the solution of optimization problems that cannot be achieved without such a stationary battery BAT. For example, in case of a decrease in the share of renewable energies in the energy mix transferred to the charging management system CMS, the charging station CS can be powered by using energy taken from the stationary battery BAT. 1 , C.S. 2 ,...CS N The stationary battery BAT can supplement the electric energy supplied to the vehicle. In contrast, in cases where an exceptionally large amount of renewable energy is available but only a low demand from users needs to be met, the stationary battery can be charged. The stationary battery can then supplement the renewable energy at a later time, when the share of renewable energy in the energy mix drops. The charging / discharging schedule of the stationary battery BAT is determined at the same time and in the same optimization process, so that separate schedules for the battery and the vehicle, which would reduce the overall performance, can be avoided.

[0048] The charging management system CMS sends a schedule request to the charging scheduler CS, which holds information about the calculation of the charging schedule. The charging scheduler CS then calculates the charging schedule for N charging stations CS for T time steps of length Δt in the future. 1 , C.S. 2 , ..., C.S. N and the schedule of charging power P of the stationary battery charger BC.

[0049]

number

[0050] The first charging schedule is calculated by formulating and solving an optimization problem of the form:

[0051]

number

[0052]

number

[0053] The first objective function f 1 is the electricity price per unit of energy at time step t t The second objective function f 2 is the number of electric vehicles (EVs) at the end of the planning period. 1 , E.V. 2...EV N The constraint in (9) is that the sum of the energy levels of electric vehicles n (charging stations CS n ) is connected to a certain maximum power

[0054]

number

[0055]

number

[0056] [Number] , in the following similar cases, specific values must be satisfied for variables such as N, T, Δt, c t , etc. Some of these values (for example, the number of time steps T planned in the future) may be configuration parameters of the charging scheduler CS. The remaining values must be provided according to the scheduling requirements of the charge management system CMS.

[0057] In this exemplary problem, the two objectives are conflicting (assuming the electricity price c t is greater than zero). This is a common situation. Some hierarchy of objectives should be considered, and it is assumed that the objectives at the higher level of this hierarchy are strictly prioritized higher than the objectives at the lower level. Without loss of generality, when i < j, the objective function f i is assumed to be at a higher level than the objective function f j . A feasible solution S * to this optimization problem is optimal with respect to a given hierarchy if this solution optimizes the highest-level objective function f 1 and there is no other solution that improves one of the objectives at the lower level (f * , i > 1) without degrading one of the objectives at the higher level (f j , j < i) compared to S i .

[0058] The electric vehicle charging system makes it possible to construct the objective hierarchy considered here to take into account the various priorities of various operators. This provides a set of basic objectives G = {g 1 ,..., g k} and a hierarchy F = (f 1,...,f M )(f for all i=1,...,M i f for ∈G and i ≠ j i ≠f j This hierarchy determines the order in which single-objective problems are solved for multiple charging objectives selected from the entire set of basic objectives. This is the order in which the entire set of basic objectives g 1 ,g 2 ,g 3 ,...g K Multiple charging purposes selected from f 1 ,f 2 ,...f M As shown in Figure 2, the operator can achieve the basic objective g 1 ,...,g k It is possible to determine which of the basic objectives are taken into account in the scheduling of charging and how these objectives are prioritized relative to each other. The overall basic objectives are defined at the design stage of the scheduler. However, since charging objectives can be selected from the basic objectives and ordered by a hierarchy, the present invention allows a very flexible adaptation of the scheduling according to the needs of the operator of the system. Since the stationary battery BAT is included in the scheduling process, this includes not only the charging of the vehicle battery, but also the charging / discharging of the stationary battery that can be used as a buffer. In order to allow the operator to adjust the hierarchy.

[0059] A monitoring system MS is connected to the charging management system CMS. Information about the first charging schedules and the boundary conditions and demands resulting from the respective first charging schedules are recorded using the memory of the monitoring system MS, but feedback from the vehicle drivers may also be recorded in the monitoring system MS. In addition, malfunctions of the charging stations can be communicated to the monitoring system MS, and the operator can then react to the information and adjust the hierarchy accordingly. However, the main function of the monitoring system MS is to allow the determination offline of an objective function with weighted objectives that assumes comparable boundary conditions and leads to a result similar to the lexicographical optimization. The process of determining the weights of such a weighted sum approach used to determine the second charging schedule in the second stage is described below.

[0060] The lexicographic optimization described above is used in the first stage to solve the charging scheduling problem for a given objective hierarchy configuration and provide a first charging schedule accordingly. Thus, it is possible to address possible configurations of charging objectives within the constraints posed by the available basic objectives. Using the lexicographic optimization technique, a set of M subproblems are solved, where the most highly prioritized objective function f 1 The subproblem that considers only and the original constraints is solved first:

[0061]

number

[0062] y 1 Let be the solution of this subproblem, i.e., the optimal objective value. The second subproblem is the second most important objective function f 2 We consider only the subproblems, and guarantee through additional constraints that the solution is optimal in the sense of the preceding subproblems:

[0063]

number

[0064] Similarly, the solution y of the second subproblem is 2 Using the objective function f 3 We construct further constraints for the third subproblem that considers only the objective f M This continues similarly up to the Mth problem, which only considers the Mth problem. This approach does not require specifying individual objective weights for each possible configuration of objectives, and allows for a system that generates and applies charging schedules quickly after the system configuration has changed or after different priorities have been given to specific objectives. Changing the overall basic objectives and / or the charging objectives selected from the objective hierarchy F only changes the objectives considered and their order in which the subproblems are solved. Although adding new basic objectives requires a redesign of the scheduler, compared to the all-weighted-sum approach, this is a much simpler approach since it requires the addition of additional basic objectives g K+1 It also makes it easy to extend the charging scheduler by providing

[0065] The considered objective hierarchy may be set by the operator in the form of a list in a configuration file. The configuration file is stored in the charging scheduler CS or transmitted with the request. Furthermore, it is also possible for the charging management system CMS to dynamically specify the objective hierarchy so that it can be considered as part of the schedule request. In addition, the charging management system CMS can automatically adapt the objective hierarchy to changing conditions and / or operator settings, where a particular setting type or frequency of those settings is assigned to a particular priority / ranking of the objectives (single-objective problem). Alternatively or additionally, the charging management system CMS can continuously estimate one or more conditions, e.g., utilization rate, peak load, number of charging stations with faults, electricity price, surplus electricity, etc., compare them with corresponding pre-set thresholds, and increase (or decrease) the priority / ranking of the objective assigned to that condition if the threshold is reached. Thus, for the scheduling problems (7)-(11), it is possible that minimizing the electricity price is automatically changed from the most important objective to the second objective when the electricity price falls below a certain value, and maximizing the sum of the energy level is automatically changed from the second objective to the most important objective. Conversely, if the photovoltaic forecast indicates a decrease in available energy, the objective regarding the stationary battery target state of charge may be moved to a lower priority or even cancelled from the list.

[0066] Typically, a larger number of objectives leads to a longer execution time, since more subproblems need to be solved. To counter this undesirable effect, it is appropriate to reduce the number of objectives in the hierarchy before the actual optimization, if this does not affect the optimization result. For example, if two objectives f i and f i+1 If the objectives f are not in conflict, they can be optimized simultaneously in one common subproblem, which is the sum of these objectives f i +f i+1 Therefore, the objective hierarchy is automatically F=(f 1 ,...,f i-1 ,f i +fi+1 , f i+2 ,..., f M can be changed to

[0067] Furthermore, since another objective f at a higher level of the hierarchy is already implied by j , the objective f i (j < i) may be redundant (for example, the objective of charging an EV as fast as possible typically implies the objective of charging as many EVs as possible), the objective f i can be automatically removed from the hierarchy. Similarly, if an objective is not applicable or is automatically satisfied for each constraint respectively (for example, the objective of minimizing discharge is automatically satisfied if the minimum charging power of all EVs is 0), this objective can be removed from the hierarchy.

[0068] The monitoring system MS has an interface for connecting it to the charging management system CMS. This interface enables the transfer of data to record the data in the memory included in the monitoring system MS. The monitoring system, like the charging scheduler CS, has a processor connected to the memory and the interface. Based on the data recorded in the memory, the processor calculates weights that enable it to generate an objective function with weighted objectives that yields the same or a sufficiently close result as the first-stage lexicographic optimization.

[0069] Based on the recorded data, optimization is performed to fit the weights of the objective function so that the simulation results using the recorded boundary conditions yield a second charging schedule that is the same or close to the first charging schedule. When an end condition is satisfied, such as the remaining difference between the first charging schedule and the second charging schedule being below a given threshold, the system switches to the second-stage operation.

[0070] The determined weights are sent to the charging scheduler CS via the charging management system CMS and the interface. In the charging scheduler CS, an objective function is created based on the objectives of the hierarchy used in the first stage. Alternatively, the objective function may be generated in the monitoring system MS and transferred to the charging scheduler CS.

[0071] Once the objective function with the weighted objectives is available to the charging scheduler CS, a second charging schedule is generated by the charging scheduler CS, which is then sent, like the first charging schedule of the first phase, to a control unit, the charging management system CMS, which then controls the charging (and possibly discharging) of the electric vehicle and, if included in the system, the stationary battery BAT.

[0072] FIG. 3 shows a block diagram of the charging scheduler CS and the monitoring system MS. The charging scheduler CS and the monitoring system MS each comprise an interface 1 through which the charging scheduler CS and the monitoring system MS are connected to the charging management system CMS, a processor 2 configured to generate the charging schedule or determine the weights, and a memory for storing data, e.g., information on the objectives and hierarchies, or recorded data. The generation of the first or second charging schedule by the processor 2 of the charging scheduler can be implemented as software in C / C++. A suitable solver and a corresponding API (Application Programming Interface) can be used to construct and solve the subproblems of lexicographic optimization according to the configured objective hierarchy. In the case of only linear objectives and constraints, the SCIP (Solving Constraint Integer Programs) solver disclosed in A. Gleixner, et al.: “The SCIP optimization suite 5.0” Tech. Rep. 17-61, ZIB, Takustr.7, 14195 Berlin, 2017 can be used. The interface 1 between the CMS can be realized based on HTTP(S) (Hypertext Transfer Protocol (Secure)) and REST (Representational State Transfer) protocols. In a second stage, the optimization of the charging schedule is performed as known in the prior art for an objective function with weighted objectives. It should be noted that the invention does not concern a specific manner of implementing the generation of the charging schedule based on the weighted sum approach, but rather the determination of weights so that, based on the first charging schedule generated in the first stage using lexicographical optimization, a second charging schedule can then be generated using the commonly known weighted sum approach. The determination of the weights can be realized offline by a separate processor, so as not to degrade the performance and functionality of the system already operating in a mode according to the first stage.

[0073] It can be expected that not all operators can determine the objective hierarchy, since this requires a basic understanding of how different hierarchies affect the scheduling result. Thus, if a request does not specify the charging objectives and their hierarchies to be used, the charging management system CMS or the charging scheduler CS can provide a suitable default hierarchy. Configurations selected by different operators can be collected in a central server to refine a default set of charging objectives to use from the basic objectives and charging objective hierarchies. From this information, a default hierarchy suitable for most users / operators can be derived. This can be, for example, the hierarchies most frequently selected by the users / operators. Another option can be to define a distance measure in the objective hierarchy and set the default hierarchies to the hierarchies that minimize the average distance to the hierarchies selected by the users / operators.

[0074] FIG. 4 shows a simplified flow chart of one embodiment of the method according to the present invention. In step S1, a basic objective g 1 ,...,g k is provided to the charging management system CMS by the scheduler. 1 ,...f MThe selection is made in step S2 based on the respective definitions in the request received from the charging management system CMS via the interface 1 in step S2, and a charging objective hierarchy corresponding to the selected objectives is generated from the configuration or based on the request received from the charging management system CMS. This hierarchy may be adapted by combining and / or deleting charging objectives as explained above. In step S3, the processor 2 generates a first charging schedule for a time interval by performing a lexicographic optimization as explained above. The first charging schedule thus obtained is forwarded to the charging management system CMS, and based on this first charging schedule, the charging station, including the stationary battery charger BC if included in the system, is controlled in step S4. The charging management system CMS monitors the current state with respect to conditions that trigger a new request to be sent to the scheduler. Steps S1 to S3 are executed again if a (new) request is received from the charging management system CMS. A new request may be sent from the charging management system CMS at the end of a time interval and / or when a valid change in the state of the charging station is recognized, e.g. a change in the number of vehicles connected to the charging station. Of course, other conditions that require the adaptation of the charging schedule may be defined, causing the charging management system CMS to send a request to the scheduler. 1 ,...f M can be automatically adjusted to changing conditions, and at least one condition of the charging management system CMS is automatically adjusted to the corresponding basic objective g shown in FIG. 1 ,...,g k , and at least one state or condition change occurs at a particular position f in the hierarchy as shown in FIG. 1 ,...,or f MThe charging management system CMS determines at least one state / change by continuously evaluating at least one condition and, when the state / change is determined, adapts the information about the charging purpose(s) selected from now on and their position in the hierarchy accordingly, if necessary. Alternatively, a pattern / set of states / changes of one or more conditions can be assigned to a particular hierarchy that is set when the pattern / set is determined.

[0075] While the system charges the electric vehicle according to the first charging schedule in the first phase, in step S5, data including information about the first charging schedule and the conditions under which the first charging schedule was generated is recorded. These data are used to determine objective weights for creating an objective function with weighted objectives that is used to determine a second charging schedule in the second phase.

[0076] The weights are determined while the system is still in the first stage so that operation of the electric vehicle charging system proceeds using lexicographic optimization. After the weights are determined in step S6, the weights or the entire objective function are transferred to the charging scheduler CS in step S7.

[0077] The charging scheduler CS then starts generating a second charging schedule based on the objective function with the weighted objectives in step S8, which is the transition of the system to the second stage. In the second stage, the charging management system CMS controls the charging of the electric vehicle according to the second charging schedule, which is generated iteratively in the same way as in the first stage, but at a higher repetition rate because the optimization routine using the weighted sum approach is faster than the lexicographic optimization. While the system is operating in the second stage, possible changes in the objectives, or in the case of no change in the selection of objectives, in the hierarchy, are monitored. This ensures that the system automatically switches to the first stage when the objectives change or when the hierarchy is adapted. This may be the case, for example, when a regenerative battery is replaced as the stationary battery BAT, thereby requiring the priorities of the objectives defining the protection of the battery to be adapted. The system automatically switches to the first stage and generates the first charging schedule again using the adapted hierarchy. At the same time, the system starts recording data again, based on which new weights can be determined to create a new objective function.

[0078] The determination of the weights can begin automatically as soon as a predetermined amount of data has been collected by recording the data obtained in the first stage.

[0079] It will be apparent to those skilled in the art that various modifications and variations can be made to the structure of the present disclosure without departing from the scope or spirit of the disclosure. In view of the foregoing, it is intended that the present disclosure covers modifications and variations of this disclosure provided they fall within the scope of the following claims and their equivalents.

Claims

1. 1. A computer-implemented method for charging an electric vehicle with an electric vehicle charging system, comprising: In a first stage, generating a first charging schedule by performing lexicographic optimization based on a hierarchy of charging objectives, obtaining the charging objectives to be considered in generating the first charging schedule for one or more charging stations of the electric vehicle charging system; and determining a hierarchy of the charging objectives; controlling charging of the electric vehicle according to the first charging schedule; In a second stage, generating a second charging schedule by optimizing based on an objective function having the combined weighted objectives, the weights are learned by offline optimization based on the recorded data from the first stage; and controlling charging of the electric vehicle according to the second charging schedule.

2. The computer-implemented method of claim 1 , wherein the first charging schedule and the second charging schedule specify at least one of a charging or discharging power, a charging current, a charging curve, and a charging amount for each charging station.

3. 3. The computer-implemented method of claim 1 or 2, wherein the method automatically switches to the first stage upon request or when a desired new tier is determined, and learns adapted weights for generating a respective new second charging schedule based on recorded data recorded after a respective new first charging schedule is generated and controlling charging of the vehicle based on the new first charging schedule is started.

4. The computer-implemented method of claim 1 , wherein weight learning is initiated when an amount of the first charging schedule generated in the first stage exceeds a predetermined threshold.

5. The computer-implemented method of claim 1 , wherein the recorded data includes the first charging schedule and system state information.

6. 2. The computer-implemented method of claim 1, wherein generating a charging schedule and charging according to the schedule comprises charging at least one stationary battery in the first phase and in the second phase.

7. A program that, when executed or loaded into a computer, causes the computer to execute the steps of the computer-implemented method according to claim 1.

8. 1. An electric vehicle charging system for scheduling charging of an electric vehicle, comprising: A processing means, obtaining charging objectives to be considered in generating a first charging schedule for one or more charging stations of the electric vehicle charging system; determining a hierarchy of the charging purpose; generating the first charging schedule by performing a lexicographical optimization based on the hierarchy of the charging objectives; providing a first charging schedule for controlling the one or more charging stations according to the first charging schedule during a first phase; A processing means configured to a memory configured to record data during the first stage; The processing means: learning weights of an objective function with the combined weighted objectives in an offline optimization based on the recorded data from the first stage; generating a second charging schedule by optimizing the objective function; providing a second charging schedule for controlling the one or more charging stations according to the second charging schedule during a second phase; The electric vehicle charging system according to claim 1, further comprising:

9. 9. The electric vehicle charging system of claim 8, comprising: a scheduling device having a processor implementing at least a portion of the processing means, the scheduling device being configured to obtain the charging intent, determine the hierarchy, and generate the first charging schedule and the second charging schedule; and a control device that controls the charging station according to the first charging schedule or the second charging schedule.

10. 10. The electric vehicle charging system of claim 9, wherein the system comprises at least one charging station for charging an electric vehicle, a stationary battery, and a stationary battery charger, and the controller controlling the charging station is further configured to control the stationary battery charger based on the first or second charging schedule.

11. the control device is configured to send a request to the scheduling device to generate the first charging schedule; The request includes at least one of information regarding the charging purpose and information regarding the tier, 11. The electric vehicle charging system of claim 9 or 10, wherein the processor is configured to determine the hierarchy of the charging intent based on the information included in the request.

12. the control device is configured to determine at least one charging purpose selected by a user and to transmit the request including information regarding the at least one selected charging purpose; 12. The electric vehicle charging system of claim 11, wherein the processor is configured to determine a hierarchy by modifying a pre-established hierarchy of charging intents based on the at least one selected charging intent.

13. The control device is configured to determine a tier selected by a user for the charging purpose and to transmit the request including information regarding the selected tier; The electric vehicle charging system of claim 11 , wherein the processor is configured to determine the tier according to the selected tier.

14. 14. An electric vehicle charging system according to claim 9, wherein the scheduling device is configured to provide information about the objectives included in the determined hierarchy to an objective function determination unit, the objective function determination unit being configured to learn the weights and to transmit the objective function together with weighted objectives or the determined weights to the scheduling device.

Images