ARRANGEMENT FOR OPERATING A TECHNICAL INSTALLATION

Abstract

The invention provides for an arrangement for operating a technical installation comprising a number of machines, the arrangement comprises a control structure, the control structure being adapted for generating a schedule for sending requests to the machines by means of a simulation and in consideration of expectation values of the machines, for monitoring the operation achieved by switching of the machines, wherein switching of the machines is performed by the machines themselves, in order to achieve a template illustrating the switching of the machines, and for comparing the generated schedule with the template.

Claims

1. Arrangement for operating a technical installation comprising a number of machines, the arrangement comprising: a control structure; the control structure being adapted for generating a schedule for sending requests to the machines by means of a simulation and in consideration of expectation values of the machines, for monitoring the operation achieved by switching of the machines, wherein switching of the machines is performed by the machines themselves, in order to achieve a template illustrating the switching of the machines, and for comparing the generated schedule with the template.

2. Arrangement according to claim 1, adapted for operating a smart grid, wherein the control structure is adapted for monitoring the energy flow achieved by switching of the machines.

3. Arrangement according to claim 1, wherein the control structure is adapted for providing a rank array considering the expectation values of the machines used for generating the schedule.

4. Arrangement according to claim 1, wherein the control is performed with help of a fuzzy logic using the expectation values as input values.

5. Arrangement according to claim 1, comprising a number of agents associated to the number of machines and in connection with the control structure.

6. Arrangement according to claim 1, wherein the control is implemented within one control center forming a central control structure.

7. Arrangement according to claim 1, wherein the control structure comprises a scheduler for generating the schedule.

8. Arrangement according to claim 1, wherein the control center comprises a realization unit for sending requests to the machines, and for monitoring the operation achieved by switching of the machines in order to achieve a template illustrating the switching of the machines, and for comparing the generated schedule with the template.

9. Arrangement according to claim 1, wherein the simulation is performed on a rate basis.

10. Method for operating a technical installation according to claim 9 using an arrangement according to claim 1, comprising the steps of: generating a schedule for sending requests to the machines by means of a simulation and in consideration of expectation values of the machines, and monitoring the energy flow achieved by switching of the machines in order to achieve a template illustrating the switching of the machines, and for comparing the generated schedule with the template.

11. Method according to claim 10, which is adapted for operating a smart grid.

12. Method according to claim 11, comprising the step of providing a rank array considering the expectation values of the machines.

13. Method according to claim 11, wherein an expectation function is generated for each machine.

14. Method according to claim 11, wherein the simulation is performed on a rate basis.

15. Method according to claim 11, wherein the simulation is performed with help of patterns each pattern representing a protocol comprising an array of requests.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0098] In the drawings,

[0099] FIG. 1 is a schematic total view of one embodiment of the method as described herein,

[0100] FIG. 2 is a schedule generated by the scheduler and published by the realization unit as an output,

[0101] FIG. 3 is a schedule as received by the realization unit,

[0102] FIG. 4 is a chart showing generation of a rank array,

[0103] FIG. 5 is a chart showing generating of expectation value and possible course of said value,

[0104] FIG. 6 is an embodiment of the smart grid according to the invention,

[0105] FIG. 7 is a possible simulation proceeding with help of a flow chart,

[0106] FIG. 8 is a flow chart illustrating the simulation,

[0107] FIG. 9 is a flow chart illustrating an operation of a scheduler,

[0108] FIG. 10 is a flow chart illustrating an embodiment of the proposed method,

[0109] FIG. 11 is a flow chart illustrating generation of a rank array,

[0110] FIG. 12 is a schematic view of an embodiment of the arrangement.

DETAILED DESCRIPTION

[0111] The figures are described cohesively and in overlapping fashion, the same reference numerals denoting identical parts.

[0112] FIG. 1 shows a schematic total view of one embodiment of the method as proposed illustrating a circuit 10 demonstrating different steps of the embodiment of the method according to the invention. The steps use external demands 12 as a demand map representing energy demands of the smart grid and internal demands as the expectation functions 14 of the machines. Thus, external and internal demands are considered when computing and generating the schedule. Furthermore, historical data of the machines and current information can be used.

[0113] In a first step 16, a rank array is generated, e.g. by the scheduler, considering the expectation functions 14. This rank array expresses the internal demands, i. e. the wants and demands of the machines.

[0114] In a second step 18, a scheduler generates a schedule using internal and external information. To verify the schedule a simulation is performed to ensure that the external demands can be fulfilled, at least a portion or rate, e.g. 80% of the demand, can be satisfied.

[0115] In a third step 20, a realization unit sends out requests to the machines based on the schedule. The machines receiving a request can follow this request or not. Therefore, the machines can switch on or switch off, e.g. on a minute-by-minute basis.

[0116] In a fourth step 22, the realization unit monitors the smart grid and the machines to record switching on and switching off performed by the machines. Thus, it is possible to generate a actual schedule which can be compared with the schedule generated in step 20 obtaining a deviation 24. The result can be compared 26 with the expectation values 14. Furthermore, it is possible to predict this A 24 by considering the expectation functions.

[0117] Finally, step 22 helps to generate and update historical data 28 which can be used in steps 16 and/or 18.

[0118] The steps above are merely numbered for illustration purposes. However, the method proposed can start at any point on the circle. As an adaptive method, the method can start at step 18 for example using default values for the rank array for generating the schedule.

[0119] FIG. 2 shows a schedule 30 generated by the scheduler and published by the realization unit as basis for the requests sent to the machines. The schedule shows the requests for ten machines 40, 42, 44, 46, 48, 50, 52, 54, 56, and 58 for eight time periods 60, 62, 64, 66, 68, 70, 72, and 74. A cross x represents the request for switching on, a circle represents a requests for switching off. These requests are sent to the respective machines. The time period can be one minute. Different time periods are possible. The requests can be sent to boxes associated to the machines and forming interfaces.

[0120] FIG. 3 shows an actual schedule 80 as received by the realization unit, showing differences to schedule 30 in FIG. 2. Furthermore, the realization unit can receive data relating to the amount of power or energy which has been provided or consumed. A comparison performed by the realization unit shows that second machine 42 in third time period 64, third machine 44 in second and third time period 62 and 64, fourth machine 46 in first and second and fifth time period 60 and 62 and 68 and fifth machine 48 in all time periods do not follow the requests. In that case the realization unit attempts to compensate this deviation by sending new requests to machines considering the rank array and the expectation values. It might be necessary to generate a new schedule. The actual schedule 80 shown can further provide information to the amount of power of energy provided or consumes.

[0121] FIG. 4 shows generation of a rank array using historical data 100, e.g. comprising former rank arrays and data relating to former behavior and reliability of generators. Furthermore, internal data 102, e.g. expectation functions 102 representing the wants and demands of the machines, is used to compute or generate a rank array 104 forming the basis for a schedule generated by the scheduler. This rank array 104 is considered when generating a schedule 106, particularly, a schedule 106 of requests.

[0122] FIG. 5 shows generating of expectation values and a possible course of said values as illustration of an expectation function. Input values 120 are data relating to the type of generator and data relating to environmental conditions, e.g. temperature.

[0123] A graph 130 shows the time dependent course 132 of the expectation value illustrating an expectation function. Values can be between 1 and +1, for example.

[0124] FIG. 6 shows an embodiment of the smart grid according to the invention overall denoted with reference number 150. The smart grid 150 comprises a number of machines 152 connected to a central control 154 having a server 156, a layer for the scheduler 158, a layer for the realization unit 160, a layer for the agents 162, and a recorder layer 164 recording the behavior of the machines 152, i.e. the switching of the machines 152. Thus, FIG. 6 shows a centralized control structure being implemented in the central control 154.

[0125] FIG. 7 is a flow chart illustrating an embodiment of the simulation performed for generating the schedule by the control structure, e.g. the scheduler.

[0126] In a first step 180, a number of patterns is generated. Each pattern is an operating schedule for operating the smart grid for the considered time period divided in time steps. Each pattern can be a realistic operating schedule considering runtime requirements of the machines provided by the agents, for example.

[0127] In a second step, the patterns are evaluated successively until there is a pattern fulfilling the demands. This pattern is selected in a final step 184 and forms the basis for a schedule comprising the requests sent to the machines.

[0128] During simulation, it is to be evaluated whether the demands for energy consumption and/or energy production can be fulfilled. The demands and a demand template, respectively, can be determined with help of historical operating schedules and current information. The simulation using the rank array checks the ability of fulfilling the demands.

[0129] Furthermore, the reliability of the machines can be considered. This reliability can already be considered in the prognosticated expectation functions. Moreover, the reliability, i. e. the probability that the machine will follow, can be considered by examining the behavior of the machine in former time periods and/or in comparing historical prognosticated expectation functions with historical expectation functions. Additionally, external circumstances can be considered.

[0130] FIG. 8 illustrates a possible simulation proceedings. In a first step 220, a number of patterns is generated. Each pattern represents a possible schedule of requests for running the smart grid. To reduce numbers of patterns external demands 222, e.g. demands regarding the machines, the circumstances and/or the demands relating to the energy to be provided or consumed, and a rank array 224 are considered. This rank array 224 can be a two-dimensional or a three-dimensional rank array. A two-dimensional rank array corresponds to a rank layer comprising ranks for the machines at the time steps. A three-dimensional rank array comprises a number of rank layers.

[0131] In a next step 230, the patterns generated in step 220 are evaluated considering external demands 232 like demands relating to the energy to be provided or consumed. A possible external demand could be provide as much energy as possible. Furthermore, historical data 234 is considered. Historical data 234 can be related to the reliability of the machines, historical data to circumstances etc. The patterns are evaluated in consideration of the information provided.

[0132] Finally, in step 240, one pattern is selected that can be the best fitting or the first pattern to fulfill the requirements, particularly, the external demands 232. This pattern forms the basis for the schedule used or corresponds to the schedule used.

[0133] FIG. 9 illustrates an operation of an embodiment of the schedule as a component of the control structure. This operation can be done by any suitable device within the control structure or by the control structure itself. In a first step 200 the scheduler provides the result to be achieved with the smart grid within the smart grid, e.g. defining the amount of energy to be produced or provided by the generators at the time steps within the considered time period. Additionally or alternatively, the amount of energy to be consumed by the consumers at the time steps within the considered time period is defined. In a particular embodiment, the amount of energy at the time steps within the considered time period is fixed and only the consumption of energy at the time steps within the time period is defined.

[0134] In a next step 202, the scheduler regards the expectation values 204, 206, 208 which can be never switch on 204, it would be nice to switch on 206, and please switch on 208. Historical data can be considered. At the end the scheduler generates a rank array 210 which is used in simulation.

[0135] FIG. 10 illustrates an embodiment of the proposed method. In a first step 250 expectation values 252 are provided by the agents of the machines. These expectation values 252 together with first external demands 254, e.g. regarding the types and conditions of the machines, are input in a rank unit 256 generating a rank array 258. This rank array considers the machines at all time steps within the agreed time period.

[0136] The rank array 258 and second external demands 260, e.g. regarding the demands regarding energy provision and consumption, are used in a simulation 262 to select a schedule 264 comprising the requests 268 to be sent to the machines. A recorder 270 within the control structure, e.g. within the realization unit, compares the requests 268 with responses 272 and records 280 these responses and the differences between requests 268 and responses 272.

[0137] FIG. 11 illustrates generation of an rank array using a rank unit 300. Expectation values 302 comprising fuzzy values 304 of these expectation values 302 and external demands 310 also comprising fuzzy values 312 are input in the rank unit 300. In this embodiment, the result is a three-dimensional rank array 320 considering the fuzziness of the input data.

[0138] Thus, a two-dimensional rank array comprises a data set to machines and time steps for one time period. A three-dimensional rank array comprises a number of data sets to the machines and time step for one time period. Therefore, the fuzziness of circumstances and behavior of the machines, e.g. expressed by historical data, can be considered.

[0139] In one embodiment, the scheduler sends special offers to the machines to try to influence their decisions, e.g. a special offer regarding the price for energy to be consumed or to be provided.

[0140] Furthermore, the scheduler can send a request connected with an amount of energy to be consumed or provided, i. e. the scheduler can send requests to a machine defining in time steps the amount of energy to be consumed or to be provided. Additionally, the expectation values associated with the machines can consider amounts of energy. For example, in time step n generator can have an expectation value of 0.5 for 20 kW and an expectation value of 0.3 for 40 kW. That means that in time step n the generator would prefer to provide 20 kW instead of 40 kW.

[0141] The arrangement for operating a smart grid and the method for operating the smart grid described herein propose a new approach to consider the wants of the members in this smart grid and to fulfill the demands. The arrangements provides a schedule comprising requests sent to the machines. However, these requests are subdominant. The machines are not committed to the requests, rather the machines can switch and provide or consume energy according to their wants.

[0142] As the willingness for providing or consuming energy by the machines is calculated in advance and communicated to the arrangement with help of expectation values, the chance for a working schedule is high. In case one or more of the machines do not follow, a Fill routine helps to compensate this. In case of extensive changes in external circumstances and/or demands, the Reschedule routine helps to generate a new schedule suitable for the new situation.

[0143] FIG. 12 shows in schematic view an embodiment of the proposes arrangement denoted with reference number 400 used for controlling the operation of a number of floor conveyors 402 (only one shown) on a factory floor representing the technical installation. A control structure 404 has to calculate a course of action for at least one of the floor conveyors to make sure that goods are delivered from a store 406 to positions in the factory floor to ensure a suitable overall procedure. Therefore, the control structure 404 uses expectation values 408, i.e. courses of expectation values 408 of some or all of the floor conveyors 402 representing their will to act in a special way, e.g. turn left or turn right, within a predetermined period. The expectation values can be calculated or determined by using historical data, that means data representing the behavior of the respective floor conveyor 402 in the past.

[0144] The expectation value for the shown floor conveyor 402 could be a probability of 0.7 to turn left and of 0.1 to turn right at time point t.sub.1. Expectation values represent the likelihood of a special behavior of a machine, here the floor conveyor 402, at a defined time point in the future.

[0145] The control structure 404 simulates operation of a technical installation, here the factory floor, comprising a number of machines, here the floor conveyors 402, and calculates behaviors or action of the machines to achieve a wanted result or, e.g. at least 80% of a wanted result. The wanted actions of the machines are translated in requests 410 sent to the floor conveyors 402. Each floor conveyor 402 can independently decide to follow this request or not and usually sends back a signal representing its will to the control structure 404. The control structure 404 evaluates the responses of the floor conveyors 402 and decides, usually on basis of a simulation, whether it is necessary to send new requests to the floor conveyors 402.