Simulation system

Abstract

To provide a simulation system capable of estimating any parameter of the status of components based on measurement data obtained from a sensor that measures the status of the component constituting a device.SOLUTION: A simulation unit 2 for reproducing the behavior of a system 1 includes a modeling unit 5 that builds a governing equation for any status parameter of a component 9 of the system 1. The modeling unit 5 builds a governing equation based on measurement data obtained from a sensor 4 that measures the status quantity of the component 9 and a basis function for the governing equation and estimates the status parameters by using the governing equation.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a simulation device that reproduces the operation of a device. [Background technology]

[0002] When controlling a machine, a method is used in which a digital twin is constructed to reproduce the operation of the machine through a simulation, the operation of the machine is estimated using the digital twin, and optimal control is realized. In order to estimate the operating state of the machine quickly and accurately using a simulation, it is necessary to faithfully simulate the physical phenomena such as the operation and reaction of the machine and its constituent parts. For example, when estimating the battery temperature in a power supply device using a lithium-ion battery, a thermal fluid simulation that models the structure of the device is performed to calculate the air temperature, flow field, and component temperature inside the device, and it is possible to estimate the battery temperature. Here, if a non-steady 3D thermal fluid analysis is performed as a thermal fluid simulation, it is possible to estimate the temperature distribution and flow field of the entire device and their time changes. However, when performing a detailed simulation, the calculation load is large, a high-speed processor is required, and a large amount of data must be handled. Therefore, a slow processor installed on the board inside the device takes a long time to estimate, and there is a problem that it cannot track the actual phenomenon.

[0003] In response to these issues, the incorporation of a calculation system that performs simulations using a simplified physical model has been considered, and Patent Document 1 describes "a calculation system having a storage device in which a simulation model including at least a physical model that outputs polishing physical quantities is stored."

[0004] In addition, application of reduced-order simulation as a simulation model is being considered. For example, Patent Document 2 states that "a reduced-order simulation is performed to express the system using a linear sum, which is the sum of products of basis vectors and corresponding weighting coefficients." [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent Publication No. 2021-146493 [Patent Document 2] Patent Publication No. 2014-26440 Summary of the Invention [Problem to be solved by the invention]

[0006] In Patent Document 1, a calculation system having a storage device in which a simulation model is stored is provided in a polishing apparatus, and the calculation system is used to estimate an estimated polishing physical quantity from actual measurement data. However, in a simulation using a pre-stored physical model, if a parameter to be estimated is changed, the physical model must be re-stored.

[0007] In Patent Document 2, a reduced-order simulation unit performs a simulation in advance using a linear sum of multiple basis vectors (basis functions) and their corresponding weighting coefficients, making it possible to perform an analysis with a desired degree of freedom without reducing the analytical degree of freedom of the analysis target. However, it is necessary to build a reduced-order simulation model in advance and store it in the device, which gives rise to the same problems as in Patent Document 1.

[0008] The present invention has been made to solve the above-mentioned problems, and has an object to provide a simulation device capable of estimating any state parameter of a part constituting an apparatus, based on measurement data obtained from a sensor that measures a state quantity of the part. [Means for solving the problem]

[0009] In order to achieve the above object, the present invention provides a simulation device that reproduces the operation of an apparatus, the simulation device comprising a modeling unit that constructs a governing equation for any state parameter of a component that constitutes the apparatus, the modeling unit constructing the governing equation based on measurement data obtained from a sensor that measures a state quantity of the component and a basis function for the governing equation, and estimating the state parameter using the governing equation. Effect of the Invention

[0010] According to the present invention, it is possible to estimate any state parameter of a part constituting an apparatus based on measurement data obtained from a sensor that measures a state quantity of the part. [Brief description of the drawings]

[0011] [Figure 1] 1 is a functional block diagram of a device equipped with a simulation device according to a first embodiment of the present invention. [Diagram 2] FIG. 11 is a functional block diagram of a device equipped with a simulation device according to a second embodiment of the present invention. [Diagram 3] FIG. 11 is a functional block diagram of a device equipped with a simulation device according to a third embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In each drawing, the same reference numerals are used to denote the same elements, and duplicated explanations will be omitted as appropriate. EXAMPLES

[0013] FIG. 1 is a functional block diagram of a device equipped with a simulation device according to a first embodiment of the present invention.

[0014] The device 1 is equipped with a sensor 4, a simulation device 2, a control and calculation unit 7, a display unit 8, and a control part 9. The simulation device 2 includes a modeling unit 5, a data assimilation unit 6, and a storage unit 3 that requires a basis function library 10.

[0015] The modeling unit 5 uses the measurement data of the control part 9 obtained from the sensor 4 and a basis function library 10 arranged in the storage unit 3 to construct a mathematical model (physical model) that reproduces the measurement data.

[0016] The data assimilation unit 6 derives parameters (measurement parameters) that reproduce the measurement data from the mathematical model constructed by the modeling unit 5, and updates the parameters (model parameters) of the mathematical model so that the measurement data obtained from the sensor 4 matches the measurement parameters. Updating the model parameters in this manner is called "data assimilation calculation." The data assimilation unit 6 estimates any parameter (state parameter) that represents the state of the control component 9 using the mathematical model with the corrected model parameters, and outputs the state parameter to the control calculation unit 7. The control calculation unit 7 outputs a control signal to the control component 9 to set the control component 9 to a desired state based on the estimated state parameter.

[0017] The data assimilation calculation in the data assimilation unit 6 applies a data assimilation calculation method represented by, for example, an ensemble Kalman filter or a particle filter. By applying these calculation methods, it is possible not only to estimate the state parameters but also to simultaneously calculate the variance or probability distribution of the state parameters, which indicates the likelihood of the estimated state parameters.

[0018] A power supply device equipped with a lithium-ion battery will be described as an example of the device 1 on which the simulation device 2 is mounted. In this case, the control part 9 is, for example, a battery, and the sensor 4 is, for example, a temperature sensor such as a thermocouple or a thermistor for measuring the temperature of the battery, and a current sensor for measuring the current of the battery. The temperature data and the current value obtained from these sensors become the measurement data. The modeling unit 5 of the simulation device 2 uses the basis function library 10 to construct a mathematical model for predicting the temperature change of the battery from the current value of the battery. In this case, the mathematical model is a governing equation in which the current value instructed by the control calculation unit 7 to the battery, which is the control part 9, is input, and the temperature data obtained by the sensor 4 is output. Here, the basis function library 10 includes, as basis functions, variables that can represent the relationship between the current and temperature of the lithium-ion battery, such as a first-order term of the battery temperature obtained from the sensor 4, a second-order term of the current for reproducing Joule heat, and a reciprocal term of the temperature for reproducing the temperature dependency of the internal resistance of the lithium-ion battery, and combinations thereof. This makes it possible to construct a governing equation for estimating the temperature change of the battery using these basis functions.

[0019] Here, the modeling unit 5 can construct a governing equation that reproduces the measurement data by using a sparse identification technique as a modeling method. As the governing equation, for example, a partial differential equation can be constructed. The procedure for constructing the governing equation using the sparse identification technique is shown below.

[0020] First, as shown in Equation 1, ordinary differential equations that are governing equations of general dynamic systems are considered.

[0021]

number

[0022] Here, X(t) represents the state of the dynamic system at time t. Furthermore, a matrix Θ(X) shown in Equation 2 is constructed as a basis function library 10 for a battery.

[0023]

number

[0024] Here, the matrix Θ(X) contains basis functions such as temperature T, current I, their quadratic functions, and the inverse of electrical resistance R.

[0025] Using this matrix Θ(X) and the coefficient vector Ξ, an ordinary differential equation representing the dynamic system shown in Equation 3 is constructed.

[0026]

number

[0027] Here, the coefficient vector Ξ is a parameter (model parameter) of the mathematical model, and can be expressed as in Equation 4.

[0028]

number

[0029] Here, each coefficient ξ of the coefficient vector Ξ is determined by regression analysis so as to reproduce the time series data of temperature obtained by the sensor 4. However, the subscript m of ξ indicates the number of columns of the matrix Θ(X), i.e., the number of assumed basis functions. By determining this coefficient vector Ξ, the differential equation of equation 1 can be obtained. In particular, in an ordinary differential equation, by leaving only the basis functions with a large contribution and setting the coefficient ξ of the basis functions with a small contribution to 0, the ordinary differential equation can be simplified.

[0030] Furthermore, by including various basis functions in the basis function library 10, even if the dynamic system has a different mounting position of the sensor 4, it is possible to reproduce any state parameter of the dynamic system by changing the combination of basis functions and adjusting the coefficients (model parameters) of those basis functions.

[0031] Furthermore, in the first embodiment, for the differential equation that reproduces this temperature change, the data assimilation unit 6 performs data assimilation calculations using measurement data sequentially obtained by the sensor 4. This makes it possible to sequentially update the mathematical model constructed by the modeling unit 5, estimate the state parameters of the control parts 9 with high accuracy, and output desired control signals to the control parts 9 by the control calculation unit 7. Furthermore, the control results of the control parts 9 and the state parameters of the control parts 9 estimated by the mathematical model can be displayed on the display unit 8.

[0032] Since these temperature changes are estimated by differential equations, not only can all temperature changes from the initial value be estimated, but also it is possible to estimate the temperature change from an arbitrary time by using the measurement data obtained at that time as the initial value. Therefore, the estimation of the temperature change by the constructed differential equation is realized by the data assimilation calculation using the temperature measurement data at the desired time. In addition, the number of control parts 9 for which the mathematical model is constructed is not limited to one. In addition, a mathematical model of the control part 9 may be constructed from a combination of two or more measurement data obtained from two or more sensors 4, and the state parameters may be estimated. In addition, a mathematical model may be constructed for a part that does not require control, and the state parameters may be estimated. The parts targeted by the basis function library 10 are not limited to electric components of the device, such as the battery, motor, inverter, compressor, and valve in the electric device. Basis functions for realizing modeling of the fan, heat sink, water-cooled jacket, and water-cooled pump that constitute the device that cools the electric components may be included in the library, and a model that reproduces not only the operating conditions of the device but also an integrated system including the cooling device may be constructed. Furthermore, the mathematical model is not limited to temperature analysis, but a model that reproduces the structure and movement may be constructed by including position and velocity as basis functions.

[0033] (summary) In the first embodiment, a simulation device 2 that reproduces the operation of an apparatus 1 is provided with a modeling unit 5 that constructs a governing equation for any state parameter of a control component 9, and the modeling unit 5 constructs the governing equation based on measurement data obtained from a sensor 4 that measures the state quantity of the control component 9 and a basis function for the governing equation, and estimates the state parameter using the governing equation.

[0034] According to the embodiment 1 configured as described above, it is possible to estimate any state parameter of the part 9 constituting the device 1 based on the measurement data obtained from the sensor 4 that measures the state quantity of the part 9.

[0035] In the first embodiment, the simulation device 2 includes a data assimilation unit 6 that updates the governing equation so that the measurement data coincides with a measurement parameter derived from the governing equation and reproducing the measurement data. This makes it possible to improve the estimation accuracy of any state parameter of the part 9.

[0036] In the first embodiment, the data assimilation unit 6 calculates a probability distribution or variance of the state parameters based on the measurement data and the state parameters estimated by the modeling unit 5, and the simulation device 2 includes a display unit 8 that displays the probability distribution or the variance. This makes it possible to quantitatively evaluate the likelihood of the state parameters estimated using the governing equations. EXAMPLES

[0037] A second embodiment of the present invention will be described, focusing on the differences from the first embodiment (shown in FIG. 1).

[0038] 2 is a functional block diagram of an apparatus 1 equipped with a simulation apparatus 2 in this embodiment. A basis function library 10 is arranged in a cloud 11. The simulation apparatus 2 further includes a communication unit 12 capable of communicating with the cloud 11. The modeling unit 5 refers to the basis functions arranged in the cloud 11 via the communication unit 12 when constructing a governing equation.

[0039] In this embodiment, the basis function library 10 is placed in the cloud 11, and the basis function library 10 can be updated based on information obtained from an apparatus other than the target apparatus 1, thereby improving the accuracy of the mathematical model constructed by the modeling unit 5 of the apparatus 1 and speeding up modeling.

[0040] In addition, in this embodiment, the basis function library 10 stored in the cloud 11 can be updated. For example, by adding a basis function to the basis function library 10 or deleting an unused basis function from the basis function library 10 in advance, the regression calculation performed by the modeling unit 5 using the measurement data obtained by the sensor 4 can be efficiently performed, and the accuracy of modeling can be improved. Furthermore, even when a simulation device 2 for a new control part 9 is installed, it is possible to use the basis function library 10 updated by the data obtained from the conventional device 1 and the control part 9, and information from multiple devices 1 can be collected. Furthermore, the mathematical model constructed by the modeling unit 5 may be placed in the cloud 11.

[0041] (summary) In Example 2, the simulation device 2 is equipped with a communication unit 12 capable of communicating with the cloud 11, and basis functions for the governing equations are placed in the cloud 11. When constructing the governing equations, the modeling unit 5 refers to the basis functions placed in the cloud 11 via the communication unit 12.

[0042] According to the embodiment 2 configured as described above, the basis function library 10 can be updated based on information obtained from an apparatus other than the target apparatus 1, thereby improving the accuracy of the mathematical model constructed by the modeling unit 5 of the apparatus 1 and speeding up the modeling. EXAMPLES

[0043] The third embodiment of the present invention will be described below, focusing on the differences from the first embodiment (shown in FIG. 1).

[0044] 3 is a functional block diagram of the device 1 equipped with the simulation device 2 in this embodiment. The simulation device 2 in this embodiment further includes a prediction unit 13. The prediction unit 13 calculates future predicted values ​​of the state parameters and the variance (variance or probability distribution) of the predicted values ​​using the governing equations (mathematical model) updated by the data assimilation unit 6. The display unit 8 displays the future predicted values ​​of the state parameters calculated by the prediction unit 13 and the variance (variance or probability distribution) of the predicted values.

[0045] (summary) The simulation device 2 in the third embodiment includes a prediction unit 13 that calculates future predicted values ​​of the state parameters using the governing equations updated by the data assimilation unit 6, and a display unit 8 that displays the predicted values ​​calculated by the prediction unit 13.

[0046] This embodiment configured as above also achieves the same effects as those of the above-mentioned embodiment 1. Moreover, since it is possible to predict future state changes of the control part 9, it becomes possible to search for optimal control conditions using a model in the control calculation unit 7, for example, and to control the control part 9 under those conditions.

[0047] The present invention is not limited to the above-described embodiment, but includes various modified examples. For example, the above-described embodiment is shown to easily explain the present invention, and is not necessarily limited to those including all of the configurations described. It is also possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. It is also possible to add, delete, or replace a part of the configuration of each embodiment with another configuration. [Explanation of symbols]

[0048] 1...device, 2...simulation device, 3...memory unit, 4...sensor, 5...modeling unit, 6...data assimilation unit, 7...control and calculation unit, 8...display unit, 9...control components, 10...basis function library, 11...cloud, 12...communication unit, 13...prediction unit.

Claims

1. In a simulation device that reproduces the operation of a device, A modeling unit that constructs a governing equation for any state parameter of a component that constitutes the device, The modeling unit constructs the governing equation based on measurement data obtained from a sensor that measures a state quantity of the part and a basis function for the governing equation, and estimates the state parameter using the governing equation. A simulation device comprising:

2. 2. The simulation device according to claim 1, a data assimilation unit that updates the governing equation so that the measurement data coincides with a measurement parameter derived from the governing equation and that reproduces the measurement data; A simulation device comprising:

3. 3. The simulation device according to claim 2, the data assimilation unit calculates a probability distribution or a variance of the state parameter based on the measurement data and the state parameter estimated by the modeling unit; The simulation device includes a display unit that displays the probability distribution or the variance. A simulation device comprising:

4. 3. The simulation device according to claim 2, a prediction unit that calculates a future predicted value of the state parameter by using the governing equation updated by the data assimilation unit; a display unit that displays the predicted value calculated by the prediction unit. A simulation device comprising:

5. 2. The simulation device according to claim 1, Equipped with a communication unit capable of communicating with the cloud, The basis functions are arranged in the cloud; The modeling unit refers to the basis functions stored in the cloud via the communication unit when constructing the governing equation. A simulation device comprising:

Images