Energy operation planning device and energy operation planning method

Abstract

To provide an energy operation planning device which can find a solution within a desired calculation required time by selecting appropriate solution means and constraint.SOLUTION: An energy operation planning device (1) concerning the present disclosure comprises: a problem information input unit (10) which inputs a variable and a desired calculation required time; a constraint determination unit (20) which estimates intensity of a constraint of an optimization problem of an energy operation plan by using a learned model for estimating the intensity of the constraint from the variable and the calculation required time to determine a constraint to be used; a solution means determination unit (30) which determines priority of solution means to be used for the optimization problem of the energy operation plan by using the learned model for determining the priority of the solution means from the variable and the calculation required time; and an energy operation plan optimization unit (40) which establishes the energy operation plan by sequentially solving the optimization problem by the solution means of high priority determined by the solution means determination unit by using the constraint determined by the constraint determination unit.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a plan to supply the required energy by combining appliances for multiple energy demands. [Background technology]

[0002] In recent years, energy management systems (EMS), which manage the supply of energy, including electricity and heat, to facilities such as factories and buildings, have been attracting attention as an example of the application of optimization solutions to the energy field from the perspective of cost reduction. EMS models energy supply systems consisting of various devices such as generators, storage batteries, turbines, and boilers, determines the optimal output value for each device, and formulates an energy operation plan.

[0003] To formulate an energy operation plan, a solver (a means for solving a problem) composed of algorithms based on mathematical theory is used. The means for solving a problem is primarily used for numerical calculations, and there is a method in which parameters are defined for each problem to be solved, solution method, and technique for multiple means for solving a problem, and algorithms are compared based on calculation time, number of times required for convergence, norm of residual, etc. (e.g., Patent Document 1). There is also a method in which multiple libraries equipped with calculation means are prepared, a user-defined classification is set for each problem to be solved, and numerical calculations are performed using a library selected according to that classification (e.g., Patent Document 2).

[0004] In the field of optimization, there is a method in which the minimum constraints required for an optimization problem to be solved are set with user-defined classifications assigned to the constraints, and the extent to which the solved result satisfies the constraints is evaluated (for example, Patent Document 3).In formulating energy operation plans, there is also a technology in which multiple optimization techniques are applied to one or more selected scenarios from multiple forecast scenarios in order to optimally control multiple energy storage systems (for example, Patent Document 4). [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 5-282353 [Patent Document 2] Japanese Patent Application Laid-Open No. 2008-4043 [Patent Document 3] Japanese Patent Application Laid-Open No. 2014-86022 [Patent Document 4] Special Publication No. 2021-505103 Summary of the Invention [Problem to be solved by the invention]

[0006] However, when formulating an energy operation plan, the required calculation time and the value of the objective function may differ depending on the model of the energy supply system being calculated. In prior art documents, the solution method and the required calculation time for numerical calculation, as well as the objective function and constraints, are compared and selected. Furthermore, when formulating an energy operation plan, the solution method and constraints have a significant impact on the required calculation time and the value of the objective function. Therefore, in order to increase the reliability of optimization calculations, it is necessary to correlate and compare the solution method, the required calculation time, the objective function, and the constraints to understand the relationship. Although users may understand this empirically, there are still issues, such as a deterioration in the value of the objective function due to the correlation between each parameter or the calculation not completing.

[0007] The present disclosure has been made to solve the above-mentioned problems, and aims to provide an energy operation planning device that can evaluate the objective function, calculation time, and constraints for each solution-finding means when formulating an energy operation plan, and can find a solution within the desired calculation time using an appropriate solution-finding means. [Means for solving the problem]

[0008] The energy operation planning device according to the present disclosure is an energy operation planning device that formulates an energy operation plan that determines control command values ​​at each time for equipment that generates energy including electricity or heat, and includes: a problem information input unit to which variables used in determining the objective function and constraints of an optimization problem of the energy operation plan and a desired calculation time required to optimize the energy operation plan are input; a constraint condition determination unit that uses a trained model that estimates the strength of the constraints from the variables and the calculation time to estimate the strength of the constraints to be used in the optimization problem from the variables input in the problem information input unit and the desired calculation time, and determines the constraints to be used in the optimization problem; a solution-finding means determination unit that uses a trained model that determines the priority of the solution-finding means from the variables and the calculation time, and determines the priority of the solution-finding means to be used in the optimization problem from the variables input in the problem information input unit and the desired calculation time; and an energy operation plan optimization unit that formulates the energy operation plan by solving the optimization problem in order using the solution-finding means with the highest priority determined in the solution-finding means determination unit using the constraints determined in the constraint condition determination unit. Another energy operation planning device according to the present disclosure is an energy operation planning device that formulates an energy operation plan that determines control command values ​​at each time for equipment that generates energy including electric power or heat, and includes a problem information input unit to which variables used to determine the objective function and constraints of an optimization problem for the energy operation plan and a desired calculation time required to optimize the energy operation plan are input, a constraint condition determination unit that uses a trained model that estimates the strength of the constraints from the variables and the calculation time to estimate the strength of the constraints to be used in the optimization problem from the variables and the desired calculation time input in the problem information input unit and determines the constraints to be used in the optimization problem, and uses the constraints determined in the constraint condition determination unit to: Multiple The system also includes an energy operation plan optimization unit that formulates an energy operation plan by solving an optimization problem using the solution-finding means.

[0009] The energy management planning method according to the present disclosure is an energy management planning device that formulates an energy management plan that determines control command values ​​at each time for equipment that generates energy including electricity or heat, and includes: a problem information input step of inputting variables to be used in determining the objective function and constraints of an optimization problem of the energy management plan and a desired calculation time required to optimize the energy management plan; a constraint condition determination step of estimating the strength of the constraints to be used in the optimization problem from the variables input in the problem information input step and the desired calculation time using a trained model that estimates the strength of the constraints from the variables and the calculation time, and determining the constraints to be used in the optimization problem; a solution-finding means determination step of determining the priority of the solution-finding means to be used in the optimization problem from the variables input in the problem information input step and the desired calculation time using a trained model that determines the priority of the solution-finding means from the variables and the calculation time; and an energy management plan optimization step of formulating the energy management plan by solving the optimization problem in order using the solution-finding means with the highest priority determined in the solution-finding means determination step using the constraints determined in the constraint condition determination step. Another energy management planning method according to the present disclosure is an energy management planning device that formulates an energy management plan for determining control command values ​​at each time for equipment that generates energy including electric power or heat, the method comprising: a problem information input step of inputting variables used to determine an objective function and constraints for an optimization problem of the energy management plan and a desired calculation time required for optimizing the energy management plan; a constraint condition determination step of estimating the strength of the constraints to be used in the optimization problem from the variables and the desired calculation time inputted in the problem information input step using a trained model that estimates the strength of the constraints from the variables and the calculation time, and determining the constraints to be used in the optimization problem; and Multiple and an energy operation plan optimization step of formulating an energy operation plan by solving the optimization problem using the solution-finding means at the same time. [Effects of the Invention]

[0010] The energy management planning device and energy management planning method according to the present disclosure use a trained model that determines the strength of constraints based on variables and required calculation time to determine the strength of constraints used in an optimization problem, thereby reducing the risk of an optimized energy management plan having no solution. Furthermore, by using a trained model that has learned the priority of solution-finding methods based on variables and desired required calculation time to determine the priority of solution-finding methods used in an optimization problem, it is possible to apply a solution-finding method appropriate for each optimization problem being handled. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a configuration diagram illustrating an energy operation planning device according to a first embodiment of the present disclosure. [Figure 2] 1 is a configuration diagram illustrating an example of a model of an energy supply system that is a calculation target in an energy operation planning device according to a first embodiment of the present disclosure. [Figure 3] 1 is a schematic diagram illustrating a hardware configuration of an energy operation planning device according to a first embodiment of the present disclosure. [Figure 4] 4 is a flowchart showing the input / output relationship of data related to the constraint condition learning process according to the first embodiment of the present disclosure. [Figure 5] 1 is a flowchart showing the input / output relationship of data related to a learning process for determining priority levels of a solution-finding means according to the first embodiment of the present disclosure. [Figure 6] 10 is a flowchart showing a method for re-determining constraint conditions and continuing to solve within a desired calculation time after prioritizing the solution-finding means to be used in the solution-finding process of the energy operation planning device according to the first embodiment of the present disclosure. [Figure 7] This is a flowchart showing a method for continuing to solve problems by prioritizing the solution-finding means to be used within a desired calculation time after determining constraints using a learning model, in the solution-finding process of an energy operation planning device according to embodiment 2 of the present disclosure. [Figure 8]11 is a flowchart showing a method of simultaneously using a plurality of solution-finding means in the solution-finding process of the energy operation planning device according to the third embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0012] Embodiment 1

[0013] 1 is a configuration diagram showing an energy operation planning device according to a first embodiment of the present disclosure. The energy operation planning device 1 includes, for example, a problem information input unit 10, a constraint condition determination unit 20, a solution-finding means determination unit 30, an energy operation plan optimization unit 40, a storage unit 50, and a calculation result display unit 60. The energy operation planning device 1 formulates an energy operation plan that determines control command values ​​at each time for equipment that generates energy, including electric power and heat.

[0014] Information used for calculating the optimization problem of the energy management plan is input to the problem information input unit 10. The information input to the problem information input unit 10 includes, for example, the objective function and constraints of the optimization problem, variables, input / output characteristic expressions, calculation frame, desired calculation time required for optimization, demand for energy including electricity or heat, fuel purchase price, and electricity purchase price or sales price.

[0015] The objective function is a function that represents, for example, the cost of fuel use in the energy supply system, the cost of electricity use, the amount of primary energy used, the amount of greenhouse gas emissions, etc. The variables input to the problem information input unit 10 are, for example, the amount of gas purchased, the amount of electricity purchased, boot A stop flag, etc.

[0016] Constraint conditions include, for example, an electricity supply and demand matching constraint in the target energy operation plan, a steam supply and demand matching constraint, upper and lower output limit constraints for model equipment, input and output ramp rate constraints for model equipment, operating time constraints for model equipment, start / stop constraints for model equipment, constraints on the number of starts and stops for model equipment, and constraints that determine the starting method for model equipment.

[0017] The energy operation planning device 1 models an energy supply system so that an objective function is minimized under predetermined constraints, solves an optimization problem, and formulates an energy operation plan. Here, the energy operation plan is a plan formulated by optimizing an energy supply system including at least one of fuel storage facilities, heat source equipment, generators, power storage facilities, thermal storage facilities, and power systems, and determining control command values ​​for each time period.

[0018] Figure 2 is a configuration diagram showing a model of an energy supply system input to a problem information input unit according to the first embodiment of the present disclosure. As an example, an optimization problem for satisfying power demand and steam demand for a model consisting of an equipment configuration of one boiler and one turbine generator as shown in Figure 2 will be given below, and a specific example of input information required for formulating an energy operation plan will be described. In the model shown in Figure 2, the objective function, constraints, and input / output characteristics for formulating an energy operation plan can be expressed, for example, by the following equations (1) to (5).

[0019]

number

[0020] Equation (1) expresses the objective function, where t is the time frame and C g (t) is the cost associated with fuel consumption, C E (t) is the cost related to the amount of electricity used. In equation (1), it is expressed in terms of the cost of fuel consumption and the amount of electricity used, but it may also be expressed in terms of greenhouse gas emissions, etc. Furthermore, although equation (1) is a first-order polynomial, the form of the objective function according to the present disclosure is not limited to this example.

[0021]

number

[0022] Equation (2) is the constraint on the supply and demand balance consisting of the amount of electricity used, the amount of turbine power generated, and the amount of electricity demand. E(t) is the amount of electricity purchased from the power grid, E turbine(t) is the turbine generator output, E L (t) is the power demand required in the energy operation plan.

[0023]

number

[0024] Equation (3) is the constraint on the supply and demand balance consisting of the amount of steam generated by the boiler, the amount of steam consumed by the turbine generator, and the steam demand. boiler (t) is the amount of steam generated in the boiler, S turbine (t) is the steam consumption of the turbine generator, S L (t) represents the steam demand in the energy operation plan. In equations (1) to (5), the objective function, constraints, and input / output characteristics are expressed using continuous variables as examples, but discrete variables may also be used.

[0025] In this example, the constraint condition is the supply and demand balance between electricity demand and steam demand, but upper and lower output limit constraints, constraints that determine the start and stop of equipment, etc. may also be used, and the embodiments of the present disclosure are not limited thereto.

[0026]

number

[0027]

number

[0028] Equation (4) and equation (5) are characteristic equations of the input and output of the turbine generator and boiler in the energy operation plan. Equation (4) expresses the power generation amount of the turbine generator, and f turbine (S turbine (t)) is a function that determines the amount of power generated by the turbine generator. turbine (t) is the amount of steam input to the turbine generator.

[0029] Equation (5) expresses the steam generation rate of the boiler, and f boiler (gboiler (t)) is a function that determines the amount of steam generated in the boiler. boiler (t) is the amount of raw gas consumed by the boiler.

[0030] In this embodiment, with regard to the input / output characteristic equations, the input / output characteristics of the turbine generator are expressed by the input steam amount, and the input / output characteristics of the boiler are expressed by the raw material gas consumption amount, but this expression method is not intended to limit the embodiments according to the present disclosure.

[0031] In the energy operation planning device 1, the input / output characteristics of the equipment are modeled, constraints and objective functions are set, and an optimization problem is calculated. In the energy operation planning device 1 of the present disclosure, parameters required for the optimization problem related to the energy operation plan are learned, thereby preparing models with strong and weak constraints to be calculated, and further assigning priorities to the solution-finding means.

[0032] The constraint condition determination unit 20 uses a trained model that estimates the strength of constraint conditions from variables and the required calculation time to estimate the strength of constraint conditions to be used in the optimization problem from the variables and the required calculation time input by the problem information input unit 10, and determines the constraint conditions to be used in the optimization problem.

[0033] The solution-finding means determination unit 30 uses a trained model that determines the priority of solution-finding means based on variables and required calculation time to determine the priority of solution-finding means to be used in the optimization problem based on the variables and desired required calculation time input by the problem information input unit 10.

[0034] In this embodiment, the solution means is an algorithm for solving the optimization problem, and examples of the solution means include the simplex method, the interior point method, the branch and bound method, and combinations of these methods.

[0035] The energy management plan optimization unit 40 includes, for example, the problem information input unit 10, the constraint condition determination unit 20, a program description unit 41 that writes a program for solving the optimization problem based on the number of variables, constraint conditions, and solution means determined by the solution means determination unit 30, a solution means execution unit 42 that has multiple solution means for performing calculations, a calculation result comparison unit 43 that compares the values ​​of the objective function of the optimization problem depending on the type of solution means and constraint conditions, a solution means learning unit 44 that creates a solution means learned model, and a constraint condition learning unit 45 that creates a constraint condition learned model.

[0036] The storage unit 50 includes, for example, an objective function storage unit 51 that stores the values ​​of the objective functions when solving past optimization problems, a constraint condition storage unit 52 that stores the constraint conditions used, a variable information storage unit 53 that stores the number and types of variables used, a required calculation time storage unit 54 that stores the calculation time required to find a solution, and a trained model that determines the priority of the solution-finding means. Remember Solution-finding means trained model storage unit 55, trained model for estimating the strength of constraint conditions Remember The system includes a constraint condition learned model storage unit 56.

[0037] After solving the optimization problem, the calculation result display unit 60 displays the compared objective function values ​​and the calculation time required according to the solution means or the type of constraint condition.

[0038] The solution-finding means learning unit 44 and the constraint condition learning unit 45 may be, for example, a means for determining output information by supervised learning such as a neural network model (not shown). Supervised learning is a method of providing a learning device with pairs of input and result data, learning the characteristics of the learning data, and inferring the result from the input. A neural network is composed of an input layer consisting of multiple neurons, an intermediate layer (hidden layer) consisting of multiple neurons, and an output layer consisting of multiple neurons. The intermediate layer may be one layer, or two or more layers.

[0039] FIG. 3 is a schematic diagram showing a hardware configuration of an energy operation planning device according to the first embodiment of the present disclosure. An example of an environment for executing a method for reducing the calculation time required for optimization calculation and improving the calculation result is a computer 100. The computer 100 includes, for example, a CPU (Central Processing Unit) 101, a main memory device 102, and an auxiliary memory device 103. For example, the main memory device 102 may be a RAM, and the auxiliary memory device 103 may be a ROM or other storage medium. An external input device 200 and an external output device 300 are provided outside the computer 100. The external input device 200 is, for example, a mouse or a keyboard. The external output device 300 is, for example, a display monitor that displays the output results from the computer 100. In addition, in FIG. 3 As shown in FIG. 1, the computer 100 may be connected to an external storage device 500 via a network 400 so that the results of calculations can be stored in an external storage medium.

[0040] The problem information input unit 10 is realized by inputting data using an external input device 200. The storage unit 50 is realized, for example, by a main storage device 102, an auxiliary storage device 103, and an external storage device 500. The constraint condition determination unit 20, the solution-finding means determination unit 30, and the energy management plan optimization unit 40 are realized by a CPU 101 that performs calculation processing. The calculation result display unit 60 is equipped with an output device such as a display monitor, making it possible to visualize the contents stored in the storage unit 50 and the calculation results of the energy management plan optimization unit 40.

[0041] 4 is a flowchart showing the input / output relationship of data related to the constraint learning process according to the first embodiment of the present disclosure. In step S41, the constraint learning unit 45 acquires variables used in past optimization problems from the variable information storage unit 53 of the storage unit 50, constraints used in the past from the constraint storage unit 52, and the calculation time required for the past optimization problems from the calculation time storage unit 54.

[0042] However, when retrieving constraints used in the past, a score is set for each type of constraint. The user sets a value for the score of the constraint to be set corresponding to the strength of the constraint for each type of constraint. One possible method for determining the value corresponding to the strength of the constraint is to set a value between 0 and 100 depending on the strength of the constraint, for example.

[0043] In step S42, the constraint condition learning unit 45 learns the constraint conditions based on the variables of the optimization problem and the required calculation time acquired from the problem information input unit 10 so that the constraint conditions approach the scores of the constraint conditions used in past optimization problems stored in the storage unit 50. In other words, the constraint condition learning unit 45 generates a learned model for determining the scores of the constraint conditions from the variables of the optimization problems calculated in the past and stored in the storage unit 50, and the required calculation time.

[0044] In step S43, the trained model generated by the constraint condition training unit 45 is stored in the constraint condition trained model storage unit 56 of the storage unit 50.

[0045] Next, the constraint condition determination unit 20 inputs the variables and the desired calculation time using the trained model generated by the constraint condition learning unit 45, and outputs a score for determining the strength of the constraint condition. In step S44, the constraint condition determination unit 20 acquires the variables and the desired calculation time input by the problem information input unit 10. In step S45, the constraint condition determination unit 20 inputs the variables and the desired calculation time to the constraint condition trained model storage unit 56 of the storage unit 50. In step S46, the constraint condition determination unit 20 outputs the variables and the desired calculation time, and a score for each type of constraint condition, such as facility information and output characteristics, from the trained model.

[0046] In step S47, the constraint condition determination unit 20 determines the strength of the constraint condition based on the score for each type of output constraint condition. One method of determining the score is to sort the scores in descending order, and treat the top one-third of the scores as weak constraint conditions, the top two-thirds as medium constraint conditions, and those that use all constraint conditions as strong constraint conditions, for example.

[0047] However, depending on the energy operation plan input to the problem information input unit 10, it is possible that a type may include a constraint that does not exist, so it is necessary to count the score ranking without including the type of constraint that does not exist. In step S48, the constraint determination unit 20 outputs the constraints to the program description unit 41 in order from weakest to strongest, based on the information on the strength of the constraints obtained up to step S47.

[0048] 5 is a flowchart showing the learning process for determining a solution-finding means and the output relationship of related data according to an embodiment of the present disclosure. In step S51, the solution-finding means learning unit 44 acquires variables used in past optimization problems from the variable information storage unit 53 of the storage unit 50. The solution-finding means learning unit 44 also acquires the names of available solution-finding means from the solution-finding means execution unit 42 of the energy management plan optimization unit 40. The solution-finding means learning unit 44 also acquires the calculation time required for past optimization problems from the calculation time storage unit 54.

[0049] In step S52, the solution-finding means learning unit 44 learns the solution-finding means based on the variables of the optimization problem and the required calculation time acquired by the problem information input unit 10 so that the solution-finding means matches the solution-finding means used for past optimization problems stored in the storage unit 50. In other words, the solution-finding means learning unit 44 generates a trained model for determining the score of the solution-finding means from the variables of the optimization problem calculated in the past and the required calculation time.

[0050] In step S53, the trained model generated by the solution-finding means training unit 44 is stored in the solution-finding means trained model storage unit 55 of the storage unit 50.

[0051] Next, the solution-seeking means determination unit 30 inputs the variables and the desired calculation time using the trained model generated by the solution-seeking means learning unit 44, and outputs a score for the solution-seeking means. In step S54, the solution-seeking means determination unit 30 acquires the type and number of variables and the desired calculation time input by the problem information input unit 10. In step S55, the solution-seeking means determination unit 30 inputs the type and number of variables to be used in calculating the optimization problem and the desired calculation time to the solution-seeking means trained model storage unit 55 of the storage unit 50. In step S56, the solution-seeking means determination unit 30 outputs a score for each solution-seeking means from the acquired variables, desired calculation time, and trained model.

[0052] In step S57, the solution-finding means determination unit 30 sorts the output scores for each type in descending order of score and ranks the scores of each solution-finding means. This determines the priority of the solution-finding means. In step S58, the solution-finding means determination unit 30 outputs the priority of the solution-finding means to the solution-finding means execution unit 42 of the energy management plan optimization unit 40.

[0053] In this embodiment, the case where supervised learning is applied to the learning algorithm used by the solution-finding means learning unit 44 and the constraint condition learning unit 45 has been described, but the present invention is not limited to this. As for the learning algorithm, reinforcement learning, unsupervised learning, semi-supervised learning, or the like can also be applied in addition to supervised learning.

[0054] 6 is a flowchart showing the processing of the energy management planning device 1 according to the first embodiment of the present disclosure. In step S61, the user mathematically models information used in calculating the optimization problem of the energy management plan and inputs the model to the problem information input unit 10. The information used in calculating the optimization problem of the energy management plan includes, for example, an objective function, variables required for calculating the optimization problem, constraints to be used, input / output characteristics of the modeled equipment, etc.

[0055] In step S62, in order to optimize the energy management plan within the calculation time desired by the user, the user inputs the calculation time to the problem information input unit 10. In step S63, the solution-finding means determination unit 30 acquires the variables of the optimization problem input to the problem information input unit 10 and the desired calculation time, and determines the priority of the solution-finding means to be used.

[0056] In step S64, the constraint condition determination unit 20 acquires the variables of the optimization problem and the desired calculation time input to the problem information input unit 10, and outputs the pattern of the weakest constraint conditions among the constraint conditions of the energy management plan to be solved, thereby determining the constraint conditions to be used for the optimization problem. In step S65, the solution-finding means execution unit 42 of the energy management plan optimization unit 40 selects the solution-finding means with the highest priority from the solution-finding means determined in step S63.

[0057] In step S66, the program description unit 41 writes a program for the energy management plan that is tailored to the solution-finding means, and the solution-finding means execution unit 42 calculates the optimization problem for the energy management plan using the solution-finding means selected in step S65.

[0058] In step S67, calculation result comparison unit 43 records the calculation time required until the optimization of the energy management plan is completed, and compares it with the desired calculation time set in problem information input unit 10. After comparing the calculation times, calculation result comparison unit 43 selects a constraint that is the next weaker constraint than the one determined in the previous step S64, and repeats the processing from step S64 to step S67 until the set calculation time is reached.

[0059] In step S68, the calculation result display unit 60 displays the calculation results calculated in the optimization of the energy operation plan. The calculation results include at least one of the cost of fuel use in the energy supply system, the cost of electricity use, the amount of primary energy used, the amount of greenhouse gas emissions, the command value for the fuel storage facility, the command value for the heat source machine, the command value for the generator, the command value for the power storage facility, the command value for the heat storage facility, and the command value for the power grid.

[0060] When the calculation results are displayed in step S68, the results that are most suited to the purpose of optimizing the energy management plan may be displayed, or the results that require the shortest calculation time may be output.

[0061] As described above, by repeating steps S61 to S67, the solution-finding means with the highest priority is selected from among multiple solution-finding means in advance, and the process of determining the constraint conditions to be used for the optimization problem in the constraint condition determination unit 20 is repeated within the desired calculation time input to the problem information input unit, thereby optimizing the energy operation plan.

[0062] By solving the optimization of the energy operation plan in this way, the energy operation planning device 1 not only reduces the risk of finding no solution for the energy operation plan that is to be optimized, but also makes it possible to improve the calculation results and shorten the calculation time.

[0063] Embodiment 2 In the first embodiment, after determining the priority of the solution-finding means to be used in step S63, the constraints to be used are determined in step S64, the solution-finding means with the highest priority is selected in step S65, calculation is performed in step S66, and if the energy operation plan is optimized within the required calculation time set in step S67, the process returns to step S64 and repeats steps S64 to S67 until the required calculation time is reached. In contrast, in the present embodiment, the constraints are determined first, the priority of the solution-finding means is determined, and then a solution-finding means is selected within the required calculation time according to the priority.

[0064] 7 is a flowchart showing the processing of the energy operation planning device according to embodiment 2. However, since steps S71, S72, S76, and S78 have already been explained in step S61, S62, S66, and S68 of embodiment 1, explanations thereof will be omitted here.

[0065] In step S73, the constraint condition determination unit 20 acquires the variables of the optimization problem and the desired calculation time input to the problem information input unit 10, and outputs the pattern of the weakest constraint conditions among the constraint conditions of the energy management plan to be solved. In step S74, the solution-finding means determination unit 30 acquires the information input to the problem information input unit 10, and determines the priority of the solution-finding means to be used for optimizing the energy management plan. In step S75, the solution-finding means execution unit 42 selects the solution-finding means to be used for the optimization problem according to the priority of the solution-finding means determined in step S74.

[0066] In step S77, the calculation result comparison unit 43 records the calculation time required until the optimization of the energy management plan is completed, and compares it with the calculation time required by the user set in the problem information input unit 10. After the calculation result comparison unit 43 compares the calculation times, if the optimization of the energy management plan is completed within the set calculation time, the process returns to step S75 and the previous step S7 5 in choice The solution-finding method with the next highest priority is selected and the optimization calculation is performed. The process from step S75 to step S77 is repeated until the set calculation time is reached.

[0067] As described above, by repeating steps S71 to S77, the energy operation planning device 1 determines constraints for the optimization problem in advance using the constraint condition determination unit 20, and can sequentially solve the optimization problem using multiple prioritized solution means. By solving the optimization of the energy operation plan in this way, as in the first embodiment, it is possible to reduce the risk of no solution being found for the energy operation plan to be optimized, improve the calculation results, and shorten the calculation time.

[0068] Embodiment 3 In the present disclosure, since a plurality of solution-finding means are provided, it is possible to optimize the energy operation plan not only by using the solution-finding means in sequence but also by using a plurality of solution-finding means simultaneously. Multiple solution methods simultaneously Solving do This is a flowchart showing a method for optimizing an energy management plan using a plurality of solution-finding means simultaneously. Steps S81, S82, and S86 have already been described in relation to steps S61, S62, and S66 in the first embodiment, and therefore will not be described here.

[0069] In step S83, the constraint condition determination unit 20 acquires the variables of the optimization problem input to the problem information input unit 10 and the desired calculation time, outputs the pattern of the weakest constraint conditions among the constraint conditions of the energy operation plan to be solved, and determines the constraint conditions to be used for the optimization problem.

[0070] In step S84, based on the constraints determined in step S83, the program description unit 41 writes a program for an energy management plan that is tailored to the solution-finding means, and the solution-finding means execution unit 42 executes calculations simultaneously using all selectable solution-finding means.

[0071] In step S85, the calculation result comparison unit 43 records the calculation time required until the optimization of the energy management plan is completed, and compares it with the desired calculation time set in the problem information input unit 10. After the calculation result comparison unit 43 compares the calculation times, if the optimization of the energy management plan is completed within the set calculation time, the calculation result comparison unit 43 returns to step S83 and determines whether the next constraint condition is determined after the one determined in the previous step S83. weak The program for the energy management plan is written so that the required calculation time is reached, and the processing from step S83 to step S85 is repeated until the set required calculation time is reached.

[0072] As described above, by repeating steps S83 to S85, the energy operation planning device 1 can repeatedly determine constraint conditions for the optimization problem related to the energy operation plan in the constraint condition determination unit 20 within the desired required calculation time input in the problem information input unit 10, and can solve the optimization problem using multiple solution means simultaneously. By solving the optimization of the energy operation plan in this manner, it is possible to reduce the risk of not finding a solution for the energy operation plan to be optimized, improve the calculation results, and shorten the required calculation time, as in the first embodiment.

[0073] The configurations described in the above embodiments are merely examples of the contents of the present disclosure, and may be combined with other known technologies. Furthermore, parts of the configurations may be omitted or modified without departing from the scope of the present disclosure.

[0074] Various aspects of the present disclosure are summarized below as appendices. (Appendix 1) an energy management planning device that formulates an energy management plan that determines control command values ​​for each time period for equipment that generates energy including electricity or heat, the energy management planning device comprising: a problem information input unit that inputs variables used to determine an objective function and constraints for an optimization problem of the energy management plan, and a desired calculation time required to optimize the energy management plan; a constraint condition determination unit that uses a trained model that estimates the strength of the constraints from the variables and the required calculation time to estimate the strength of the constraints to be used in the optimization problem from the variables and the desired calculation time input by the problem information input unit, and determines the constraints to be used in the optimization problem; a solution finding means determination unit that uses a trained model that determines the priority of solution finding means from the variables and the required calculation time to determine the priority of solution finding means to be used in the optimization problem from the variables and the desired calculation time input by the problem information input unit; and an energy management plan optimization unit that formulates the energy management plan by solving the optimization problem in order using the constraints determined by the constraint condition determination unit using the solution finding means with the highest priority determined by the solution finding means determination unit. (Appendix 2) 2. The energy management planning device according to claim 1, wherein the energy management plan optimization unit formulates the energy management plan by simultaneously using a solution-finding means to solve the optimization problem. (Appendix 3) An energy operation planning device as described in Appendix 1 or 2, comprising a constraint condition learning unit that generates the trained model that determines the strength of the constraint condition from the variables used in the optimization problem and the calculation time required. (Appendix 4) An energy operation planning device according to any one of appendices 1 to 3, comprising a solution-finding means learning unit that generates the trained model that determines the priority of solution-finding means based on variables used in the optimization problem and the calculation time required. (Appendix 5) 5. The energy operation planning device according to claim 1, wherein the objective function of the optimization problem includes at least one of the cost of fuel usage, the cost of electricity usage, the amount of primary energy usage, and the amount of greenhouse gas emissions in the energy operation plan. (Appendix 6) 6. The energy operation planning device according to any one of appendixes 1 to 5, wherein the equipment used in the energy operation plan includes at least one of a fuel storage facility, a heat source machine, a generator, a power storage facility, a heat storage facility, and a power system. (Appendix 7) 7. The energy operation planning device according to claim 1, wherein the solution-finding means determination unit determines in advance a solution-finding means to be used, and the energy operation plan optimization unit repeats a process of determining the constraint conditions to be used in the optimization problem in accordance with the strength of the constraint conditions to be used in the constraint condition determination unit within a desired calculation time input to the problem information input unit. (Appendix 8) 7. The energy operation planning device according to claim 1, wherein the constraint condition determination unit determines constraint conditions to be used in advance, and the energy operation plan optimization unit repeats a process in which the solution-finding means determination unit determines the solution-finding means to be used in accordance with the priority of the solution-finding means within the desired calculation time input to the problem information input unit. (Appendix 9) 7. The energy operation planning device according to claim 1, wherein the constraint condition determination unit repeats a process of determining the constraint conditions to be used in the optimization problem within the desired calculation time, and the energy operation plan optimization unit executes calculations simultaneously using a plurality of solution-finding means. (Appendix 10) an energy management planning device that formulates an energy management plan that determines control command values ​​for each time period for equipment that generates energy including electricity or heat, the energy management plan comprising: a problem information input step of inputting variables to be used in determining an objective function and constraints for an optimization problem of the energy management plan, and a desired calculation time required to optimize the energy management plan; a constraint condition determination step of determining constraint conditions to be used in the optimization problem from the variables and the desired calculation time inputted in the problem information input step, using a trained model that estimates the strength of constraints from the variables and the calculation time; a solution-finding means determination step of determining priorities of solution-finding means to be used in the optimization problem from the variables and the desired calculation time inputted in the problem information input step, using a trained model that determines the priorities of solution-finding means from the variables and the calculation time; and an energy management plan optimization step of formulating the energy management plan by solving the optimization problem in order using the solution-finding means with the highest priorities determined in the solution-finding means determination step, using the constraints determined in the constraint condition determination step. (Appendix 11) 11. The energy management planning method according to claim 10, wherein the energy management plan optimization step formulates the energy management plan by solving the optimization problem simultaneously using a solution-finding means. [Explanation of symbols]

[0075] 1 Energy operation planning device, 10 Problem information input unit, 20 Constraint condition determination unit, 30 Solution-finding means determination unit, 40 Energy operation plan optimization unit, 41 Program description unit, 42 Solution-finding means execution unit, 43 Calculation result comparison unit, 44 Solution-finding means learning unit, 45 Constraint condition learning unit, 50 Memory unit, 51 Objective function memory unit, 52 Constraint condition memory unit, 53 Variable information memory unit, 54 Calculation required time memory unit, 55 Solution-finding means learned model memory unit, 56 Constraint condition learned model memory unit, 60 Calculation result display unit, 100 Computer, 101 CPU, 102 Main memory unit, 103 Auxiliary memory unit, 200 External input device, 300 External output device, 400 Network, 500 External memory unit

Claims

1. 1. An energy operation planning device that formulates an energy operation plan that determines control command values ​​for devices that generate energy including electric power or heat at each time, a problem information input unit into which variables used to determine an objective function and constraint conditions for the optimization problem of the energy management plan and a desired calculation time required for optimizing the energy management plan are input; a constraint condition determination unit that estimates the strength of the constraint condition to be used in the optimization problem from the variables and the desired required calculation time input by the problem information input unit using a trained model that estimates the strength of the constraint condition from the variables and the required calculation time, and determines the constraint condition to be used in the optimization problem; a solution-finding means determination unit that determines a priority order of solution-finding means to be used for the optimization problem based on the variables and the desired required calculation time input by the problem information input unit, using a trained model that determines a priority order of solution-finding means based on the variables and the required calculation time; an energy management plan optimization unit that formulates the energy management plan by solving the optimization problem in order using the solution-finding means with high priority determined by the solution-finding means determination unit using the constraint conditions determined by the constraint condition determination unit; and An energy operation planning device comprising:

2. 1. An energy operation planning device that formulates an energy operation plan that determines control command values ​​for devices that generate energy including electric power or heat at each time, a problem information input unit into which variables used to determine an objective function and constraint conditions for the optimization problem of the energy management plan and a desired calculation time required for optimizing the energy management plan are input; a constraint condition determination unit that estimates the strength of the constraint condition to be used in the optimization problem from the variables and the desired required calculation time input by the problem information input unit using a trained model that estimates the strength of the constraint condition from the variables and the required calculation time, and determines the constraint condition to be used in the optimization problem; an energy management plan optimization unit that formulates the energy management plan by solving the optimization problem using the constraints determined by the constraint condition determination unit and simultaneously using a plurality of solution means; An energy operation planning device comprising:

3. The energy operation planning device according to claim 1 or 2, further comprising a constraint condition learning unit that generates the learned model that determines the strength of the constraint condition from the variables used in calculating the optimization problem and the calculation time required.

4. The energy operation planning device according to claim 1 , further comprising a solution-finding means learning unit that generates the trained model that determines the priority of solution-finding means based on variables used in calculating the optimization problem and the calculation time required.

5. 3. The energy operation planning device according to claim 1, wherein the objective function of the optimization problem includes at least one of the cost related to fuel usage, the cost related to electricity usage, the amount of primary energy usage, and the amount of greenhouse gas emissions in the energy operation plan.

6. The equipment used in the energy operation plan includes at least one of a fuel storage facility, a heat source machine, a generator, a power storage facility, a heat storage facility, and a power system. The energy operation planning device according to claim 1 or 2.

7. the solution-finding means determination unit determines in advance a priority order of solution-finding means to be used for the optimization problem; the constraint condition determination unit repeats a process of determining the constraint condition to be used in the optimization problem according to the strength of the constraint condition, and the energy management plan optimization unit repeats a process of executing a calculation for the optimization problem using the solution-finding means having the highest priority until the required calculation time input to the problem information input unit is reached; The energy operation planning device according to claim 1 or 4, wherein the constraint condition determination unit selects the constraint conditions in ascending order of weakness.

8. the constraint condition determination unit determines in advance the constraint conditions to be used in the optimization problem according to the strength of the constraint conditions, and the solution-finding means determination unit determines in advance the priority of the solution-finding means to be used in the optimization problem, 2. The energy operation planning device according to claim 1, wherein the energy operation plan optimization unit repeats the process of performing calculations for the optimization problem using the solution-finding means selected according to the priority order until the required calculation time inputted to the problem information input unit is reached.

9. the constraint condition determination unit repeats a process of determining the constraint conditions to be used in the optimization problem according to the strength of the constraint conditions, and the energy management plan optimization unit repeats a process of simultaneously using a plurality of solution-finding means to perform calculations for the optimization problem until the required calculation time input to the problem information input unit is reached; The energy operation planning device according to claim 2 , wherein the constraint condition determination unit selects the constraint conditions in ascending order of weakness.

10. 1. An energy operation planning device that formulates an energy operation plan that determines control command values ​​for devices that generate energy including electric power or heat at each time, a problem information input step of inputting variables used to determine an objective function and constraint conditions for the optimization problem of the energy management plan and a desired calculation time required for optimizing the energy management plan; a constraint determination step of estimating the strength of the constraint to be used in the optimization problem from the variables and the desired required calculation time input in the problem information input step using a trained model that estimates the strength of the constraint from the variables and the required calculation time, and determining the constraint to be used in the optimization problem; a solution-finding means determination step of determining a priority order of solution-finding means to be used for the optimization problem from the variables and the desired required calculation time input in the problem information input step, using a trained model that determines a priority order of solution-finding means from the variables and the required calculation time; an energy operation plan optimization step of formulating the energy operation plan by solving the optimization problem in order using the solution-finding means with high priorities determined in the solution-finding means determination step using the constraint conditions determined in the constraint condition determination step; An energy operation planning method comprising:

11. 1. An energy operation planning device that formulates an energy operation plan that determines control command values ​​for devices that generate energy including electric power or heat at each time, a problem information input step of inputting variables used to determine an objective function and constraint conditions for the optimization problem of the energy management plan and a desired calculation time required for optimizing the energy management plan; a constraint determination step of estimating the strength of the constraint to be used in the optimization problem from the variables and the desired required calculation time input in the problem information input step using a trained model that estimates the strength of the constraint from the variables and the required calculation time, and determining the constraint to be used in the optimization problem; an energy operation plan optimization step of formulating the energy operation plan by solving the optimization problem using the constraints determined in the constraint condition determination step and simultaneously using a plurality of solution means; An energy operation planning method comprising:

Images