Linearization processing method, device and storage medium for nonlinear model

Abstract

A linearization processing method, device, and storage medium for a nonlinear model. The method includes: determining the value range of each input parameter of each device's nonlinear model (101); dividing the value range of each input parameter into a plurality of subintervals based on a plurality of interpolation points (102); determining a plurality of input sample values in each subinterval (103); traversing the input sample value combinations of each input parameter of the model, and obtaining the output sample value combination corresponding to each input sample value combination using the nonlinear model (104); and generating a tensor table (105) using all input sample value combinations and their corresponding output sample value combinations. The above method can realize linearization processing of nonlinear models for some models including nonlinear physical processes.

Description

Technical Field

[0001] The present invention relates to the industrial field, and in particular to a linearization processing method, device and computer-readable storage medium for a nonlinear model in an integrated energy system. Background Art

[0002] Distributed energy systems (DES) are considered an effective solution to addressing the volatile consumption of renewable energy. DES are being deployed around the world, including in my country, and the demand for operational optimization models and the overall system energy production and utilization is increasing. Previous researchers have developed specific operational optimization models that are applicable only to systems with specific energy sources and components. Therefore, these models, in practical implementation, do not meet the requirements of the emerging integrated energy services landscape.

[0003] Therefore, it is necessary to provide a universal integrated energy system that fully utilizes various resource forms such as renewable energy, fossil fuels, residual heat and pressure, and new energy, and coordinates them. Through the flexible operation of sources, networks, loads, and storage, innovative business mechanisms can be established and intelligent means can be used to achieve high-quality, high-efficiency, and most economical and environmentally friendly regional supply of multiple loads such as electricity, heat, cooling, and gas, meeting the requirements of random fluctuations in terminal loads. Integrated energy promotes the absorption capacity of renewable energy and improves the comprehensive utilization rate of energy.

[0004] However, implementing this integrated energy system requires modeling many devices. These devices often involve numerous nonlinear physical processes (also known as fundamental processes), such as the flow-pressure relationship in a gas turbine compressor and the conversion of mechanical energy into pressure energy. Therefore, the models for these devices are typically nonlinear. Nonlinear models are often complex but offer high accuracy. However, directly using these nonlinear models to simulate the integrated energy system can affect the real-time performance of the entire simulation due to their relatively slow execution speed. Summary of the Invention

[0005] In view of this, the embodiments of the present invention propose, on the one hand, a linearization processing method for a nonlinear model, and on the other hand, a linearization processing device for a nonlinear model and a computer-readable storage medium, which are used to realize linearization processing of nonlinear models for some models that include nonlinear basic processes, and then used for the construction of an integrated energy system.

[0006] A linearization processing method for a nonlinear model proposed in an embodiment of the present invention includes: determining the value range of each input parameter of the nonlinear model of each device; dividing the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points; evenly determining a plurality of input sample values in each sub-interval; traversing the input sample value combinations of each input parameter of the model, and using the nonlinear model to obtain the output sample value combination corresponding to each input sample value combination; and generating a tensor table using all input sample value combinations and their corresponding output sample value combinations.

[0007] In one embodiment, dividing the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points includes: dividing the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points based on a balance criterion.

[0008] In one embodiment, the balancing and determining the plurality of input sample values in each subinterval is: balancing and determining the plurality of input sample values in each subinterval based on a balancing criterion.

[0009] In one embodiment, when performing simulation, the tensor table is searched according to the current value of each input parameter, and the corresponding data found in the tensor table are interpolated to obtain the corresponding output value.

[0010] In one embodiment, the nonlinear model of each device is modeled by the following method: for each target nonlinear basic process of each device, its complete design point data is determined; a description formula of the nonlinear basic process is established by using the ratio of the similarity number supported by the similarity criterion to the similarity number based on the design point data, and a universal model of the nonlinear basic process is obtained; the universal model includes variable parameters that change nonlinearly with the change of actual operating parameters; a machine learning algorithm is constructed between the actual operating parameters and the variable parameters, and an association relationship is established between the machine learning algorithm and the universal model; the universal model of all target nonlinear universal processes of each device and its associated machine learning algorithm are obtained. The machine learning algorithm constitutes a general model of the equipment; for each target nonlinear basic process of a specific equipment of the equipment, the historical data of actual operating parameters and variable parameters corresponding to the target nonlinear basic process of the specific equipment are obtained, and the machine learning algorithm is trained using the historical data to obtain a variable parameter training model of the target nonlinear basic process; the variable parameter training model of the target nonlinear basic process is substituted into the general model of the target nonlinear basic process to obtain a trained model of the target nonlinear basic process of the specific equipment; the trained models of all target nonlinear basic processes of the specific equipment constitute the trained model of the specific equipment.

[0011] In one embodiment, the variable parameter has a preset default value.

[0012] In one embodiment, the equipment includes: a gas turbine, a heat pump, an internal combustion engine, a steam turbine, a waste heat boiler, an absorption refrigerator, a heating machine, a multi-effect evaporator, water electrolysis to produce hydrogen, hydrogen to chemical equipment, reverse osmosis, a fuel cell, and a boiler; the target nonlinear basic process of each device includes one or more of the following processes: a heat transfer process, a process of converting thermal energy into kinetic energy, a pipeline resistance process, a process related to flow and pressure, a process of converting thermal energy into mechanical energy, a process of converting electrical energy into cooling, a thermal energy process, a distillation process, an evaporation process, and a filtration process.

[0013] A linearization processing device for a nonlinear model proposed in an embodiment of the present invention includes: a first processing module, used to determine the value range of each input parameter of the model for the nonlinear model of each device; a second processing module, used to divide the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points; a third processing module, used to evenly determine a plurality of input sample values in each sub-interval; a fourth processing module, used to traverse the input sample value combinations of each input parameter of the model, and use the nonlinear model to obtain the output sample value combination corresponding to each input sample value combination; and a fifth processing module, used to generate a tensor table using all input sample value combinations and their corresponding output sample value combinations.

[0014] In one embodiment, the second processing module divides the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points based on a balance criterion.

[0015] In one embodiment, the third processing module determines a plurality of input sample values in each subinterval in a balanced manner based on a balanced criterion.

[0016] In one embodiment, the system further includes: a sixth processing module, configured to perform interpolation processing on the tensor table according to current values of various input parameters to obtain corresponding output values during simulation.

[0017] In one embodiment, the invention further comprises: a first modeling module for determining the complete design point data for each target nonlinear basic process of each device; establishing a description formula of the nonlinear basic process by using the ratio of the similarity number supported by the similarity criterion to the similarity number based on the design point data, and obtaining a general model of the nonlinear basic process; the general model includes variable parameters that change nonlinearly with the change of actual operating parameters; constructing a machine learning algorithm between the actual operating parameters and the variable parameters, and establishing an association relationship between the machine learning algorithm and the general model; the general model of all target nonlinear general processes of each device and its associated machine learning algorithm constitute the general model. A general model of a type of equipment; and a second modeling module, which is used to obtain the historical data of actual operating parameters and variable parameters corresponding to the target nonlinear basic process of a specific device of the type of equipment for each target nonlinear basic process of the specific device, and use the historical data to train the machine learning algorithm to obtain a variable parameter training model of the target nonlinear basic process; substitute the variable parameter training model of the target nonlinear basic process into the general model of the target nonlinear basic process to obtain a trained model of the target nonlinear basic process of the specific device; the trained models of all target nonlinear basic processes of the specific device constitute the trained model of the specific device.

[0018] Another linearization processing device for a nonlinear model proposed in an embodiment of the present invention includes: at least one memory and at least one processor, wherein: the at least one memory is used to store a computer program; the at least one processor is used to call the computer program stored in the at least one memory to execute the linearization processing method for the nonlinear model described in any of the above embodiments.

[0019] The computer-readable storage medium proposed in the embodiment of the present invention stores a computer program thereon; the computer program can be executed by a processor and implement the linearization processing method described in any of the above embodiments.

[0020] As can be seen from the above scheme, the present invention utilizes piecewise linearization technology to process the nonlinear model, generating a tensor table. During simulation, interpolation operations are performed based on this tensor table to obtain the required simulation data. This linearized device model runs faster, meeting the real-time requirements of simulation.

[0021] Furthermore, when modeling the equipment model, a descriptive formula for each target nonlinear fundamental process of each type of equipment is established by using the ratio of the similarity number supported by the similarity criterion to the similarity number based on the design point data. This results in a universal model for the nonlinear fundamental process, allowing the model to be applied to a class of equipment as a universal model. Furthermore, the universal model includes variable parameters that vary nonlinearly with actual operating parameters, and by constructing a machine learning algorithm between the actual operating parameters and the variable parameters, the variable parameters can be obtained through machine learning, thereby endowing the universal model with self-learning capabilities.

[0022] In addition, for a specific device of this type of equipment, by obtaining historical data of actual operating parameters and variable parameters corresponding to the target nonlinear basic process of the specific device for each of its target nonlinear basic processes, and using the historical data to train the machine learning algorithm, a variable parameter training model of the target nonlinear basic process is obtained, and the variable parameter training model of the target nonlinear basic process is substituted into the general model of the target nonlinear basic process, so as to obtain a trained model of the target nonlinear basic process of the specific device, that is, an instantiated model that conforms to the characteristics of the specific device.

[0023] Furthermore, by presetting default values for the variable parameters, the nonlinear model can be made usable even when the variable parameters are not available for training on site, for example, when there is insufficient historical data.

[0024] Finally, the modeling method in the embodiment of the present invention can be applied to various nonlinear processes of various devices, which is not only convenient to implement but also has high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, so that those skilled in the art will understand the above and other features and advantages of the present invention more clearly. In the accompanying drawings:

[0026] Figure 1 This is an exemplary flow chart of a linearization processing method for a nonlinear model in an embodiment of the present invention.

[0027] Figure 2 This is an exemplary flow chart of a modeling method for a nonlinear model in an embodiment of the present invention.

[0028] Figure 3 This is an exemplary structural diagram of a linearization processing device for a nonlinear model in an embodiment of the present invention.

[0029] Figure 4 This is an exemplary structural diagram of another linearization processing device for a nonlinear model in an embodiment of the present invention.

[0030] Figure 5 This is an exemplary structural diagram of another linearization processing device for a nonlinear model in an embodiment of the present invention.

[0031] The accompanying drawings are numerals as follows:

[0032] Label meaning 101-105，201-207 step

[0033] 301 First processing module 302 Second processing module 303 The third processing module 304 Fourth processing module 305 Fifth processing module 306 Sixth processing module 307 The first modeling module 308 Second modeling module 51 Memory 52 processor 53 bus DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail with reference to the following examples.

[0035] Figure 1 FIG. 1 is an exemplary flow chart of a linearization processing method for a nonlinear model in an embodiment of the present invention. Figure 1 As shown, the method may include the following steps:

[0036] Step 101: for each device's nonlinear model, determine the value range of each input parameter of the model.

[0037] For example, the effective power range is 50% to 110% of the rated operating conditions. In addition, there is also the range of local ambient temperature changes, the range of ambient pressure changes, etc.

[0038] Step 102: Divide the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points.

[0039] In this step, the value range of each input parameter can be divided into multiple sub-intervals based on multiple interpolation points based on the balance criterion. The interpolation points can be determined by setting more interpolation points for areas with drastic nonlinear changes and setting fewer interpolation points for areas with slow nonlinear changes.

[0040] For example, the power range is inserted into 40 points, the ambient temperature is inserted into 20 points, the ambient pressure is inserted into 5 points, and so on.

[0041] Step 103: Balance and determine a plurality of input sample values in each subinterval.

[0042] In this step, a plurality of input sample values may be determined to be balanced within each subinterval based on a balance criterion.

[0043] For example, power and ambient temperature can be divided equally in the actual domain.

[0044] Step 104 , traverse the input sample value combinations of the input parameters of the model, and use the nonlinear model to obtain the output sample value combination corresponding to each input sample value combination.

[0045] For example, for each set of input sample values obtained by traversal, there is a corresponding set of outputs, such as efficiency output, or fuel consumption, emission output, operating cost output, etc.

[0046] Step 105: Generate a tensor table using all input sample value combinations and their corresponding output sample value combinations.

[0047] For example, if the values of temperature, pressure, and power are known as mentioned above, the efficiency value can be obtained by interpolation by looking up the tensor table.

[0048] Specifically, when using the model of the device for simulation, the tensor table can be searched based on the current value of each input parameter, and the corresponding value found in the tensor table can be interpolated to obtain the corresponding output value. The current value of each input parameter can be a real value or a hypothetical value.

[0049] For example, a device may have one or more tables, such as a table that maps temperature, pressure, and power to efficiency, a table that maps temperature, pressure, and power to emissions, or any other desired parameter. The interpolation algorithm can be selected based on the actual situation, for example, linear interpolation or nonlinear interpolation can be used. In one example, linear interpolation can be used for closely adjacent points, while nonlinear interpolation can be used for more distant points.

[0050] The values of other output variables corresponding to the required temperature, pressure, and performance can be obtained through spline interpolation of three-dimensional temperature, pressure, and power.

[0051] This general approach uses a general program, and any specific model, such as heat pump, internal combustion engine, heat exchanger, etc., can be processed by this code tool.

[0052] Figure 2 FIG. 1 is an exemplary flow chart of a method for modeling a nonlinear model in an embodiment of the present invention. Figure 2 As shown, the method may include the following steps:

[0053] Step 201: determine complete design point data for each target nonlinear basic process of each device.

[0054] In this step, fundamental processes are sometimes referred to as physical processes, such as heat transfer processes, electrical energy conversion processes, and the aforementioned processes related to flow and pressure. For each device, the fundamental processes of interest, i.e., the fundamental processes that need to be modeled, can be determined. These fundamental processes that need to be modeled are referred to as target fundamental processes, and nonlinear target fundamental processes are referred to as target nonlinear fundamental processes. For example, for a gas turbine, target nonlinear fundamental processes may include: processes related to flow and pressure in an expansion turbine, processes for converting thermal energy into mechanical energy, etc.; for a heat pump, target nonlinear fundamental processes may include: heat transfer processes, processes for converting electrical energy into thermal energy, chemical processes for separating solution substances using high-temperature thermal energy, electrochemical processes, pipeline resistance processes, processes related to flow and pressure, etc. In addition, for equipment such as internal combustion engines, steam turbines, waste heat boilers, absorption refrigerators, heating machines, multi-effect evaporators, water electrolysis hydrogen production, hydrogen production of chemicals equipment, reverse osmosis, fuel cells, boilers, etc., the target nonlinear basic process of each device may include: flow and pressure related processes, thermal energy to mechanical energy processes, electrical energy to cooling, thermal energy processes, pipeline resistance processes, heat transfer processes, distillation processes, evaporation processes, filtration processes, chemical reaction processes, electrochemical processes, etc. One or more of the following.

[0055] For each nonlinear fundamental process, complete design point data can be restored based on publicly available design parameters and common design point information provided by the manufacturer. For example, for a general model of a process related to flow and pressure, the design point data may include pressure ratio and air flow rate. From this design point data, relevant design parameters not available to the user, such as efficiency, inlet resistance, and air extraction capacity, can be derived.

[0056] In step 202, a description formula of the nonlinear basic process is established by using the ratio of the similarity number supported by the similarity criterion to the similarity number based on the design point data, thereby obtaining a general model of the nonlinear basic process; the general model includes variable parameters that change nonlinearly with changes in actual operating parameters.

[0057] Actual operating parameters refer to parameters of a specific device that vary with actual operating parameters. Examples include dimensional changes due to mechanical wear over time, seasonal temperature changes, or parameters that vary with different operating conditions. These variable parameters may have pre-set default values.

[0058] Since there may be different models for each device, for example, taking a compressor as an example, there may be compressors with different powers such as 5M, 50M, and 500M, so in order to establish a universal model for the compressor, it is necessary to use similarity numbers supported by the similarity criterion to replace specific parameter values. For example, still taking the universal model of the flow and pressure correlation process mentioned above as an example, similarity numbers supported by the similarity criterion of flow, pressure, and power are used to replace specific parameters. For example, the similarity criterion of flow can be shown as follows (1):

[0059]

[0060] Among them, G1 is flow rate, T1 is temperature, P1 is pressure, G0 is the flow rate of the corresponding design point, T0 is the temperature of the corresponding design point, and P0 is the pressure of the corresponding design point.

[0061] Accordingly, the general model of the correlation process between flow and pressure can be expressed as follows:

[0062]

[0063] Where f() is a function, and coefficients a and b are variable parameters that change nonlinearly with actual operating parameters. In practical applications, default values can also be set for the variable parameters a and b. IGV is the inlet guide vane angle.

[0064] Step 203: construct a machine learning algorithm between the actual operating parameters and the variable parameters.

[0065] In this step, a machine learning algorithm between the actual operating parameters and the variable parameters can be constructed based on a machine learning big data analysis method such as a neural intelligent network or a support vector machine.

[0066] In step 204 , the general model of all target nonlinear basic processes of each type of equipment and its associated machine learning algorithm constitutes a general model of the equipment.

[0067] It can be seen that through the above process, a nonlinear universal model for each device can be established. Based on these universal models, an integrated energy system platform can be constructed.

[0068] In actual applications, after purchasing the integrated energy system platform, users need to build their own integrated energy system. At this time, each universal model needs to be associated with specific equipment on site, so the universal model needs to be instantiated. Accordingly, the method can further include the following steps:

[0069] Step 205: For each target nonlinear basic process of a specific device of the device, obtain historical data of actual operating parameters and variable parameters corresponding to the target nonlinear basic process of the specific device, use the historical data to train the corresponding machine learning algorithm, and obtain a variable parameter training model of the target nonlinear basic process.

[0070] In this step, during specific training, a set of historical data of actual operating parameters is used as input sample values, and the historical data of variable parameters corresponding to the set of historical data of actual operating parameters is used as output sample values. By using a large number of input sample values and corresponding output sample values to train the machine learning algorithm, a self-learning model of the variable parameters, also called a training model, can be obtained.

[0071] For example, still taking the above-mentioned flow and pressure correlation process as an example, the historical data of the actual operating parameters of the on-site gas turbine and the historical data of its corresponding variable parameters can be obtained to obtain the input and output sample sets, and after training, the training model of the variable parameters a and b can be obtained.

[0072] Step 206: Substitute the variable parameter training model of the target nonlinear basic process into the universal model of the target nonlinear basic process to obtain a trained model of the target nonlinear basic process of the specific device. The trained model is a self-learning model with learning ability.

[0073] In this step, the variable parameter training model of the target nonlinear basic process can be substituted into the general model of the target nonlinear basic process according to the association relationship between the machine learning algorithm and the general model.

[0074] For example, still taking the above-mentioned process of flow and pressure correlation as an example, by inputting the current training model of the variable parameters a and b into the above formula (2), a general model of the process of flow and pressure correlation of the compressor of the on-site gas turbine can be obtained.

[0075] Step 207: The trained models of all target nonlinear basic processes of the specific device constitute the trained model of the specific device.

[0076] In actual use, the input parameters of the trained model may include the input parameters required by the trained model of all target nonlinear basic processes.

[0077] The above describes in detail a method for linearizing a nonlinear model and one modeling method thereof in embodiments of the present invention. Next, a device for linearizing a nonlinear model and one modeling device thereof in embodiments of the present invention are described in detail. The device in embodiments of the present invention can be used to implement the method in embodiments of the present invention. Details not disclosed in the device embodiments of the present invention are referred to the corresponding description in the method embodiments of the present invention and will not be detailed here.

[0078] Figure 3 FIG. 1 is an exemplary structural diagram of a linear processing device for a nonlinear model according to an embodiment of the present invention. Figure 3 As shown, the apparatus may include: a first processing module 301 , a second processing module 302 , a third processing module 303 , a fourth processing module 304 and a fifth processing module 305 .

[0079] The first processing module 301 is used to determine the value range of each input parameter of the nonlinear model of each device.

[0080] The second processing module 302 divides the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points. In specific implementation, the second processing module 302 may divide the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points based on a balance criterion.

[0081] The third processing module 303 is configured to determine a plurality of input sample values in a balanced manner within each subinterval. In specific implementation, the third processing module 303 may determine a plurality of input sample values in a balanced manner within each subinterval based on a balanced criterion.

[0082] The fourth processing module 304 is used to traverse the input sample value combinations of various input parameters of the model, and obtain the output sample value combination corresponding to each input sample value combination using the nonlinear model.

[0083] The fifth processing module 305 is configured to generate a tensor table using all input sample value combinations and their corresponding output sample value combinations.

[0084] In other embodiments, the linearization processing device of the nonlinear model can be as follows: Figure 3 The dotted part in FIG further includes: a sixth processing module 306, which is used to perform interpolation processing on the tensor table according to the current value of each input parameter to obtain a corresponding output value during simulation.

[0085] Figure 4 FIG. 1 is an exemplary structural diagram of another linear processing device for a nonlinear model in an embodiment of the present invention. Figure 4 As shown, the device can be Figure 3Based on the device shown, it further includes: a first modeling module 307 and a second modeling module 308.

[0086] Among them, the first modeling module 307 is used to determine the complete design point data for each target nonlinear basic process of each equipment; adopt the ratio of the similarity number supported by the similarity criterion to the similarity number based on the design point data to establish a description formula of the nonlinear basic process and obtain a general model of the nonlinear basic process; the general model includes variable parameters that change nonlinearly with the change of actual operating parameters; construct a machine learning algorithm between the actual operating parameters and the variable parameters, and establish an association relationship between the machine learning algorithm and the general model; the general model of all target nonlinear general processes of each equipment and its associated machine learning algorithm constitute the general model of this type of equipment.

[0087] The second modeling module 308 is used to obtain the historical data of actual operating parameters and variable parameters corresponding to each target nonlinear basic process of a specific device of this type of equipment, use the historical data to train the machine learning algorithm to obtain a variable parameter training model of the target nonlinear basic process; substitute the variable parameter training model of the target nonlinear basic process into the general model of the target nonlinear basic process to obtain the trained model of the target nonlinear basic process of the specific device; the trained models of all target nonlinear basic processes of the specific device constitute the trained model of the specific device.

[0088] Figure 5 FIG. 1 is a schematic structural diagram of another linear processing device for a nonlinear model according to an embodiment of the present invention. Figure 5 As shown, the system may include at least one memory 51 and at least one processor 52. In addition, other components may be included, such as communication ports, etc. These components communicate via a bus 53.

[0089] Wherein: at least one memory 51 is used to store a computer program. In one embodiment, the computer program can be understood to include Figure 3 or Figure 4 In addition, at least one memory 51 can also store an operating system, etc. Operating systems include but are not limited to: Android operating system, Symbian operating system, Windows operating system, Linux operating system, etc.

[0090] At least one processor 52 is configured to call a computer program stored in at least one memory 51 to execute the linearization processing method for a nonlinear model described in an embodiment of the present invention. The processor 52 may be a CPU, a processing unit / module, an ASIC, a logic module, or a programmable gate array. The processor 52 may receive and transmit data via the communication port.

[0091] It should be noted that not all steps and modules in the above processes and structure diagrams are required, and certain steps or modules can be omitted based on actual needs. The execution order of the steps is not fixed and can be adjusted as needed. The division of the modules is merely for the convenience of describing the functional division adopted. In actual implementation, a module can be implemented by multiple modules, and the functions of multiple modules can be implemented by the same module. These modules can be located in the same device or in different devices.

[0092] It is understood that the hardware modules in the above-mentioned embodiments can be implemented mechanically or electronically. For example, a hardware module may include a specially designed permanent circuit or logic device (such as a dedicated processor, such as an FPGA or ASIC) for performing a specific operation. The hardware module may also include a programmable logic device or circuit (such as a general-purpose processor or other programmable processor) temporarily configured by software to perform a specific operation. As for whether to implement the hardware module mechanically, or using a dedicated permanent circuit, or using a temporarily configured circuit (such as configured by software), it can be decided based on cost and time considerations.

[0093] In addition, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, the computer program being executable by a processor and implementing the linearization processing method for a nonlinear model described in the embodiment of the present invention. Specifically, a system or device equipped with a storage medium can be provided, wherein the storage medium stores software program code that implements the functions of any of the above embodiments, and the computer (or CPU or MPU) of the system or device reads and executes the program code stored in the storage medium. In addition, the operating system, etc., operating on the computer can be caused to perform some or all of the actual operations based on instructions based on the program code. The program code read from the storage medium can also be written to a memory provided in an expansion board inserted into the computer or to a memory provided in an expansion unit connected to the computer, and then, based on the instructions of the program code, the CPU, etc. installed on the expansion board or expansion unit can be caused to perform some or all of the actual operations, thereby implementing the functions of any of the above embodiments. The storage medium embodiment for providing the program code includes a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code may be downloaded from a server computer via a communications network.

[0094] As can be seen from the above scheme, the present invention utilizes piecewise linearization technology to process the nonlinear model, generating a tensor table. During simulation, interpolation operations are performed based on this tensor table to obtain the required simulation data. This linearized device model runs faster, meeting the real-time requirements of simulation.

[0095] When modeling equipment, for each target nonlinear fundamental process of each type of equipment, a description formula for the nonlinear fundamental process is established by using the ratio of the similarity number supported by the similarity criterion to the similarity number based on the design point data. This results in a universal model for the nonlinear fundamental process, making the model applicable to a class of equipment. Furthermore, the universal model includes variable parameters that vary nonlinearly with actual operating parameters. By constructing a machine learning algorithm between the actual operating parameters and the variable parameters, the variable parameters can be obtained through machine learning, thereby endowing the model with self-learning capabilities.

[0096] In addition, for a specific device of this type of equipment, by obtaining historical data of actual operating parameters and variable parameters corresponding to the target nonlinear basic process of the specific device for each of its target nonlinear basic processes, and using the historical data to train the machine learning algorithm, a variable parameter training model of the target nonlinear basic process is obtained, and the variable parameter training model of the target nonlinear basic process is substituted into the general model of the target nonlinear basic process, so as to obtain a trained model of the target nonlinear basic process of the specific device, that is, an instantiated model that conforms to the characteristics of the specific device.

[0097] Furthermore, by presetting default values for the variable parameters, the nonlinear model can be made usable even when the variable parameters are not available for training on site, for example, when there is insufficient historical data.

[0098] Finally, the modeling method in the embodiment of the present invention can be applied to various nonlinear processes of various devices, which is not only convenient to implement but also has high accuracy.

[0099] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A linearization processing method for a nonlinear model in an integrated energy system, characterized in that: include: For each device's nonlinear model, determine a value range of each input parameter of the model (101); Divide the value range of each input parameter into a plurality of subintervals based on a plurality of interpolation points (102); Evenly determining a plurality of input sample values within each subinterval (103); Traversing the input sample value combinations of the input parameters of the model, and using the nonlinear model to obtain the output sample value combination corresponding to each input sample value combination (104); and Generate a tensor table (105) using all input sample value combinations and their corresponding output sample value combinations; The nonlinear model of each device is modeled using the following method: Determine complete design point data (201) for each target nonlinear basic process of each equipment; A description formula of the nonlinear basic process is established by using a ratio of a similarity number supported by the similarity criterion to a similarity number based on design point data, thereby obtaining a general model of the nonlinear basic process; the general model includes variable parameters (202) that change nonlinearly with changes in actual operating parameters; Constructing a machine learning algorithm between the actual operating parameters and the variable parameters, and establishing an association relationship between the machine learning algorithm and the general model (203); The general model of all target nonlinear general processes of each device and its associated machine learning algorithm constitutes the general model of the device (204); For each target nonlinear basic process of a specific device of the device, obtaining historical data of actual operating parameters and variable parameters corresponding to the target nonlinear basic process of the specific device, using the historical data to train the machine learning algorithm to obtain a variable parameter training model of the target nonlinear basic process (205); and Substituting the variable parameter training model of the target nonlinear basis process into the general model of the target nonlinear basis process to obtain a trained model of the target nonlinear basis process of the specific device (206); The trained models of all target nonlinear basic processes of the specific device constitute the trained model (207) of the specific device.

2. The linearization processing method of a nonlinear model according to claim 1, characterized in that: The dividing the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points is: based on a balance criterion, dividing the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points.

3. The linearization processing method of a nonlinear model according to claim 1, characterized in that: The method of determining the multiple input sample values in a balanced manner within each subinterval comprises: determining the multiple input sample values in a balanced manner within each subinterval based on a balanced criterion.

4. The linearization processing method of a nonlinear model according to any one of claims 1 to 3, characterized in that: When performing simulation, the tensor table is searched according to the current value of each input parameter, and the corresponding data found in the tensor table are used for interpolation processing to obtain the corresponding output value.

5. The nonlinear modeling method according to claim 1, wherein: The variable parameters have preset default values.

6. The nonlinear modeling method according to claim 1 or 5, characterized in that: The equipment includes: gas turbines, heat pumps, internal combustion engines, steam turbines, waste heat boilers, absorption refrigerators, heating machines, multi-effect evaporators, water electrolysis hydrogen production, hydrogen-to-chemical equipment, reverse osmosis, fuel cells, boilers: The target nonlinear basic process of each device includes one or more of the following processes: heat transfer process, process of converting thermal energy into kinetic energy, pipeline resistance process, flow and pressure related process, thermal energy into mechanical energy process, electrical energy into cooling, thermal energy process, distillation process, evaporation process and filtration process.

7. A linearization processing device for a nonlinear model in an integrated energy system, characterized in that: include: A first processing module (301) is used to determine the value range of each input parameter of the nonlinear model of each device; A second processing module (302) divides the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points; A third processing module (303) is used to determine a plurality of input sample values in each subinterval in a balanced manner; A fourth processing module (304) is used to traverse the input sample value combinations of each input parameter of the model, and obtain the output sample value combination corresponding to each input sample value combination using the nonlinear model; and A fifth processing module (305) is configured to generate a tensor table using all input sample value combinations and their corresponding output sample value combinations; Further comprising: a first modeling module (307), for determining complete design point data for each target nonlinear basic process of each type of equipment; establishing a description formula of the nonlinear basic process by using a ratio of a similarity number supported by a similarity criterion to a similarity number based on the design point data, and obtaining a general model of the nonlinear basic process; the general model includes variable parameters that change nonlinearly with changes in actual operating parameters; constructing a machine learning algorithm between the actual operating parameters and the variable parameters, and establishing an association relationship between the machine learning algorithm and the general model; the general model of all target nonlinear general processes of each type of equipment and the associated machine learning algorithm constitute the general model of the equipment; and The second modeling module (308) is used to obtain the historical data of actual operating parameters and variable parameters corresponding to each target nonlinear basic process of a specific device of the device, and use the historical data to train the machine learning algorithm to obtain a variable parameter training model of the target nonlinear basic process; substitute the variable parameter training model of the target nonlinear basic process into the general model of the target nonlinear basic process to obtain a trained model of the target nonlinear basic process of the specific device; the trained models of all target nonlinear basic processes of the specific device constitute the trained model of the specific device.

8. The linearization processing device for a nonlinear model according to claim 7, characterized in that: The second processing module divides the value range of each input parameter into a plurality of sub-intervals based on a plurality of interpolation points based on a balance criterion.

9. The linearization processing device for a nonlinear model according to claim 7, characterized in that: The third processing module determines a plurality of input sample values in a balanced manner within each subinterval based on a balanced criterion.

10. The linearization processing device for a nonlinear model according to any one of claims 7 to 9, characterized in that: Further including: The sixth processing module (306) is used to perform interpolation processing on the tensor table according to the current value of each input parameter during simulation to obtain a corresponding output value.

11. A linearization processing device for a nonlinear model in an integrated energy system, characterized in that: include: At least one memory (51) and at least one processor (52), wherein: The at least one memory (51) is used to store a computer program; The at least one processor (52) is configured to call a computer program stored in the at least one memory (51) to execute the linearization processing method for a nonlinear model according to any one of claims 1 to 6.

12. A computer-readable storage medium having a computer program stored thereon; characterized in that: The computer program can be executed by a processor and implements the linearization processing method of a nonlinear model according to any one of claims 1 to 6.

Images