JP6509537B2 - Plant simulator and operation assignment method of plant simulator - Google Patents

Description

  The present invention relates to a plant simulator and an operation assignment method of the plant simulator.

  Conventionally, in power generation plants such as thermal power plants and nuclear power plants, simulators for operation training used for education and training of operators have been developed (see, for example, Patent Documents 1 and 2). When such operation training simulator is implemented by parallel processing by a plurality of processors, calculation is performed by creating a plant model in which components of a power generation plant are connected and assigning a mathematical model of an equipment unit to each processor. Had to run.

JP, 2014-6178, A JP, 2011-75672, A

In recent years, there has been a demand for further speedup of simulation speed (for example, 10 times the speed of the actual behavior), and there is a need to develop a simulator that satisfies this demand. However, according to the conventional method of assigning a mathematical model to each processor on an equipment basis, since the calculation load assigned to each processor varies, parallel processing can not be efficiently progressed, and time loss is caused. There has been a problem that speeding up to satisfy the above requirements can not be expected. For example, if the processing load of each processor varies, the time to finish the processing varies among the processors, and the processor that finished the process has to wait until the processing of the other processor finishes, which is a waste Waiting time had occurred.
In addition, simulation speed has a limit in leveling the processing of each processor by allocating a mathematical model in units of equipment to each processor, so sufficient waiting time until the end of processing of other processors is sufficient. In order to realize the speeding up of the simulation speed, the cost of the simulator has been increased, for example, by further increasing the number of processors and increasing the number of parallel processes.

  The present invention has been made in view of such circumstances, and it is an object of the present invention to provide a plant operation simulator capable of achieving a further increase in simulation speed and a method thereof.

A first aspect of the present invention is a plant simulator that simulates the behavior of a power plant by a plurality of processors performing simulation operations in parallel, and a plurality of mathematical models for computing the behavior of the plant are data types. And mathematical operation storing means stored in association with each other and arithmetic assignment means for assigning the mathematical model to the processor for each data type, one of the data types being a nonlinear model, the operation assignment The means extracts the mathematical model having a predetermined computational load or more, and preferentially assigns the mathematical model among the extracted mathematical models in descending order of the computational load, and the computational load of each processor is approximately as will be equally, have row assignment of the mathematical model, the arithmetic assignment means, nonlinear A non-linear operation assignment unit for assigning the mathematical model of Dell to the processor, wherein the non-linear operation assignment unit extracts the mathematical model having a computational load equal to or greater than a preset operation load from the mathematical model of the non-linear model Extracting a predetermined number of the mathematical models from the extracted mathematical models in descending order of operation load, and assigning each of the extracted mathematical models to the individual processors; and the mathematical model of the non-linear model being assigned. And a step of assigning a mathematical model of another non-linear model so as to substantially equalize the operation load of each of the processors that execute the simulation operation including the processors .

According to this aspect, the mathematical model storage means stores a plurality of mathematical models representing the behavior of the plant in association with the data types and is classified and stored, and the calculation load of each processor is substantially equal, A mathematical model is assigned to a plurality of processors for each data type by the operation assigning means. As a result, the utilization rates of the processors, in other words, the computation time of each processor can be made substantially uniform, and the simulator speed can be improved.
Furthermore, according to the present aspect, since mathematical model allocation in which the processing content is divided for each data type is performed, not for each device conventionally performed, it is possible to perform allocation according to the peculiarity of operation. Become. In addition, since mathematical models above the operation load set in advance are extracted, and among the extracted mathematical models, the mathematical models are preferentially assigned to the processors in descending order of the operation load, so that the operation load of each processor becomes substantially equal. It is possible to relatively easily assign such mathematical models, and it is possible to eliminate the need to increase the number of processors more than necessary.
Further, according to this aspect, although the mathematical model of the nonlinear model has a high operation load and a longer processing time than the linear model, it preferentially assigns the mathematical model with the high operation load to the processor. It is possible to relatively easily assign a mathematical model so that the calculation load of each processor is substantially equal.

  In the plant simulator, the operation assignment unit may assign the mathematical model so that a mathematical model of the same or similar data type is assigned to one processor.

  According to the above configuration, by assigning mathematical models of the same or similar data types to one processor, it is possible to repeat similar processing. Thereby, the improvement of the calculation speed can be expected.

  In the plant simulator, the operation assignment unit may determine the operation load based on the size of the operation matrix.

  If the size of the operation matrix is the same, the operation time can be considered to be substantially the same, so by using the size of the operation matrix as an index of the size of the operation load, it is easily and efficiently It becomes possible to assign mathematical models.

In the above-mentioned plant simulator, the operation assignment unit is any one of linear models to which no mathematical model of the non- linear model extracted by the non-linear operation assignment unit is assigned. It may have pressure operation assignment means assigned to the processor.

The pressure calculation of the steam system occupies an important position in the mathematical model of the linear model, and the computation time is also longer compared to other linear models due to the relatively large computational load. Therefore, with regard to such a mathematical model of pressure calculation of steam system, the operation load is distributed by giving priority to any processor that does not calculate the non-linear model extracted by the non- linear operation allocation means. Is possible.

In the above-mentioned plant simulator, the operation assignment unit is a mathematical model of volume calculation of a combustion system among linear models, and also a mathematical model of the non-linear model extracted by the non-linear operation assignment unit. A volume operation assignment means may be provided which is assigned to any of the processors which is not assigned a model.

The volume calculation of the combustion system (for example, volume calculation of combustion air, exhaust gas, fuel gas, etc.) occupies an important position in the mathematical model of the linear model, and the calculation load is also relatively large. The computation time is also longer than other linear models. Therefore, with regard to such mathematical model of volume calculation of the combustion system, the calculation of the non-linear model extracted by the non- linear calculation allocation means is also performed by preferentially allocating to the processor which does not calculate the pressure calculation of the steam system. It is possible to distribute the burden.

In the plant simulator, the arithmetic assignment means, of the linear model, the processor mathematical model of the heat transfer calculation for the pipe, a mathematical model of the heat transfer calculations for the valve, and a mathematical model of the heat transfer calculations for the pump fan different The heat transfer calculation assignment means may be distributed and assigned to

  Electric heating calculation is one of linear models. Electric heat calculation needs to be performed for each device, but the number of pipes, valves, and pump fans is particularly large, so the operation load is large and the processing time is also relatively long. Therefore, the processor operation can be performed by distributing and preferentially assigning the mathematical model of heat transfer calculation for piping, the mathematical model of heat transfer calculation for valves, and the mathematical model of heat transfer calculation for pump fans to different processors. It is possible to distribute the burden.

A second aspect of the present invention is the operation assignment method applied to a plant simulator that simulates the behavior of a power plant by a plurality of processors performing simulation operations in parallel, wherein the plant simulator is the power plant And a mathematical model storing unit in which a plurality of mathematical models for computing the behavior of the mathematical model are stored in association with the data type, and a computer assigns an arithmetic assignment step of assigning the mathematical model to the processor for each data type. One of the data types is a non-linear model, and in the operation assignment step, the mathematical model above the preset operation load is extracted and the operation load is large among the extracted mathematical models In addition to assigning the mathematical model in order of priority, each processor As calculation load is substantially uniform, have row assignment of the mathematical model, the arithmetic assignment step, the mathematical model of the nonlinear model includes a nonlinear operation allocation step of allocating to said processor, said non-linear operation assignment step And extracting the mathematical model having a predetermined operation load or more from the mathematical model of the non-linear model, and extracting a predetermined number of the mathematical models from the extracted mathematical models in descending order of the operation load. Step of assigning each of the extracted mathematical models to the individual processors, and the processor to which the mathematical model of the non-linear model is assigned, so that the operation load of each of the processors executing the simulation operation becomes substantially equal. Assign mathematical models of other nonlinear models, and A calculation method of assigning a plant simulator.

  According to the present invention, there is an effect that the simulation speed can be further increased.

It is a figure showing an example of a simulation system diagram of a thermal power plant considered as a simulation object of a plant simulator concerning one embodiment of the present invention. It is a figure showing hardware constitutions of a plant simulator concerning one embodiment of the present invention. It is a functional block diagram of a plant simulator concerning one embodiment of the present invention. It is the flow chart which showed the procedure of the processing performed by the plant simulator concerning one embodiment of the present invention. When the plant simulator which concerns on one Embodiment of this invention performs the process shown in FIG. 4, it is the figure which showed typically an example of the relationship between the mathematical model allocated to each CPU, and each apparatus.

  Hereinafter, a plant simulator and an operation assignment method of the plant simulator according to an embodiment of the present invention will be described with reference to the drawings. The plant simulator in the present invention simulates the behavior of a power plant such as a thermal power plant or a nuclear power plant, for example, and is, for example, an operation training simulator used for operation training. The plant simulator and the operation assignment method of the plant simulator according to the present invention can be widely applied to the simulation of a power plant, but particularly includes a boiler, a turbine, and a condenser as main components, and state quantities of flow rate, temperature and pressure. Is suitable for being applied to the simulation of a power plant having as an element. In the following embodiment, although the case where a thermal power plant is simulated is illustrated for convenience of explanation, the object by which the behavior is simulated by the plant simulator of the present invention is not limited to a thermal power plant.

FIG. 1 is a diagram showing an example of a simulated system diagram of a thermal power plant. As shown in FIG. 1, in the plant model simulation system diagram, various components required to simulate the behavior of the plant are connected by a block diagram based on the relationship in the actual thermal power plant, and the steam system , Cooling water system, fuel system etc. are formed.
In FIG. 1, the steam generated by the boiler 51 is supplied to the high pressure turbine 52. The steam that has worked in the high pressure turbine 52 is sent to the reheater 53 and heated, and then supplied to the medium pressure turbine 54 and the low pressure turbine 55 to perform work. The steam that has worked in the low pressure turbine 55 is cooled in the condenser 56 to be condensed water 61. The condensed water 61 is sent to the deaerator 57 to remove impurities, and then returned to the boiler 51 again as the water supply 62.

  In addition, high pressure steam is partially taken out in the vicinity of the reheater 53 and the high pressure turbine 52 and is used as the auxiliary steam 66. Further, the rotational power of the high pressure turbine 52, the intermediate pressure turbine 54 and the low pressure turbine 55 is transmitted to the generator 58, whereby the generator 58 generates power. A portion of the steam in the intermediate pressure turbine 54 is supplied to a steam driven feed water pump 59 and used to drive the feed water pump 59.

  Further, a portion of the water supply 62 and a portion of the condensate 61 are returned again as condensate from the condenser 56 to the deaerator 57 via the heater drain 60. Furthermore, in FIG. 1, fuel air 63, exhaust gas 64, and fuel 65 are elements of the fuel system, and supplemental water 67, seawater 68, vacuum 69, axial cold water 70, bearing oil 71, and the like are elements of the cooling water system. It is shown.

  FIG. 2 is a diagram showing a hardware configuration of a plant simulator according to the present embodiment. As shown in FIG. 2, the plant simulator is a so-called computer, and eight CPUs (processors) 11a to 11h and a main memory 12 functioning as a work area at the time of execution of each program are connected to a host bus 18 It is done. On the other hand, a hard disk drive (hereinafter referred to as "HDD") 14 in which various programs executed by the CPUs 11a to 11h are stored in the I / O bus 19 connected to the host bus 18 via the I / O bridge 13. An input unit 15 to which an input command is input by being operated by an operator (for example, a trainee), and a display unit 16 for displaying a simulation result and the like are connected.

  Eight CPUs 11a to 11h may be realized as a multi-core single processor in which a plurality of processor cores are mounted in one processor package, and one processor core is mounted in one processor package It may be realized by a single core multiprocessor comprising a plurality of single cores. Also, it may be realized as a multi-core multiprocessor having a plurality of multi-core processors. Although FIG. 2 illustrates the case where eight CPUs 11a to 11h are provided, the number of CPUs is not limited. Further, with regard to the number of CPUs, in order to improve the processing capacity more than necessary, it is preferable to reduce the number of CPUs rather than having too much room for necessary processing, in order to reduce the cost.

  FIG. 3 is a functional block diagram of a plant simulator according to the present embodiment. As shown in FIG. 3, the plant simulator includes eight simulation execution units 21a to 21h that execute computations related to simulation, and a mathematical model storage unit 25 in which a plurality of mathematical models for computing the behavior of the plant are stored. And an operation assignment unit 30 for assigning a mathematical model to each of the simulation execution units 21a to 21h.

  Each of the simulation execution units 21a to 21h is in one-to-one correspondence with each of the CPUs 11a to 11h. That is, in the present embodiment, eight CPUs 11a to 11h function as the simulation execution units 21a to 21h, respectively, and when the parallel processing is performed, the simulation calculation proceeds.

  The mathematical model storage unit 25 stores a plurality of mathematical models for calculating the behavior of the plant in association with the data types. The data type includes, for example, whether or not it is a non-linear model, and information (pressure, volume, temperature, etc.) of the state quantity to be calculated.

  The operation assignment unit 30 assigns mathematical models so that the operation load of each of the simulation execution units 21a to 21h is substantially equal. Here, "substantially equal" means that the difference from the average value of the calculation load assigned to each of the simulation execution units 21a to 21h is within the predetermined allowable range. The operation assignment unit 30 further assigns a mathematical model so that data types of the mathematical model assigned to each of the simulation execution units 21a to 21h are the same or similar as much as possible.

  The operation assignment unit 30 assigns, for example, a first operation assignment unit 31 (nonlinear operation assignment means) that assigns a mathematical model of a non-linear model, and a mathematical model other than heat transfer calculation (for example, pressure, volume, etc.) among linear models. Operation assignment unit 32 (pressure operation assignment unit, volume operation assignment unit) for assigning the second operation, and a third operation assignment unit 33 (heat transfer calculation assignment unit) for assigning the mathematical model of heat transfer calculation among linear models ).

  The first operation assignment unit 31 extracts, for example, a mathematical model whose operation load is equal to or greater than a predetermined operation load threshold among mathematical models of non-linear models, and is specified in descending order of operation load from the extracted mathematical models. Mathematical models of numbers are extracted, and the extracted mathematical models are assigned to different simulation execution units 21a to 21h. For example, when the predetermined number is set to “four”, the first operation allocation unit 31 extracts four mathematical models in descending order of the operation load, and calculates each mathematical model as four simulation execution units, for example, It allocates to simulation execution part 21a, 21b, 21e, 21f, respectively. If the mathematical model whose computational load is at least the computational load threshold is less than the "predetermined number", for example, the mathematical model whose "predetermined number" is "four" and the computational load is at least the computational load threshold is "2 In the case of “1”, these two mathematical models are respectively assigned to two simulation execution units.

  Then, with regard to the remaining mathematical models in the non-linear model, the calculation load of each of the simulation executing sections 21a to 21h including the simulation executing sections 21a, 21b, 21e and 21f to which the mathematical model of the non-linear model is already assigned Make the assignment so that is almost even. Here, that each operation load is substantially equal means that the upper and lower limits of each operation processing time of each of the simulation execution units 21a to 21h are ± 10 with respect to the average operation processing time calculated from each operation processing time. %, More preferably ± 5%.

  The “calculation load” can be determined based on the size of the calculation matrix. That is, since the mathematical model is represented by a matrix such as a Jacobian matrix, a predetermined number (for example, four) may be selected from the top according to the size (the number of column rows) of the matrix. If the size of the matrix is the same, the operation time can also be regarded as almost the same, and by assigning the mathematical model based on the matrix size, the mathematical model can be assigned easily and efficiently. . In this case, the operation load threshold is also set using the size of the matrix. Also, the “calculation load threshold” and the “predetermined number” may be implemented so as to be able to be set dynamically on the program, or may be set in advance according to the scale of the plant and simulation. Alternatively, the configuration may be such that the operator can input as appropriate. As described above, any method may be used to set the “calculation load threshold” and the “predetermined number”.

The second operation assignment unit 32 is a linear model, and among mathematical models other than heat transfer calculation, a mathematical model of pressure calculation of a steam system and a mathematical model of volume calculation of a combustion system (for example, combustion air, exhaust gas And a simulation execution unit to which the mathematical model of the non-linear model extracted by the first operation assignment unit 31 is not assigned. Assign each to 21h. Subsequently, the second calculation allocation unit 32 allocates the remaining mathematical models so that the calculation load of each of the simulation execution units 21a to 21h is substantially equal.

The pressure calculation of the steam system and the volume calculation of the combustion system, to which the second calculation allocation unit 32 preferentially allocates, occupy an important position in understanding the behavior of the power plant, and the calculation load is also relatively large. Because it is large, the operation time is also longer than the mathematical models of other linear models. Therefore, the mathematical model relating to the pressure calculation of the steam system and the volume calculation of the combustion system may be individually assigned to any simulation execution unit that does not calculate the non-linear model extracted by the first operation assignment unit 31. By this, it is possible to distribute the computational load efficiently. In addition, in these simulation execution units, since the same calculation is repeatedly performed, improvement in calculation processing speed can be expected.

The third operation assignment unit 33 is a linear model and extracts heat transfer calculation for piping, heat transfer calculation for valves, and heat transfer calculation for pump fan from mathematical models for electric heat calculation, and these are respectively individual Assign to the simulation execution unit preferentially. At this time, depending on the volume of each heat transfer calculation, a simulation execution unit to which a mathematical model of the non-linear model extracted by the first operation assignment unit 31 is allocated, that is, allocation to the simulation execution units 21a, 21b, 21e, 21f It may be Subsequently, mathematical models relating to the remaining heat transfer calculation are assigned so that the calculation load of each simulation execution unit becomes substantially equal.

  In the heat transfer calculation, the number of installed pipes, valves, and pump fans is large, and the calculation becomes complicated. For this reason, the calculation load becomes relatively large, and the calculation time becomes long. Therefore, with regard to mathematical models for heat transfer calculation of piping, valves, and pump fans, by preferentially assigning them to different simulation execution units, it becomes possible to disperse computational load efficiently.

Next, the process performed by the plant simulator according to the present embodiment will be described with reference to FIG.
First, at the start of simulation, a predetermined CPU (for example, the CPU 11a) reads out the main program stored in the HDD 14 to the main memory 12 and executes it so that the CPUs 11a to 11h can execute parallel processing normally. Perform preparation process to For example, the CPU 11a allocates programs for causing the CPUs 11a to 11h to function as the above-described simulation execution units 21a to 21h to the CPUs 11a to 11h, and allocates threads for division operation to the CPUs 11a to 11h (step SA1). Subsequently, the CPU 11a issues a calculation start event to the allocated thread and program (step SA2). Thus, preparation for each of the CPUs 11a to 11h to function as the simulation execution units 21a to 21h is completed.

Next, the CPU 11 a functions as a first operation assignment unit 31, a second operation assignment unit 32, and a third operation assignment unit 33 of the operation assignment unit 30 by executing various programs stored in the HDD 14.
Specifically, first, four mathematical models are extracted from the mathematical models of non-linear models stored in the mathematical model storage unit 25 in descending order of operation load, for example, in order of large size of Jacobian matrix, and extracted The respective mathematical models are respectively assigned to individual simulation execution units (first operation assignment unit: step SA3). Thus, for example, the mathematical model with the largest calculation amount is the simulation execution unit 21a, the mathematical model with the second largest calculation amount is the simulation execution unit 21b, and the mathematical model with the third largest calculation amount is the simulation execution unit 21e. The fourth largest mathematical model is assigned to the simulation execution unit 21 f.

  Next, the CPU 11a allocates the mathematical models of the remaining non-linear models stored in the mathematical model storage unit 25 so that the calculation load of the simulation execution units 21a to 21h is substantially equal (first calculation allocation unit: Step SA4).

  Next, the CPU 11a extracts a mathematical model related to pressure calculation of the steam system and a mathematical model related to volume calculation of the combustion system among the linear models stored in the mathematical model storage unit 25, and simulates the extracted mathematical model separately Assign to each execution unit. At this time, in step SA3, a simulation execution unit 21c, 21d, 21g, 21h different from the simulation execution units 21a, 21b, 21e, 21f to which a mathematical model of a nonlinear model with a relatively large operation load is allocated is allocated. (Second operation allocation unit: step SA5). Thereby, for example, a mathematical model of pressure calculation of the steam system is allocated to the simulation execution unit 21g, and a mathematical model of volume calculation of the combustion system is allocated to the simulation execution unit 21h. Subsequently, the CPU 11a allocates the remaining mathematical models so that the calculation load on the simulation execution units 21a to 21h is substantially equal (second calculation allocation unit: step SA6).

  Next, among the linear models stored in the mathematical model storage unit 21, the CPU 11a assigns the heat transfer calculation for the piping, the heat transfer calculation for the valve, and the heat transfer calculation for the pump fan to individual simulation execution units (No. 3 Operation allocation unit: step SA7). Thus, for example, the mathematical model of heat transfer calculation for piping is to the simulation execution unit 21g, the mathematical model of heat transfer calculation for the valve is to the simulation execution unit 21a, and the mathematical model of heat transfer calculation for the pump fan is to the simulation execution unit 21b. Assign each one. Subsequently, the CPU 11a allocates the remaining mathematical models of heat transfer calculation (for example, heat transfer calculation in a drum, deaerator, etc.) so that the calculation load of each of the simulation execution units 21a to 21h becomes substantially equal (third Operation assignment unit: step SA8).

In this way, when the assignment is completed for all the mathematical models stored in the mathematical model storage unit 25, the computations of the mathematical models by the simulation execution units 21a to 21h are performed in parallel (step SA9). And each simulation execution part 21a-21h notifies that to CPU11a, if calculation of all the mathematical models allocated to self is complete | finished.
When the CPU 11a receives the notification of completion of operation from all the CPUs 11a to 11h (step SA10), it returns to step SA2 and issues a calculation start event to each of the CPUs 11a to 11h (simulation execution units 21a to 21h). Start processing by an operation cycle (thread).

  FIG. 5 is a view schematically showing an example of the relationship between the mathematical model assigned to each of the CPUs 11a to 11h by executing the above-mentioned processing and each device. As can be seen from the boiler 51, the reheater 53, etc. in FIG. 5, it can be seen that even with one device, computations are assigned to a plurality of CPUs according to the computation type. Further, for example, the pressure calculation of the steam system for the reheater 53, the intermediate pressure turbine 54, and the low pressure turbine 55 is assigned to the CPU 11g. As described above, by causing one CPU 11 g to take charge of computations related to a series of flows of steam, it is possible to improve the efficiency of the computation processing. In particular, with regard to the intermediate pressure turbine 54, the supply destination is divided into the low pressure turbine 55 and the feed water pump 59 while the number of steam supply sources is one, and the calculation becomes complicated. Therefore, it is expected that further speeding up of the operation can be achieved by concentrating and executing the complicated operation on one CPU.

As described above, according to the plant simulator and the operation assignment method of the plant simulator according to the present embodiment, the following effects can be obtained.
The various mathematical models stored in the mathematical model storage unit 25 are assigned to the CPUs 11a to 11h (simulation execution units 21a to 21h) in units of data types, so, for example, the same data among a plurality of devices constituting the power generation plant It becomes possible to merge the operation contents belonging to the type and assign them to one processor.

  The assignment of the mathematical model is performed so that the calculation load of each of the CPUs 11a to 11h is substantially equal. Therefore, the calculation time of each of the CPUs 11a to 11h can be substantially equalized, and the simulator speed can be improved. . The fact that the calculation load here is substantially even means that the upper limit and the lower limit of each calculation processing time of each simulation execution unit is within ± 10% of the average calculation processing time calculated from each calculation processing time, more preferably Indicates a state within ± 5%. Further, since a mathematical model having a large computational load is preferentially allocated to the processors, it is possible to relatively easily allocate the mathematical model so that the computational load of each processor is substantially equal.

  The assignment of the mathematical model by the operation assignment unit 30 is performed at the start of operation of the plant model, that is, at the start of the simulation. Therefore, the assignment can be dynamically changed in accordance with the operation state of the power plant from time to time. As described above, even when the operating state of the power generation plant changes, the mathematical model is allocated so that the calculation load of the CPUs 11a to 11h becomes even following the change, thereby depending on the operating state of the power generation plant As a result, it is possible to avoid phenomena such as a significant decrease in simulation speed and to maintain a stable simulation speed.

  In the present embodiment, a mathematical model is assigned to the CPU 11a functioning as the operation assignment unit 30 as well as the other CPUs 11b to 11h and functions as the simulation execution unit 21a, but the CPU 11a is used from the OS In many cases, it is often used to execute processes other than simulation calculations, as compared with other CPUs 11 b to 11 h. From this, for example, the CPU (CPU 11a in the above example) that executes the main program or assigns a mathematical model functions as a simulation execution unit so as not to inhibit the movement of the OS or the like. A flag may be provided to eliminate the As a result, when the flag is set, it is possible to prohibit the function as a simulation execution unit, and thus it is possible to avoid the inconvenience such as the movement of the OS being hindered.

  In the embodiment described above, as described with reference to FIG. 4, the first operation assignment unit 31 to the third operation assignment unit 33 perform the processing in order, but the first operation assignment unit 31 to Each of the third operation assignment units 33 may execute processing in parallel. Thus, by performing parallel processing, it is possible to reduce the allocation time.

  The present invention is not limited to the above-described embodiments, and various modifications may be made without departing from the scope of the invention, for example, by partially or wholly combining the above-described embodiments. It is.

11a to 11h CPU
12 main memory 13 I / O bridge 14 HDD
15 input unit 16 display units 21a to 21h simulation execution unit 25 mathematical model storage unit 30 operation allocation unit 31 first operation allocation unit 32 second operation allocation unit 33 third operation allocation unit 51 boiler 52 high pressure turbine 53 reheater 54 Pressure turbine 55 Low pressure turbine 56 Condenser 58 Generator

Claims (7)

A plant simulator that simulates the behavior of a power plant by a plurality of processors performing simulation operations in parallel,
Mathematical model storage means in which a plurality of mathematical models for calculating the behavior of the plant are stored in association with the data type;
Operation assigning means for assigning the mathematical model to the processor for each data type;
One of the data types is a non-linear model,
The operation assignment means extracts the mathematical model having a predetermined operation load or more, and preferentially assigns the mathematical model among the extracted mathematical models in descending order of the operation load, and the operation of each processor so that the load is substantially uniform, we have row assignment of the mathematical model,
The operation assignment means includes non-linear operation assignment means for assigning the mathematical model of a non-linear model to the processor,
The non-linear operation assignment means
The mathematical model having a preset operation load or more is extracted from the mathematical models of the non-linear model, and a predetermined number of the mathematical models are extracted from the extracted mathematical models in descending order of the operation load, and extracted Assigning each of said mathematical models to a respective one of said processors;
Assigning mathematical models of other non-linear models such that the processors of the non-linear models, to which the mathematical models are assigned, are approximately equal in operation load of each of the processors executing the simulation operation;
Plant simulator to run .   The plant simulator according to claim 1, wherein the operation assignment unit assigns the mathematical model such that a mathematical model of the same or similar data type is assigned to one processor. The plant simulator according to claim 1 or 2 , wherein the operation allocation unit determines the operation load based on the size of the operation matrix. The operation assignment means assigns a pressure operation to any one of the linear models to which the mathematical model of the pressure calculation of the steam system is not assigned, the mathematical model of the non-linear model extracted by the non-linear operation assignment means. The plant simulator according to any one of claims 1 to 3, further comprising an assignment means. Among the linear models, the operation assignment means is assigned a mathematical model of volume calculation of the combustion system, and a mathematical model of the non- linear model extracted by the non-linear operation assignment means is also assigned a mathematical model of pressure calculation of the steam system. The plant simulator according to claim 4 , further comprising volume operation assignment means for assigning to any one of the processors. Assigning said computation assignment means, of the linear model, heat transfer calculation mathematical model, the mathematical model of the heat transfer calculations for the valve, and each was dispersed in different said processor a mathematical model of the heat transfer calculations for the pump fan about the pipe plant simulator according to any one of claims 1 to 5 comprising a heat transfer calculation allocating means. An operation assignment method applied to a plant simulator that simulates the behavior of a power generation plant by a plurality of processors performing simulation operations in parallel.
The plant simulator includes mathematical model storage means in which a plurality of mathematical models for calculating the behavior of the power plant are stored in association with data types.
A computer executes an operation assigning step of assigning the mathematical model to the processor for each of the data types;
One of the data types is a non-linear model,
In the operation assignment step, the mathematical model having a preset operation load or more is extracted, and the mathematical model is preferentially assigned in descending order of the operation load among the extracted mathematical models, and the operation of each processor is performed. so that the load is substantially uniform, we have row assignment of the mathematical model,
The operation assignment step comprises a non-linear operation assignment step of assigning the mathematical model of a non-linear model to the processor;
The non-linear operation assignment step
The mathematical model having a preset operation load or more is extracted from the mathematical models of the non-linear model, and a predetermined number of the mathematical models are extracted from the extracted mathematical models in descending order of the operation load, and extracted Assigning each of said mathematical models to a respective one of said processors;
Assigning mathematical models of other non-linear models such that the processors of the non-linear models, to which the mathematical models are assigned, are approximately equal in operation load of each of the processors executing the simulation operation;
Method of operation assignment of a plant simulator having

Images