JP6225655B2 - Model predictive control apparatus, method and program - Google Patents

Description

  The present invention relates to a model predictive control apparatus, method, and program.

  In a hybrid electric vehicle (HEV), for example, power is generated by a gasoline engine, and a motor is driven by electricity generated by the engine. Of the generated electricity, surplus electricity that is not used for driving the motor is accumulated by charging the storage battery. For example, when the engine output is insufficient, the HEV fuel efficiency is improved by driving the motor with a storage battery.

  For example, a lithium ion battery is used as the storage battery. However, a lithium ion battery deteriorates when it is in an overcharged state in which the charging rate indicating the state of the battery (SoC: State of Charge) is too high, and when the battery is in an overdischarged state in which the charging rate is too low. Is known to deteriorate. Therefore, in order to prevent the deterioration of the lithium ion battery and extend its life, it is desirable to use the lithium ion battery so as not to be overcharged or overdischarged.

An example of conventional storage battery monitoring type control will be described. In this example, the target power (hereinafter referred to as “target power”) P _d is determined from, for example, the HEV motor speed P _m , and the charge / discharge plan generator creates a charge / discharge plan that minimizes fuel consumption for a certain period. Create P (k). In the following model predictive control problem, Q _k represents a fuel consumption function, and P _m (k−1) represents the rotational speed of the motor at time k−1.

minΣQ _k (P _d , P _m (k−1), P (k))
When the overcharge or overdischarge occurs based on the planned charge / discharge plan P (k) based on the SoC from the storage battery, the deterioration prevention corrector stops charging so as not to overcharge, Make corrections to stop the discharge so that it does not occur. The engine power generation planner creates an engine power generation plan P _e (k) based on the corrected charge / discharge plans P ^to (k) and supplies the engine power generation plan P _e (k) to the engine. The corrected charge / discharge plans P ^to (k) are also supplied to the storage battery. Here, P _m (k) indicates the rotational speed of the motor at time k.

P _e (k) = P _d −P ^to (k)
In this example of the storage battery monitoring type control, the input restriction that SoC is 0 to 100%, the state restriction, that is, the OCV (Open Circuit Voltage) safety restriction are not taken into consideration. Moreover, since discharge is stopped when the SoC is in an overdischarge state, and charge is stopped when the SoC is in an overcharge state, there is a possibility that the discharge is stopped excessively or the charge is stopped excessively. Stopping the discharge excessively or stopping the charging excessively results in insufficient engine output, resulting in overdischarge or overcharge, and HEV fuel consumption decreases to avoid overdischarge or overcharge. May end up.

On the other hand, in addition to the SoC, the number of overcharge states, the number of overdischarge states, and the number of times between overdischarge states and overcharge states are provided in the state quantity of the control model. A system that improves the fuel efficiency of HEV for a certain period in consideration of the restriction of discharge has been proposed (for example, Non-Patent Document 1). In this proposed system, the number of times of overcharge / discharge is N, the count value of a coulomb counter that integrates the current flowing into the storage battery and the current flowing out of the storage battery is i, and the state is x (k) = [P _m , SoC, i , N] ^T , input is u (k) = [P (k), Pe (k)], auxiliary parameter for frequency measurement is δ, x (k + 1) = Ax (k) + B ₁ u (k) + B ₂ Suppose that δ (k), A, B ₁ and B ₂ are variables, the fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)) is defined by a quadratic expression of x and u. it can. Therefore, the model predictive control problem can be expressed as follows, and is the same as the above example of the conventional storage battery monitoring type control, but the constraints are u _min ≦ u (k) ≦ u _max , x _min ≦ x (K) ≦ x _max .

minΣQ _k (P _d , P _m (k−1), P (k))
Therefore, since the proposed system handles discrete states, the combinatorial optimization problem is solved, and the amount of calculation of the model predictive control problem increases. When the calculation amount of the model predictive control problem increases, a high-performance and expensive computer such as an ECU (Engine Control Unit) is required to execute the calculation, and the cost of the entire system also increases.

Masakazu Mukai and Taketoshi Kawamata, “A Study on Model Predictive Control of Electric Vehicle Battery System”, 11th Control Division Conference, March 16-18, 2011

  As in Non-Patent Document 1, in a system that improves the fuel efficiency of HEV for a certain period in consideration of the restrictions of overcharge and overdischarge, the amount of calculation of the model predictive control problem increases.

  Accordingly, in one aspect, an object of the present invention is to provide a model predictive control apparatus, method, and program capable of suppressing the amount of calculation of a model predictive control problem.

In one embodiment, a model predictive control device for a system in which a motor is driven by electricity generated by an engine, and among the electricity generated by the engine, electricity not used for driving the motor is accumulated by charging a storage battery And, based on the target power P _d , the rotational speed P _{m of the} motor and the state SoC of the storage battery, the overcharge limit according to the overcharge level and the overdischarge limit according to the overdischarge level are set. wherein, and, and the charge-discharge plan creation means fuel of a constant period of creating a smallest charge planning P (k), based on the charge planning P (k) and the target power P _d, the engine power program create a P e _(k) an engine power program creating means for supplying to said engine, said charge and discharge plan creation means, electrodeposition flowing out current and the battery which has flowed into the accumulator The count value of the coulomb counter for integrating the i, state _{x (k) = [P m} , SoC, i] T, enter the u (k) = [P ( k), Pe (k)], x (k + 1 ) = When Ax (k) + Bu (k ), a, and B variable, the target power _{P d} and on the basis of the rotational speed _{P m,} x, fuel efficiency function _Q k which is defined by a quadratic equation of u ( Based on the fuel consumption function generation unit that generates P _d , P _m (k−1), P _m (k)), and the target power P _d and the rotation speed P _m , u _min ≦ u (k) ≦ u _Based on the constraint condition generator that generates the constraint conditions u _min and u _max satisfying _max and the state SoC of the storage battery from the storage battery, to what extent the state SoC of the storage battery is less than the lower limit of overdischarge Tolerance and to what extent is it allowed to exceed the upper limit of overcharge The representative penalty function p (x (k)) and the penalty function generation unit for generating said fuel function _{Q k (P d, P m} (k-1), P m (k)), the constraint _{u min,} Based on u _max and the penalty function p (x (k)), an optimization problem that generates an optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) And the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) are solved and the charge / discharge plan P (k) is output. A model predictive control device having an optimization problem solving unit is provided.

  The computational complexity of the model predictive control problem can be reduced.

It is a block diagram which shows an example of the control system of HEV in one Example. It is a block diagram which shows an example of a charging / discharging plan creation device. It is a block diagram which shows an example of an engine power generation planner. It is a figure explaining an example of a penalty function. It is a figure explaining the other example of a penalty function. It is a block diagram which shows an example of the model prediction control apparatus in one Example. It is a flowchart explaining an example of a process of a model prediction control apparatus.

In the model predictive control device, method, and program in the embodiment, the overcharge limit and overdischarge level according to the overcharge level are set based on the target power P _d , the motor rotation speed P _m, and the storage battery state SoC. A charge / discharge plan P (k) that includes a corresponding overdischarge limit and that minimizes the fuel consumption for a certain period is created. Further, an engine power generation plan P _e (k) is created based on the charge / discharge plan P (k) and the target power P _d and supplied to the engine. A motor is driven by electricity generated by the engine, and electricity that is not used for driving the motor among the electricity generated by the engine is accumulated by charging the storage battery.

  Embodiments of the model predictive control device, method, and program will be described below with reference to the drawings as embodiments of the invention.

  FIG. 1 is a block diagram illustrating an example of an HEV control system according to an embodiment. In FIG. 1, the HEV control system 1 includes a charge / discharge plan generator 11, an engine power generation planner 12, a storage battery 13, an engine 14, and a motor 15. The storage battery 13 is, for example, a lithium ion battery. The engine 14 is not limited to a gasoline engine, and may be a diesel engine, a liquefied gas engine, or the like.

The control system 1 determines the target power P _d from, for example, the rotation speed P _m of the HEV motor 15, and the charge / discharge plan generator 11 sets the target power P _d , the SoC from the storage battery 13 and the rotation speed P from the motor 15. Based on _m , overcharge limit and overdischarge limit in consideration of overcharge level and overdischarge level, that is, overcharge limit according to overcharge level and overdischarge according to overdischarge level The charging / discharging plan P (k) that includes the above-described limitation and minimizes the fuel consumption for a certain period is created. Target power _{P d,} for example described below ECU may determine. The charge / discharge plan P (k) is supplied to the engine power generation planner 12 and the storage battery 13. The target power P _d is also supplied to the engine power generation planner 12. The engine power generation planner 12 creates an engine power generation plan P _e (k) based on the charge / discharge plan P (k) and the target power P _d and supplies it to the engine 14. Electricity is generated by the engine 14 and the motor 15 is driven by electricity generated by the engine 14. Of the electricity generated by the engine 14, surplus electricity that is not used for driving the motor 15 is accumulated by charging the storage battery 13. For example, when the output of the engine 14 is insufficient, the motor 15 is driven by the storage battery 13 to improve the fuel efficiency of the HEV.

The charge / discharge plan creation device 11 is an example of a charge / discharge plan creation means for executing a charge / discharge plan creation process for creating a charge / discharge plan P (k). The engine power generation planner 12 is an example of an engine power generation plan creation unit that executes an engine power generation plan creation process for creating the engine power generation plan P _e (k). The charge / discharge plan generator 11 and the engine power generation planner 12 form an example of a model predictive control device. The functions of the charge / discharge plan creator 11 and the engine power generation planner 12 may be realized by an ECU (not shown) that controls the HEV engine 14 as will be described later.

  FIG. 2 is a block diagram illustrating an example of a charge / discharge plan creator. In FIG. 2, the charge / discharge plan generator 11 includes a fuel consumption function generator 21, a constraint condition generator 22, a penalty function generator 23, an optimization problem generator 24, and an optimization problem solver (or an optimization problem solver). 25.

The count value of the coulomb counter that integrates the current flowing into the storage battery 13 and the current flowing out of the storage battery 13 in the charge / discharge plan generator 11 is i, the state is x (k) = [P _m , SoC, i] ^T , input Where u (k) = [P (k), Pe (k)], x (k + 1) = Ax (k) + Bu (k), and A and B are variables, the fuel efficiency function Q _k (P _d , P _m (K-1), P _m (k)) can be defined by a quadratic expression of x and u. The fuel consumption function generation unit 21 generates a fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)) based on the target power P _d and the rotation speed P _m of the motor 15. The constraint condition generator 22 generates constraint conditions u _min and u _max that satisfy u _min ≦ u (k) ≦ u _max based on the target power P _d and the rotational speed P _m of the motor 15. The penalty function generating unit 23 generates a penalty function p (x (k)) based on the SoC from the storage battery 13. The penalty function p (x (k)) represents the extent to which SoC is allowed to be less than the lower limit value of overdischarge and to what extent is allowed if the SoC exceeds the upper limit value of overcharge. If it is a function, it will not be specifically limited.

The optimization problem generation unit 24 includes the fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)) from the fuel consumption function generation unit 21, and the constraint condition u _min , from the constraint condition generation unit 22. Based on u _max and the penalty function p (x (k)) from the penalty function generator 23, the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k )). The optimization problem solving unit 25 solves the above-described optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)), and the charge / discharge plan P (k) Is output.

Fuel function generating unit 21 is an example of a fuel economy function generating means for executing a fuel function generation process for generating a fuel economy function Q _k. The constraint condition generation unit 22 is an example of a constraint condition generation unit that executes a constraint condition generation process for generating the constraint conditions u _min and u _max . The penalty function generation unit 23 is an example of a penalty function generation unit that executes a penalty function generation process for generating a penalty function p (x (k)). The optimization problem generating unit 24 performs optimization problem generation processing for generating an optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)). It is an example of a problem generation means. The optimization problem solving unit 25 executes an optimization problem solving process for solving the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) for a certain period. An optimization problem for obtaining a charge / discharge plan P (k) that includes an overcharge limit according to the overcharge level and an overdischarge limit according to the overdischarge level, and that minimizes the fuel consumption for a certain period. It is an example of a solving means.

FIG. 3 is a block diagram illustrating an example of an engine power generation planner. In FIG. 3, the engine power generation planner 12 has an engine power generation plan creation unit 31. The engine power generation plan creation unit 31 creates the engine power generation plan P _e (k) based on the target power P _d and the charge / discharge plan P (k) from the charge / discharge plan creator 11. In this example, the engine power generation plan creation unit 31 calculates P _d −P (k) = P _e (k).

  FIG. 4 is a diagram illustrating an example of a penalty function. In FIG. 4, the vertical axis represents the penalty function p (x (k)), and the horizontal axis (that is, the x axis) represents SoC (%). The penalty function p (x (k)) shown in FIG. 4 can be expressed by the following equation. Here, r represents an arbitrary function.

In FIG. 4, the alternate long and short dash lines indicate cases where SoC is 0, 20%, 80%, and 100%. For example, in Non-Patent Document 1 described above, SoC ranges from 0 to 20% and ranges from 80% to 100%. Then, overdischarge and overcharge are stopped. On the other hand, the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) using the penalty function p (x (k)) shown in FIG. When the charge / discharge plan P (k) is obtained by solving, as shown by the solid line, the penalty function p (x (k)) increases as the SoC decreases when the SoC is in the range of 0 to 20%. In the range of 80% to 100%, the penalty function p (x (k)) increases as the SoC increases. For example, if over several times, the overdischarge and the upper limit (less than the lower limit (20%) that can be tolerated) Overcharge exceeding 80%) can be allowed.

FIG. 5 is a diagram for explaining another example of the penalty function. In FIG. 5, the vertical axis represents the penalty function p (x (k)), and the horizontal axis (that is, the x axis) represents SoC (%). The penalty function p (x (k)) shown in FIG. 5 can be expressed by the following equation including an exponential function. Here, r ₁ and r ₂ each represent an arbitrary function.

In FIG. 5, the alternate long and short dash lines indicate cases where SoC is 0, 20%, 80%, and 100%. For example, in Non-Patent Document 1 described above, SoC ranges from 0 to 20% and ranges from 80% to 100%. Then, overdischarge and overcharge are stopped. On the other hand, the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) using the penalty function p (x (k)) shown in FIG. When the charge / discharge plan P (k) is obtained by solving, as shown by the solid line, the penalty function p (x (k)) increases as the SoC decreases when the SoC is in the range of 0 to 20%. In the range of 80% to 100%, the penalty function p (x (k)) increases as the SoC increases. For example, if over several times, the overdischarge and the upper limit (less than the lower limit (20%) that can be tolerated) Overcharge exceeding 80%) can be allowed.

  Note that the upper limit of overcharge and the lower limit of overdischarge for the storage battery 13 are not limited to the above SoC values, and can be appropriately set according to the storage battery 13. Further, the penalty function p (x (k)) can also be appropriately set according to the storage battery 13, and is not limited to a specific function.

According to the above embodiment, the constraints u _min and u _max satisfying u _min ≦ u (k) ≦ u _max with respect to the input u (k) are used, and x _min ≦ x (k) ≦ x _max Since the constraint condition for x (k) is not used, the amount of calculation of the model predictive control problem can be suppressed as compared with, for example, the proposed system of Non-Patent Document 1 described above. For this reason, it is not necessary to use a high-performance and expensive computer such as an ECU in order to execute the calculation of the model predictive control problem, and an increase in the cost of the entire system can be suppressed.

  Further, when the upper limit of overcharge of the lithium ion battery is 80%, for example, and the lower limit of overdischarge is 20%, for example, it is desirable to avoid overcharge exceeding 80%. Even if this occurs, the lithium ion battery does not deteriorate immediately. Although it is desirable to avoid overdischarge below 20%, for example, even if 19% overdischarge occurs several times, the lithium ion battery does not deteriorate immediately. On the other hand, for example, when 98% overcharge or 2% overdischarge occurs, the lithium ion battery deteriorates even if it is overcharged or overdischarged once.

  For example, in the proposed system of Non-Patent Document 1, since the overcharge level (that is, the degree) is not taken into consideration, if the upper limit of overcharge is 80%, both overcharge of 81% and overcharge of 98% It will be treated as causing the same deterioration to the lithium ion battery. Similarly, since the level of overdischarge is not taken into account, if the lower limit of overdischarge is 20%, for example, 19% overdischarge and 2% overdischarge are treated as causing the same deterioration to the lithium ion battery. It will be broken. Therefore, in this example, 81% overcharge or 19% overdischarge, such as overcharge or overdischarge at a level that can be tolerated for several times, is limited. Since charging and discharging may be excessively restricted to prevent this, it is difficult to further improve the HEV fuel efficiency in the proposed system of Non-Patent Document 1.

  On the other hand, according to the above-described embodiment, it is possible to limit overcharge and overdischarge in consideration of the overcharge level and the overdischarge level. It is possible to further improve HEV fuel efficiency by allowing overcharge or overdischarge at a level that can be tolerated if it is several times, such as 19% overdischarge.

  FIG. 6 is a block diagram illustrating an example of a model prediction control apparatus according to an embodiment. In FIG. 6, the same parts as those in FIG. In FIG. 6, the model predictive control device 41 includes an ECU (Engine Control Unit) 51 that controls the engine 14 and a memory 52. The ECU 51 is an example of a computer such as a computer or a processor that implements the functions of the charge / discharge plan creator 11 and the engine power generation planner 12 shown in FIG. The memory 52 is an example of a storage unit that stores a program executed by the ECU 51, intermediate results such as calculations executed by the ECU 51, various parameters, a penalty function, and the like. The memory 52 may form an example of a computer-readable storage medium that stores a program that causes the ECU 51 to execute the model prediction control process. Computer-readable storage media include portable recording media such as CD-ROM (Compact Disk-Read Only Memory), DVD (Digital Versatile Disk), USB (Universal Serial Bus) memory, semiconductor memory such as flash memory, and hard disk It can be formed by a storage device such as a hard disk drive (HDD).

The ECU 51 can function as a target power determination unit that determines the target power P _d from, for example, the rotation speed P _m of the HEV motor 15, and is based on the target power P _d , the SoC from the storage battery 13, and the rotation speed P _m from the motor 15. Thus, a charge / discharge plan P (k) that includes overcharge limit and overdischarge limit considering the overcharge level and overdischarge level, and that minimizes the fuel consumption for a certain period, is created in the storage battery 13. It can function as a charging / discharging plan creation means to be supplied. Further, the ECU 51 can function as an engine power generation plan means that creates an engine power generation plan P _e (k) and supplies it to the engine 14 based on the charge / discharge plan P (k) and the target power P _d . As a result, power is generated by the engine 14 and the motor 15 is driven by electricity generated by the engine 14. Of the electricity generated by the engine 14, surplus electricity that is not used for driving the motor 15 is accumulated by charging the storage battery 13. For example, when the output of the engine 14 is insufficient, the motor 15 is driven by the storage battery 13 to improve the fuel efficiency of the HEV.

  FIG. 7 is a flowchart illustrating an example of processing of the model prediction control apparatus. The process shown in FIG. 7 can be executed by the ECU 51 in FIG. In FIG. 7, a target power determination process is executed in step S1, a charge / discharge plan creation process is executed in steps S2 to S6, and an engine power generation plan creation process is executed in step S7. The charge / discharge plan creation process and the engine power generation plan creation process are included in the model predictive control process. The target power determination process may be included in the model predictive control process.

In FIG. 7, in step S < _b > 1, the ECU 51 executes target power determination processing, and determines the target power P _d from, for example, the rotational speed P _m of the HEV motor 15 and stores it in the memory 52.

In step S2, the ECU 51 executes a fuel efficiency function generation process, and based on the target power P _d and the rotational speed P _m of the motor 15, the fuel efficiency functions Q _k (P _d , P _m (k−1), P _m (k )) Is generated and stored in the memory 52. In step S3, the ECU 51 executes a constraint condition generation process, and based on the target power P _d and the rotational speed P _m of the motor 15, the constraint conditions u _min and u _max satisfying u _min ≦ u (k) ≦ u _max are set. Generated and stored in the memory 52. In step S <b> 4, the ECU 51 executes a penalty function generation process, generates a penalty function p (x (k)) based on the SoC from the storage battery 13, and stores it in the memory 52. The execution order of steps S2, S3, and S4 is not particularly limited, and at least some of them may be executed in parallel.

In step S5, the ECU 51 executes an optimization problem generation process, and the fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)) stored in the memory 52 and the constraints u _min , u _{Based on max} and the penalty function p (x (k)), an optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) is generated to generate the memory 52. To remember. In step S < _b > 6, the ECU 51 executes an optimization problem solving process, and the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) stored in the memory 52. And charge / discharge plan P (k) is output.

In step S7, the ECU 51 executes engine power generation plan creation processing, creates the engine power generation plan P _e (k) based on the target power P _d and the charge / discharge plan P (k) stored in the memory 52, and stores the memory. 52. In this example, in the engine power generation plan creation process, P _d −P (k) = P _e (k) is calculated and supplied to the engine 14.

  According to the above embodiment, charging and discharging are not excessively restricted in order to prevent the deterioration of the lithium ion battery, and the calculation amount of the model predictive control problem can be suppressed. Moreover, according to the said Example, the electric power generation amount by the engine of HEV and the charging / discharging amount of a lithium ion battery can be determined so that deterioration of a lithium ion battery may be suppressed.

The following supplementary notes are further disclosed with respect to the embodiments including the above examples.
(Appendix 1)
A motor is driven by electricity generated by an engine, and among the electricity generated by the engine, electricity that is not used for driving the motor is a model predictive control device that accumulates by charging a storage battery,
Based on the target power P _d , the rotational speed P _{m of the} motor and the state SoC of the storage battery, including an overcharge limit according to the overcharge level and an overdischarge limit according to the overdischarge level; and Charging / discharging plan creation means for creating a charging / discharging plan P (k) that minimizes fuel consumption for a certain period;
An engine power generation plan creating means for creating an engine power generation plan P _e (k) based on the charge / discharge plan P (k) and the target power P _d and supplying the engine power generation plan P _e (k) to the engine Predictive control device.
(Appendix 2)
The charge / discharge plan creation means includes:
The count value of the coulomb counter that integrates the current flowing into the storage battery and the current flowing out of the storage battery is i, the state is x (k) = [P _m , SoC, i] ^T , and the input is u (k) = [P (k), Pe (k) ], x (k + 1) = Ax (k) + Bu (k), a, when the a variable B, and based on the target power _{P d} and the rotational speed _{P m,} x, u A fuel consumption function generation unit that generates a fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)) defined by the quadratic expression of
A constraint condition generator that generates constraint conditions u _min and u _max satisfying u _min ≦ u (k) ≦ u _max based on the target power P _d and the rotation speed P _m ;
Based on the state SoC of the storage battery from the storage battery, to what extent is the state allowed when the state SoC of the storage battery is less than the lower limit value of overdischarge, and to what extent is allowed when the upper limit value of overcharge is exceeded A penalty function generator that generates a penalty function p (x (k)) representing
Based on the fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)), the constraints u _min , u _max and the penalty function p (x (k)), an optimization problem an optimization problem generator that generates minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k));
An optimization problem solving unit that solves the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) and outputs the charge / discharge plan P (k) The model predictive control device according to appendix 1, characterized by comprising:
(Appendix 3)
The engine power generation plan creation means creates an engine power generation plan P _e (k) such that P _d −P (k) = P _e (k) based on the target power P _d and the charge / discharge plan P (k). The model predictive control apparatus according to Supplementary Note 1 or 2, further comprising an engine power generation plan creation unit that performs the following.
(Appendix 4)
And further comprising a target power determination means for determining the target power P _d from the rotational speed P _m, the model predictive control apparatus according to any one of Supplementary Notes 1 to 3.
(Appendix 5)
A motor is driven by electricity generated by an engine, and among the electricity generated by the engine, electricity that is not used for driving the motor is a model predictive control method for accumulating by charging a storage battery,
A computer for controlling the engine
Based on the target power P _d , the rotational speed P _{m of the} motor and the state SoC of the storage battery, including an overcharge limit according to the overcharge level and an overdischarge limit according to the overdischarge level; and Create a charge / discharge plan P (k) that minimizes fuel consumption over a period of time,
A model predictive control method, comprising: executing an engine power generation plan P _e (k) to be supplied to the engine based on the charge / discharge plan P (k) and the target power P _d .
(Appendix 6)
Creation of the charge / discharge plan P (k)
The count value of the coulomb counter that integrates the current flowing into the storage battery and the current flowing out of the storage battery is i, the state is x (k) = [P _m , SoC, i] ^T , and the input is u (k) = [P (k), Pe (k) ], x (k + 1) = Ax (k) + Bu (k), a, when the a variable B, and based on the target power _{P d} and the rotational speed _{P m,} x, u A fuel consumption function generation process for generating a fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)) defined by the quadratic equation of
A constraint condition generation process for generating constraint conditions u _min and u _max satisfying u _min ≦ u (k) ≦ u _max based on the target power P _d and the rotation speed P _m ;
Based on the state SoC of the storage battery from the storage battery, to what extent is the state allowed when the state SoC of the storage battery is less than the lower limit value of overdischarge, and to what extent is allowed when the upper limit value of overcharge is exceeded A penalty function generation process for generating a penalty function p (x (k)) representing
Based on the fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)), the constraints u _min , u _max and the penalty function p (x (k)), an optimization problem optimization problem generation processing for generating minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k));
An optimization problem solving process for solving the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) and outputting a charge / discharge plan P (k) The model predictive control method according to appendix 5, wherein the model predictive control method is included.
(Appendix 7)
The generation of the engine power generation plan P _e (k) is based on the target power P _d and the charge / discharge plan P (k), and the engine power generation plan P _{e is} P _d −P (k) = P _e (k). The model predictive control method according to appendix 5 or 6, characterized by including an engine power generation plan creation process for creating (k).
(Appendix 8)
Said computer, said from the rotation speed P _m target power P _d processing for determining further characterized by executing, model predictive control method according to any one of Appendices 5-7.
(Appendix 9)
A program that drives a motor with electricity generated by an engine and causes a computer to execute model predictive control processing of a system in which electricity that is not used for driving the motor among the electricity generated by the engine is stored by charging a storage battery Because
Based on the target power P _d , the rotational speed P _{m of the} motor and the state SoC of the storage battery, including an overcharge limit according to the overcharge level and an overdischarge limit according to the overdischarge level; and Create a charge / discharge plan P (k) that minimizes fuel consumption over a period of time,
Generating an engine power generation plan P _e (k) to be supplied to the engine based on the charge / discharge plan P (k) and the target power P _d , causing the computer to execute the model predictive control process; ,program.
(Appendix 10)
Creation of the charge / discharge plan P (k)
The count value of the coulomb counter that integrates the current flowing into the storage battery and the current flowing out of the storage battery is i, the state is x (k) = [P _m , SoC, i] ^T , and the input is u (k) = [P (k), Pe (k) ], x (k + 1) = Ax (k) + Bu (k), a, when the a variable B, and based on the target power _{P d} and the rotational speed _{P m,} x, u A fuel consumption function generation process for generating a fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)) defined by the quadratic equation of
A constraint condition generation process for generating constraint conditions u _min and u _max satisfying u _min ≦ u (k) ≦ u _max based on the target power P _d and the rotation speed P _m ;
Based on the state SoC of the storage battery from the storage battery, to what extent is the state allowed when the state SoC of the storage battery is less than the lower limit value of overdischarge, and to what extent is allowed when the upper limit value of overcharge is exceeded A penalty function generation process for generating a penalty function p (x (k)) representing
Based on the fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)), the constraints u _min , u _max and the penalty function p (x (k)), an optimization problem optimization problem generation processing for generating minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k));
An optimization problem solving process for solving the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) and outputting a charge / discharge plan P (k) The program according to appendix 9, wherein the program is included.
(Appendix 11)
The generation of the engine power generation plan P _e (k) is based on the target power P _d and the charge / discharge plan P (k), and the engine power generation plan P _{e is} P _d −P (k) = P _e (k). The program according to appendix 9 or 10, including an engine power generation plan creation process for creating (k).
(Appendix 12)
The program, the processing of the rotation speed P _m to determine the target power P _d, characterized in that it further executes the computer, model predictive control method according to any one of Appendices 9-11.

  Although the disclosed model predictive control apparatus, method, and program have been described with the embodiments, the present invention is not limited to the above embodiments, and various modifications and improvements can be made within the scope of the present invention. Needless to say.

DESCRIPTION OF SYMBOLS 1 Control system 11 Charging / discharging plan preparation machine 12 Engine power generation planner 13 Storage battery 14 Engine 15 Motor 41 Model prediction control apparatus 51 ECU
52 memory

Claims (4)

A motor is driven by electricity generated by an engine, and among the electricity generated by the engine, electricity that is not used for driving the motor is a model predictive control device that accumulates by charging a storage battery,
Based on the target power P _d , the rotational speed P _{m of the} motor and the state SoC of the storage battery, including an overcharge limit according to the overcharge level and an overdischarge limit according to the overdischarge level; and Charging / discharging plan creation means for creating a charging / discharging plan P (k) that minimizes fuel consumption for a certain period;
On the basis of the charge planning P (k) and the target power P _d, and an engine power program creating means for supplying to said engine to create an engine power program P _{e (k),}
The charge / discharge plan creation means includes:
The count value of the coulomb counter that integrates the current flowing into the storage battery and the current flowing out of the storage battery is i, the state is x (k) = [P _m , SoC, i] ^T , and the input is u (k) = [P (k), Pe (k) ], x (k + 1) = Ax (k) + Bu (k), a, when the a variable B, and based on the target power _{P d} and the rotational speed _{P m,} x, u A fuel consumption function generation unit that generates a fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)) defined by the quadratic expression of
A constraint condition generator that generates constraint conditions u _min and u _max satisfying u _min ≦ u (k) ≦ u _max based on the target power P _d and the rotation speed P _m ;
Based on the state SoC of the storage battery from the storage battery, to what extent is the state allowed when the state SoC of the storage battery is less than the lower limit value of overdischarge, and to what extent is allowed when the upper limit value of overcharge is exceeded A penalty function generator that generates a penalty function p (x (k)) representing
Based on the fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)), the constraints u _min , u _max and the penalty function p (x (k)), an optimization problem an optimization problem generator that generates minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k));
An optimization problem solving unit that solves the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) and outputs a charge / discharge plan P (k) A model predictive control device comprising: The engine power generation plan creation means creates an engine power generation plan P _e (k) such that P _d −P (k) = P _e (k) based on the target power P _d and the charge / discharge plan P (k). and having an engine power generation planning unit that, the model predictive control apparatus according to claim 1. A motor is driven by electricity generated by an engine, and among the electricity generated by the engine, electricity that is not used for driving the motor is a model predictive control method for accumulating by charging a storage battery,
A computer for controlling the engine
Based on the target power P _d , the rotational speed P _{m of the} motor and the state SoC of the storage battery, including an overcharge limit according to the overcharge level and an overdischarge limit according to the overdischarge level; and A process of creating a charge / discharge plan P (k) that minimizes the fuel consumption for a certain period;
The charge planning P (k) and on the basis of the target power P _d, and executes the process of creating a supply engine power program P _{e (k)} to the engine,
The process of creating the charge / discharge plan P (k) is as follows:
The count value of the coulomb counter that integrates the current flowing into the storage battery and the current flowing out of the storage battery is i, the state is x (k) = [P _m , SoC, i] ^T , and the input is u (k) = [P (k), Pe (k) ], x (k + 1) = Ax (k) + Bu (k), a, when the a variable B, and based on the target power _{P d} and the rotational speed _{P m,} x, u A fuel consumption function generation process for generating a fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)) defined by the quadratic equation of
A constraint condition generation process for generating constraint conditions u _min and u _max satisfying u _min ≦ u (k) ≦ u _max based on the target power P _d and the rotation speed P _m ;
Based on the state SoC of the storage battery from the storage battery, to what extent is the state allowed when the state SoC of the storage battery is less than the lower limit value of overdischarge, and to what extent is allowed when the upper limit value of overcharge is exceeded A penalty function generation process for generating a penalty function p (x (k)) representing
Based on the fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)), the constraints u _min , u _max and the penalty function p (x (k)), an optimization problem optimization problem generation processing for generating minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k));
An optimization problem solving process for solving the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) and outputting a charge / discharge plan P (k) is performed. A predictive control method for a model, comprising : Driving the motor with electricity generated by the engine, of the electricity generated by the engine, electricity is not used for driving the motor in order to run the model predictive control process of the system that are accumulated by charging the battery in the computer The program of
Based on the target power P _d , the rotational speed P _{m of the} motor and the state SoC of the storage battery, including an overcharge limit according to the overcharge level and an overdischarge limit according to the overdischarge level; and Create a charge / discharge plan P (k) that minimizes fuel consumption over a period of time,
On the basis of the charge planning P (k) and the target power P _d, to create a supply engine power program P _{e (k)} to the engine,
Causing the computer to perform processing ;
The process of creating the charge / discharge plan P (k) is as follows:
The count value of the coulomb counter that integrates the current flowing into the storage battery and the current flowing out of the storage battery is i, the state is x (k) = [P _m , SoC, i] ^T , and the input is u (k) = [P (k), Pe (k) ], x (k + 1) = Ax (k) + Bu (k), a, when the a variable B, and based on the target power _{P d} and the rotational speed _{P m,} x, u A fuel consumption function generation process for generating a fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)) defined by the quadratic equation of
A constraint condition generation process for generating constraint conditions u _min and u _max satisfying u _min ≦ u (k) ≦ u _max based on the target power P _d and the rotation speed P _m ;
Based on the state SoC of the storage battery from the storage battery, to what extent is the state allowed when the state SoC of the storage battery is less than the lower limit value of overdischarge, and to what extent is allowed when the upper limit value of overcharge is exceeded A penalty function generation process for generating a penalty function p (x (k)) representing
Based on the fuel consumption function Q _k (P _d , P _m (k−1), P _m (k)), the constraints u _min , u _max and the penalty function p (x (k)), an optimization problem optimization problem generation processing for generating minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k));
An optimization problem solving process for solving the optimization problem minΣQ _k (P _d , P _m (k−1), P (k)) + p (x (k)) and outputting a charge / discharge plan P (k) is performed. A program characterized by including .

Images