JPH048282A - Method for controlling solid culture apparatus - Google Patents

Abstract

PURPOSE:To save labor and electric power and improve quality by controlling culture conditions according to fuzzy control using culture factors as input variables. CONSTITUTION:Conditioned air is fed from an air conditioner 3 into a KOJI (yeast) layer 2 on a rotating disk and a wind pressure sensor 5 is installed in an air feed chamber 4. A material temperature sensor 6 is provided in the layer 2 to input measured results of the sensor 6 to a state recognition input device 7 by on-line operation. The program set value of the material temperature is inputted through a keyboard to the device 7 so that the elapsed time of culture may be measured. Thereby, an apparatus 1 for producing soy sauce KOJI, etc., is obtained. The device 7 is then connected to a manipulated variable output device 8, an arithmetic part 9 of fuzzy inference and a regulating part 10 of the fuzzy control to carry out program setting of input variables ( TEMP, STEMP, wind pressure and TIME) and output variables (wind flow rate, wind temperature and handling). A command is outputted from the device 8 to an inverter 11, a fan 12, a wind temperature regulator 13, the air conditioner 3, a heating heater 14, etc., and control is performed to operate a handler 17 and control solid culture.

Description

[Detailed description of the invention]

[Industrial Application Field 1] The present invention optimizes the culture conditions of a solid culture device, such as a culture device used for making koji, culturing for enzyme preparations, producing antibiotics, etc., without relying on the intuition of experts. This invention relates to an automatic control method. [Prior Art 1] A conventional control method for a solid-state culture device is explained below using the example of soy sauce koji. ■To ensure that the air temperature and product temperature are in line with the product temperature. The amount of air blown is controlled according to a preset wind temperature line. ■Amount of air blown The amount of heat generated from koji differs depending on the season, so change the amount of air blown depending on the time setting. ■Care period Care is performed by setting a time that matches the calorific value of the koji. Problem 1 to be Solved by the Invention The above control method has the following problems. ■Since the air temperature is not controlled based on the product temperature, the product temperature does not follow the product temperature line. (5) Case of controlling air temperature based on product temperature 5 With conventional PID control, there is a time delay of 15 to 30 minutes, and the temperature often becomes abnormal, and control is sometimes impossible. Therefore, they had no choice but to control the air temperature and air volume so that the air temperature and air volume were set in advance. ■The air volume is determined based on the air temperature, but since both are preset, the air volume is not optimal for the heat generation status of the koji, and the blower consumes a lot of power. ■The maintenance period is determined by the actual state of the koji, the expected state of the koji regardless of the calorific value of the koji, and the calorific value of the koji, so it is not always possible to maintain it at the appropriate time. Not exclusively. If automatic maintenance is carried out at a timing earlier than the state of the koji, the temperature of the koji will become too high later on, and wind alone will not be able to lower the temperature of the koji. If the timing is too late, the temperature of the product will become too high. Normally, it is best to perform maintenance at the last stage when the wind can suppress the heat generation, and this depends on the intuition of an expert. [Means for solving the problem] As a result of studying the above problems, we found that a control method that determines the manipulated variable to match the control target by feeding back the measured values of the process, typified by conventional PID control, is based on microbial culture. It has become clear that the complex factors associated with this cannot be addressed. Therefore, we developed a method for controlling a solid-state culture device, which is characterized by controlling culture conditions by fuzzy control using culture factors as input quantities. In addition to fuzzy control, it is possible to use optimal control, expert system control, sampling control, etc. However, optimal control is a control method that determines the state equation of the process and uses it to determine the manipulated variable so that the error with the evaluation standard is minimized, but the problem is whether the state equation can be found. Even if it is possible to obtain it, there is also the question of whether linearization is possible. Control by an expert system (excluding fuzzy control) is performed by regulating the amount of operation in advance according to each state. This makes the rules very complex. Sampling control is a method that does not perform PID control continuously, but performs PID control intermittently according to dead time, but it is difficult to stabilize if the system is complex. Therefore, fuzzy control is the most preferred mode. The fuzzy control referred to here is a control method based on the generally known fuzzy theory. The aim is to use fuzzy theory to model the operation method that was previously performed by operators during cultivation, which was controlled through flexible adaptability based on ambiguous information, and to implement it using computers. In more detail, for example, if there is a difference between the JI product temperature and the program setting value, the process situation can be determined based on the condition fate (
Antecedent Subcommittee M) considers that when the proposition is certain, the criteria for determining the process situation and the details of the operation are specified in control rules that conclude with an operation method such as changing the air flow rate through the koji layer (consequent proposition). It is treated as an ambiguous quantity, and the ambiguous quantity is quantified by a membership function, such as when the temperature of the koji product starts to rise slightly (antecedent proposition), and the air temperature is lowered slightly (consequent proposition). This is a control method in which the amount of operation is determined as a comprehensive judgment weighted by the suitability of the antecedent proposition of the control rule. As a comprehensive judgment procedure for this purpose, first, it is preferable to adopt the Max-Min method, which is generally used as an inference method. For example, among the control rules during culture, the first rule in Table 2, which will be described later, is: IF TIME=TM-A, ΔTEMP=NB T
) IEN air temperature = PB air volume = SA. In this control rule, the input/output variable names and input/output variables are shown in Table 1. When TIME is the elapsed culture time and ΔTEMP is the difference between the koji product temperature and the program setting value, TIME
If is TM-A and ΔTENP is NB, this means that the manipulated variables of the consequent part, the air temperature, are PR and the air volume is SA. NB, PB, and SA are fuzzy variables each representing an ambiguous state. Examples of the antecedent part fuzzy variables and membership functions are shown in FIG. 2 of the embodiment described later, and the consequent part fuzzy variables and membership functions are also shown in FIG. This membership function is a set of fuzzy variables for quantifying ambiguous quantities. The fitness of each fuzzy variable is determined by the antecedent membership function according to the input value, and based on this, the membership function regarding the output is determined by calculation. The operating force value is obtained as a point that bisects the area of the membership function regarding the output according to the usual method. These are calculated by a microcomputer or a control computer when measured values are periodically input. This operation output value is flexible and highly adaptable using ambiguous information, and if there are factors that cannot be measured online among the judgment factors of a skilled operator, it is processed along with the keyboard input to determine the appropriate solid state culture device. is controlled. The solid culture device referred to here is a device for culturing microorganisms, animal and plant cells on a solid medium (substrate) such as agar, gelatin, silica gel, wheat bran, soybean, rice, barley, etc., and is used for culturing cells of microorganisms, animals and plants on a solid medium (substrate) such as miso, soy sauce, sake, shochu, etc. Koji production equipment refers to equipment that produces microorganisms themselves, enzymes, antibiotics, etc. by culturing microorganisms. In that case, the culture factors are the amount of change in temperature of one product over time, the difference between the product temperature and the desired elapsed value of product temperature, the amount of change over time, the weight of the substrate, the amount of change over time, the amount of variation in product temperature, Wind pressure, elapsed culture time, air volume, air temperature, wandering air temperature, air humidity, exhaust air humidity, air conditioner shower water temperature, humidity inside the culture device, calorific value per unit, amount of change in heat generation over time, maintenance period, carbon dioxide concentration, substrate Moisture 5 The degree of penetration of microorganisms into the substrate, amount of fragrance components, outside temperature, outside air humidity, amount of electricity used, amount of carbon dioxide generated, etc. The culture conditions include air flow rate, air temperature, air humidity, maintenance period, carbon dioxide concentration, humidity in the culture apparatus, air conditioner shower water temperature, and the like. Please note that the results of online measurement are input into the computer, but there are cases where online measurement is not possible. It is also possible to input the operator's operation judgment directly from the keyboard. Further, control using a combination of fuzzy control and a conventional control method is also possible.

[Effect]

According to the control method of the solid-state culture device as described above, miso,
When a culture factor is input in the production of koji such as soy sauce, sake, etc. using solid culture equipment, a comprehensive judgment is made as a fuzzy inference using the suitability of the antecedent proposition of each control rule as weight. The amount is output. In the fuzzy inference used, the criteria for determining the process situation and the contents of the operation are treated as ambiguous quantities, but these are quantified using membership functions, processed by the computer, and controlled for optimal cultivation. is done automatically.

【Example】

The present invention will be explained in detail below with reference to Examples. Example (i) Example of culture control of soy sauce koji FIG. 1 is a system diagram of a koji production apparatus for soy sauce koji. An air conditioner (
Conditioned air is supplied from 3). There is a wind pressure sensor (5) in the air supply chamber (4) below the rotating disk.
Moreover, there is a product temperature sensor (6) in the koji layer (2). The measurement results of the product temperature sensor (6) are sent to the status recognition input device (7)
It is now possible to enter information online. The programmed product temperature setting value is entered into the state recognition input device (7) using a keyboard. Furthermore, the elapsed culture time is also measured. There is a calculation section (9) for fuzzy inference between the state recognition input device (7) and the manipulated variable output device f (8), and this is connected to the fuzzy control rule section (10). Manipulated amount output bag! From (8), the air volume and air temperature of the koji making equipment,
Output commands are issued for control of all operating parts related to maintenance. Regarding the air volume, the rotation of the fan (12) is controlled via the inverter (11). The wind temperature is measured by the air conditioner (3) via the wind temperature controller (13).
) Controls the heater (14) to control the temperature of the shower water supplied to the inside of the fin cooler (14), which takes in outside air as needed
15) and the exhaust duct (16) so that the opening degree of the damper can be controlled. For maintenance, use the manipulated variable output device (
3) and the operating motor of the care machine (17) of the koji making apparatus are connected to receive control commands. The input/output variable names (labels) and input/output variables in the computer control by the fuzzy control rule unit (10) and the calculation unit (9) for fuzzy inference in this embodiment are
Shown in the table. The fuzzy control rules are as shown in Table 2 below. The antecedent fuzzy variables and membership functions in fuzzy partitioning are shown in Figure 2 (a3 to (d)). Figures 3 (a) to (c) also show the same length antecedent fuzzy variables and membership functions. When the above-mentioned culture factors are input online, a comprehensive judgment is made in the device as fuzzy inference using the suitability of the antecedent proposition of each control rule as a weight, as described above. The manipulated variable is output. Using the above solid culture device and control method, the culture type,
Table 3 shows the results of a fuzzy control test with varying input and output amounts. Below is a blank space in Table 2.Example (...) Example of making malt for rice koji for sake In the case of rice koji for sake, the operating conditions must be determined depending on the growth state of the koji mold. Therefore, in the past, the brewer made koji by determining the calorific value of the koji, the water content of the koji, the level of decontamination of the koji, the amount of flavor components in the koji, and other factors. However, when transitioning to machine-made koji, there was no problem if the brewer decided the operating conditions based on his judgment, but when fully automated, it is considered impossible to operate in accordance with the growth state of the koji mold. However, by inputting these parameters by measuring them online (or by measuring them offline and inputting them manually), it is possible to comprehensively measure the product temperature, humidity, and carbon dioxide amount by manually inputting other conventional mechanical koji sensors. By making inferences and controlling the air volume, air source, air humidity, maintenance period, carbon dioxide concentration, humidity inside the koji making equipment, and air conditioner shower water temperature, it is possible to produce koji that is equivalent to the operation determined by the master brewer. did it. Table 4 shows the results of a fuzzy control test using the solid culture device and control method by changing the culture type, input amount, and output amount.

【Effect of the invention】

As described above, the control method for a solid-state culture device of the present invention has limitations in using a mathematical model, and furthermore, it requires input of various ambiguous culture factors that cannot cope with the complex factors that are based on the use of microorganisms during culture. By making it possible to control culture conditions using fuzzy control, the following effects were obtained. ■The product temperature during culturing followed the product temperature setting curve. ■Therefore, by simply inputting the desired product temperature, there is no need for manual correction control or fine-tuning (manual) based on seasonal factors (outside temperature, etc.). ■By inputting the amount of electricity used, operating conditions that can produce the same amount of koji even with a smaller amount of electricity used can be selected, reducing running costs. ■Now is the ideal time to take care of your hair. ■From the above, it has become possible to operate unmanned, which saves labor, and has the effect of improving the quality of cultured products. ■It is now possible to handle things that were previously treated as ambiguous quantities and could not be reflected as input quantities for automatic control.
Quality has improved. ■By regularizing the control method based on the experience of manufacturing engineers, it is reflected in automatic control, and it has become possible to produce products that are equivalent to or better than manual control by manufacturing engineers.

[Brief explanation of the drawing]

FIG. 1 is a system diagram of a soy sauce koji production apparatus. Second
Figures (a) to (d) and Figures 3 (a) to (C) are diagrams showing membership functions. (1) Koji making equipment (2) Koji layer (3) Air conditioner (4) Air supply room (5) Wind pressure sensor (6) Product temperature sensor (7) Condition recognition input device! (8) Manipulated amount output device (9) Fuzzy control calculation section (10) Fuzzy control regulation section (11) Inverter (12) Fan (13) Air temperature controller (14) Heater (15) Fin cooler (16) Exhaust duct (17) Care machine Fig. 1, p. J2 (Q) Fig. 2 (b)

Claims (1)

[Claims] 1. A method for controlling a solid-state culture device, characterized in that culture conditions are controlled by fuzzy control using culture factors as input amounts.