JPH0438434A - Simulation method of air current - Google Patents

Abstract

PURPOSE:To facilitate an inputting operation for anyone on an image indication by deviding input data into three groups according to importance and successively asking for demand factors on culculation at three stages. CONSTITUTION:An input measure 10, an information base 12, a reasoning measure 14 and an output measure 16 for a simulation method of air current to be carried out are provided. The input measure 10 can input a data necessary to properly control an air current of an objective space and is devided to the first to the third input measures according to importance. Namely, the first input measure, a condition being a major premise, can input an objective space dimension, a minimum exhaust hole dimension and objective things (a clean room, an atomic energy building, a food work, a general building, etc.). The second input measure, a condition being a medium premise, can input a kind of air currents and an analysis time desired. Further, the third input measure, a detail item, can input an arraingement of block classes. Therefore, input data are devided into three groups according to importance and important factors on culculation can be successively obtained for at three stages.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a fluid stain 5 radiation method, and more particularly to a method for simulating airflow necessary for detailed design of clean rooms, food factories, nuclear power buildings, etc.

[Conventional technology]

In general, when designing detailed designs for clean rooms, food factories, nuclear power buildings, etc., airflow shimilation is done in one go, and in order to generate an appropriate airflow, the position of the outlet or inlet, the presence or absence of a resistor, the wind speed, etc. Consider.

[Problem to be solved by the invention]

However, in order to converge the calculations, the airflow simulation method requires appropriate settings of various parameters (mesh division time increments, mesh division number, number of time repetitions, etc.), and these settings are performed by skilled experts. is needed.

The present invention has been made in view of these circumstances, and it is an object of the present invention to provide an airflow simulation method that allows even a designer other than a skilled expert to easily simulate the airflow in a target space.

[Means to solve problems]

In order to achieve the above object, the present invention repeats the simulation until the airflow of each cell in which the target space is mesh-divided reaches a steady state, and the resistance of the target space is adjusted so that the steady airflow in each cell becomes an appropriate value. In an airflow simulation method that allows setting conditions such as the number of simulation repetitions, the calculation time of the airflow in the mesh-divided target space, and the time until the airflow in each mesh-divided cell reaches a steady state. Preliminarily store each relationship in the knowledge means, and calculate the number of mesh divisions according to conditions such as the resistor in the target space so that the airflow is in a steady state based on the relationship between the three conditions. It is characterized by

In addition, in order to achieve the above object, the present invention divides the setting conditions of the resistor etc. in the target space into three in order of priority, the space dimension of the first input means, the minimum exhaust hole opening dimension, the symmetrical object, The time step for mesh division is calculated from each data of the airflow type, the number of mesh divisions is calculated from the data of the airflow type of the second condition, the desired analysis time, and the time step, and the number of mesh divisions is calculated from the third condition.
The present invention is characterized in that the precise time increments are determined from the arrangement data of the blocks of detailed items of the conditions, and the number of repetitions of the simulation is determined from the time increments.

[Effect]

According to the present invention, when setting conditions such as resistors when designing a target space, the data necessary for airflow simulation can be stored in advance in the knowledge means, so that conventionally skilled experts can enter the data at the time of simulation. Airflow simulations can be performed without inputting the previously used data.

In addition, since the human-powered data such as resistors is divided into three priority levels, it is possible to obtain the most suitable data at the required time.

〔Example〕

Preferred embodiments of the airflow simulation method according to the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 shows an input means 10, a knowledge base 12, an inference means 14, and an output means 16 for executing the airflow simulation method according to the present invention.

The input means 10 can manually enter data necessary for optimizing the airflow in the target space, and as shown in FIG. It is divided. That is, the main conditions, the first input means 18, can input the target space dimensions, minimum exhaust hole opening dimensions, and target objects (clean room, nuclear power building, food factory, general building, etc.). Also, the second input means 2
0 can be manually set the prerequisite conditions, type of airflow, and desired analysis time. Furthermore, the third manual means 22 can input the arrangement of blocks, which is a detailed item. Therefore, human data can be grouped into three groups according to their importance, and important factors for calculation can be sequentially determined in three stages.

Various parameters are recorded in advance as data in the knowledge base 12, and the first to fourth knowledge bases 24, 26, 28
．． It is divided into 30 parts. And the first knowledge base 2
4 records the empirical formula for calculating the reference time repetition number and time increments, and the second knowledge base 26 contains a frame that summarizes the relationship among airflow, time increments, analysis time, and mesh division number in a table. 8 self-recorded. In addition, the third knowledge base 28 records rules relating to reference length, speed, aperture ratio, and time increments, and the fourth knowledge base 30 records rules relating to time increments and the number of time repetitions. ing.

The inference means 14 receives data inputted from the input means IO,
Necessary data can be calculated from the parameters recorded in the knowledge base 12.

This reasoning means 14 includes first to fourth reasoning means 32.3.
It is divided into 4.36.38. First reasoning means 32
is the data of the first manual means 18 and the first knowledge base 2
The time step for mesh division can be calculated from the parameter 4.

Further, the second inference means 34 inputs the data of the second input means,
The number of mesh divisions can be inferred from the parameters of the 12 knowledge bases 26 and the time step for mesh division inferred by the first inference means 32. Third reasoning means 3
6 can infer a calculation time step based on the data of the third input means, the parameters of the third knowledge base 28, and the number of mesh divisions inferred by the second inference means 34. Furthermore, the fourth inference means 38 can infer the number of time iterations based on the parameters of the fourth knowledge base 30 and the calculation time steps inferred by the third inference means 36.

Hereinafter, the airflow stain-screening method according to the present invention will be explained based on FIG. 2.

First, the target space dimensions, minimum exhaust hole opening dimensions, and data of the target property are input using the first input means 18 to set the target space 50 shown in FIG. 3(B). In addition, 52 on Figure 3 (B)
indicates a resistor in the target space 50, A indicates air supply, and B indicates exhaust air.

Next, air supply A is started in the target space 50 under no-wind conditions, and the momentary airflow generated by the air supply is analyzed in accordance with a predetermined calculation formula. This analysis shows that the airflow is steady as shown in Figure 3 (A)! ! Basically, it is repeated until (T2) is reached.

At this time, the first inference means 32 calculates the time step Δt for mesh division based on the empirical formula for calculating the number of time repetitions n1 time steps recorded in the first knowledge base 24.

At this time, the time step Δt for division is set finely (for example, 0.0
1SeC) improves accuracy, and roughness (for example 0.1
sec), the accuracy will decrease. Also, the number of time repetitions n
An output that is closely related to the time (Tc) until the airflow reaches a steady state and the division time step Δt can be obtained.

However, Tc varies depending on the target space, and if the number of time repetitions n is set too small after setting the division time division Δt, the analysis time T1 will be shortened as shown in Figure 3 (A), but the final The output is in an unsteady state and the analysis accuracy decreases. On the other hand, if the number of time repetitions n is set too high, the final output will only be slightly higher than the output at Tc, even though the analysis time T2 becomes longer, and the analysis time will be wasted.

Therefore, it is important and difficult to maintain this balance. In particular, when the difference between the initial room temperature and the supply air temperature is large, the convection phenomenon becomes more intense, and the more complex the shape of the target space, the more difficult it becomes to estimate Tc. Therefore, conventionally, a skilled expert was required to set Tc, etc., but by recording the expert's know-how in the knowledge base 12, this is no longer necessary.

Next, the number of mesh divisions is determined using the second inference base 34. In this way, the reason why the mesh division number is determined after the mesh division time step Δt is to specify the analysis time of the gas simulation and perform the fluid sea% 3 ray a of the target space at an appropriate time. This is because if the number of mesh divisions into two is determined first regardless of the mesh division time increment, the calculation may not converge (fall within the allowable range of appropriate virtual values).

By the way, it is not necessarily the case that the accuracy of the output obtained as a result of simulation is good; there are many cases where it is desired to obtain the output quickly even if the accuracy is somewhat poor. For this reason,
The desired analysis time can be input using the second input means 20. By inputting these conditions manually using the input means 20, various calculation factors (time step Δt1 for division, number of meshco divisions, number of time repetitions n) are calculated using the knowledge base 12.
The inference means 14 selects the value of the most appropriate condition (condition that satisfies Tc and desired analysis time).

However, in the present invention, the desired analysis time is determined by the first input means 1.
8. In other words, it is not a major premise. The reason for this is that in order to make the airflow simulation effective, it is necessary to
The data items shown in the input means 18 are the main premise,
The desired analysis time is positioned as the second manual means 20, that is, as a medium premise. If it is determined based on the knowledge base 12 that an effective output cannot be obtained within a predetermined time, a screen instruction may be given to extend the analysis time.

Next, the third inference base 3 is based on the number of mesh divisions.
Step 6 calculates the time step for calculation. The reason for this is that the previously determined mesh division time step Δt is rough because it is used to determine the number of mesh divisions. Therefore, since the flow rate is constant, if there is anything that acts as a resistance to the airflow, such as actual equipment, gratings, shelves, or narrowed areas, the airflow will flow outward from that area, increasing the flow rate and causing the solution to diverge. Also, the number of mesh divisions cannot be determined without the division time step Δt,
Furthermore, for the above-mentioned reason, the time step for calculation cannot be determined until after mesh division.

When calculating this calculation time step, the third knowledge base 2
The reason why the reference length, velocity, aperture ratio, and time step recorded in 8 are related is that Courant's condition exists as a stability condition when numerically solving the difference equation of airflow simulation.

The coolant conditions are C1Δt/Δx<1 (C: flow rate,
Δt: time step width, ΔX: spatial step width), and based on this Courant condition, Δt is calculated for cases with and without grating.

Note that the reference length is the shorter of the mesh widths in the X direction and the y direction, the speed is the maximum inflow and outflow speed, and the opening ratio is the smallest area occupation rate among the grating floors, shelves, etc.

Next, consider the ventilation of the target space more than once, and if a convection phenomenon is expected, such as when there is a difference between the initial indoor temperature and the supply air temperature, decide to perform the calculation until several ventilations are performed. Execute.

When simulating airflow using CAE (Computer Aided Experiment) in this way, input data can be input using a keyboard or a mouse. In addition, since the mouse is easy to operate manually, it is possible to prevent careless mistakes when using the mouse manually.

〔Effect of the invention〕

As described above, according to the airflow simulation method according to the present invention, human data can be grouped into three groups according to their importance, and calculational demand factors can be sequentially determined in three stages. Depending on the desired analysis time, the most suitable level (M degrees) of output can be input in that time, and anyone can easily input the input based on image instructions.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of a device that executes the airflow simulation method according to the present invention, Fig. 2 is a detailed block diagram thereof, Fig. 3m(A) is a graph showing the state of airflow during simulation, and Fig. 3 [ B) is a schematic plan view of the target space in which the simulation is executed. 10--Input means, 12--Knowledge base, 14-
... Reasoning means, 16... Output means. 1 Figure 10...Input means, 14...Inference means, 12.Knowledge base, J6 Output means,

Claims (2)

[Claims] (1) The simulation is repeated until the airflow in each cell in which the target space is divided into meshes reaches a steady state, and conditions such as resistors in the target space can be set so that the steady airflow in each cell reaches an appropriate value. In the airflow simulation method that can be used, the relationship between the number of simulation iterations, the calculation time of the airflow in the target space divided into meshes, and the time it takes for the airflow in each meshed cell to reach a steady state is determined in advance by means of knowledge. A method for simulating airflow, which comprises storing the information and calculating the number of mesh divisions according to conditions such as a resistor in the target space so that the airflow is in a steady state based on the relationship between the three conditions. (2) Divide the setting conditions for resistors, etc. in the target space into three groups in order of priority, and use the data of the first input means, such as space dimensions, minimum exhaust hole opening dimensions, symmetrical object, and airflow type, to create mesh division-like settings. The time increments are calculated, and the number of mesh divisions is determined from each data of the airflow type and desired analysis time in the second condition, and the time increments, and the precise An airflow simulation method, characterized in that the time step is determined, and the number of simulation repetitions is determined from the time step.