JP3755026B2 - Method for estimating rice polishing ratio or broken rice rate of brewing rice - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention performs a spectral analysis of the acoustic power when milling rice for brewing raw materials, thereby estimating and controlling the degree of milling of rice, the rate of pulverization, and the operating status of the milling machine, thereby improving the quality of the product. It also relates to a technology for efficiently polishing rice .
[0002]
[Prior art]
Conventionally, rice brewed from brewing rice requires 10 to 13 hours to obtain 70% polished rice, and 50 to 70 hours to obtain 40% polished rice for ginjo sake. During this period, the rice milling ratio is estimated by stopping the milling machine, collecting white rice during the milling process, measuring the total weight with a load cell or the like, and calculating the apparent rice milling ratio. What can be known by this weight measurement is the apparent rice milling ratio. To know the true rice milling ratio excluding broken rice, the size of white rice is selected with the naked eye during milling and measured with a precision balance. There is nothing else. Since this operation is extremely complicated, it is rare to measure the polished rice ratio.
[0003]
Also, as a skillful standard for rice milling operations, the shape of white rice is closer to that of brown rice, and the incidence of cracking and broken rice is low, but the operating method for determining these conditions is that of the Kongo roll inside the rice milling machine. The rotation speed, the load resistance of the discharge port, and the opening degree of the discharge port are adjusted. Since these usually differ depending on the properties of the raw rice and factors such as the temperature and humidity in the milled rice chamber, the milling process over a long period of time is not constant and needs to be adjusted for each milled rice.
[0004]
However, at present, it is impossible to continuously monitor rice quality changes during the milling process without interrupting the milling operation. It is almost impossible to make a judgment on the way, and the actual situation is that the operation of the rice milling machine depends on the experience and intuition of skilled workers. In addition, when it is necessary to perform high-grade whitening such as raw materials for ginjo sake, it is necessary to monitor the rice milling machine all night, so that at the sake brewery, skilled rice polishing technicians and technical successors are due to labor issues. There is a shortage, it is forced to outsource rice milling to the outside, it is difficult to stably obtain raw rice with good brewing suitability.
[0005]
[Problems to be solved by the invention]
An object of the present invention is to provide a technique for improving the brewing suitability of raw cereals such as brewing raw rice and efficiently scouring. The present invention makes it possible to control a rice milling machine, which has been performed based on the experience and intuition of a conventional expert, in a real-time state by a computer. The degree can be measured. In addition, it is possible to measure the rate of crushed rice and the ratio of true rice during the milling of brewing rice, which has been difficult to measure in the past, so that it is possible to remotely determine the suitability of the rice milling operation and the operating state of the rice mill.
[0006]
[Means for Solving the Problems]
In order to achieve the above-mentioned object, the present inventors have conducted various studies from various directions, but eventually did not achieve success. In response to the need for a radical change in ideas, he conducted extensive and earnest study, research, and focused on the “sound” of rice milling.
[0007]
Then, the present inventors sampled the sound generated during the milling of the brewing raw rice rice with a precision sound level meter and digitally recorded it with DAT. This acoustic data was decomposed into a power spectrum by performing a fast Fourier transform. A frequency component having a high correlation with the rice polishing rate was extracted from the sound pressure level (dB) at a frequency of 50 Hz from 0 to 12,000 Hz. Furthermore, the frequency which has a high correlation with the rice polishing rate and the broken rice rate was selected from these frequencies by multiple regression analysis, and the prediction formula of the rice polishing rate and the broken rice rate based on the sound pressure level at this frequency was made. Furthermore, based on the principal component analysis, we extracted the principal components and created a grouping map based on the percentage of polished rice.
[0008]
That is, the present invention analyzes the acoustic data at the time of rice milling to estimate the rice milling ratio and the brewing suitability of the brewing raw material rice in real time, and to control the rice milling machine based on this result. It is.
Hereinafter, the present invention will be described by taking rice brewed rice as an example.
[0009]
DETAILED DESCRIPTION OF THE INVENTION
An ordinary condenser microphone is sufficient as the sound data sampling apparatus used in the present invention. The recording device may be anything as long as it has a low noise. For the calculation for decomposing the acoustic data into the power spectrum, an ordinary personal computer is sufficient as long as it has general-purpose software capable of fast Fourier transform.
[0010]
When estimating the polishing ratio and broken rice rate in real time during the brewing Rice polished rice, it is necessary to create a predictable expression. In the rice milling of brewing rice, some sounds during the milling are recorded according to the change in the milling rate, and the frequency is highly correlated with the milling rate by decomposing it into a power spectrum. The frequency measurement range at this time is suitably 0 to 20,000 Hz, and the frequency step size is suitably 50 to several hundred Hz. Generally, with the progress of rice milling, the sound pressure level in the high frequency band of 5,000 Hz or higher is relatively increased, and the sound pressure level in the low frequency band is relatively decreased. Multiple regression equations are created using several wavelengths that are highly correlated with the rice milling rate and the broken rice rate, respectively, and are used as prediction equations. At the time of rice polishing, (1) recording of acoustic data, (2) acquisition of power spectrum of acoustic data by fast Fourier transform, (3) extraction of sound pressure level of specific frequency, (4) rice polishing rate or broken rice by multiple regression equation Perform rate calculation.
[0011]
Examples of the frequency having a high correlation with the rice polishing rate include, for example, frequencies of 500 to 1500 Hz, preferably 700 to 1200 Hz, and more preferably around 900 Hz. Moreover, as a frequency with a high correlation with the broken rice incidence , for example, a frequency of 3000 to 6000 Hz, preferably 3500 to 5500 Hz, more preferably 4000 to 5200 Hz, for example, 4750 Hz is exemplified. Note that these frequencies are shown as an example, and the present invention is not limited to these frequencies, and these may be combined.
[0012]
In the present invention, the rice milling rate can be estimated or measured by analyzing the acoustic data as described above, and the rice milling machine can be controlled based on the result.
[0013]
At this time, the control factors of the rice mill are the rotation speed of the Kongo roll, the load resistance of the discharge port, the opening degree of the discharge port, and the like. These values are generally relatively large up to 70% of the rice polishing ratio, but are relatively small when the target rice polishing ratio is 70% or less, or when the raw rice is easily broken. If it is determined that the broken rice rate is high in the middle polished rice by the measurement of the polished rice sound, if the polishing ratio is very different to the standard course, the rotation speed of Kongo roll when it is determined that abnormal operation of other rice mill, leaving the exhaust It is possible to create a program for automatically changing at least one of the load resistance of the mouth, the opening of the discharge port, etc. and continuing the rice milling.
[0014]
The following examples further illustrate the present invention.
[0015]
[Example 1]
Estimating the Rice Milling Ratio by Power Spectrum Analysis of Rice Milling Sound Using a Test Mill Rice milled rice was brewed using a test mill manufactured by Satake Corporation, and white rice having an apparent rice milling ratio of 30 to 90% was prepared. In the milling by the test mill, brown rice of about 150 g is usually used as a sample, but the weight of the white rice decreases with the progress of the milling and becomes 1 / 2.5 of the initial weight at a polishing rate of 40%. Therefore, in order to eliminate the influence on the sound pressure level due to the decrease in the weight of the white rice, a rice milling ratio is performed in advance, and after preparing a white rice sample having a different rice milling ratio, a certain amount (weight 50 g) is taken out and then milled again with a test mill. A spectrum analysis of the generated sound power was performed. The sound power spectrum at the time of rice milling by the test mill exhibits a flat sound pressure level distribution as a whole, and rapidly decreases in a frequency region exceeding 11,000 Hz. In addition, the sound pressure level generated by the rotation of the motor rapidly decreases in the frequency range exceeding 1,500 Hz, so that it does not greatly affect the extraction of the sound component derived from rice.
[0016]
When high-correlation frequencies were extracted for apparent rice milling ratio including broken rice and true rice milling ratio by only sizing except for broken rice, the sound pressure level at 900 Hz was the most correlated for both (both correlation coefficient = 0.95, number of samples = 18), and the rice polishing rate could be estimated with high accuracy by the regression equation (FIG. 1). From this result, there is a specific frequency in which the sound pressure level changes with the shape change of polished rice due to polished rice among the components of the sound generated during polished rice by the test mill. From the regression equation based on power spectrum analysis, the apparent polished rice ratio and true polished rice It can be concluded that the commission can be estimated. In the examples, for the sake of strictness, white rice of a certain weight was taken out from a sample that had been previously polished and then polished again, but even when normal polishing was performed, a regression equation that estimates the polishing rate with high accuracy. Can be created.
[0017]
[Example 2]
Estimating the rate of crushed rice by power spectrum analysis of polished rice sound The ratio of crushed rice contained in brewing raw rice is a very important factor that determines the brewing suitability of brewing raw rice and is also an index for judging the skill of rice brewing operation. Therefore, it was investigated whether the rate of broken rice could be estimated by sound power measurement. Milled rice with a proportion of 70% of polished rice and crushed rice were mixed in various proportions, and 100 g of white rice was milled with a test mill manufactured by Satake Corporation in the same manner as in Example 1, and the spectrum of the generated acoustic power was analyzed to break the crushed rice. A frequency with a large correlation coefficient with the rate was identified. At this time, when a regression equation was created based on the sound pressure level of 4,750 Hz, which had the largest correlation with the broken rice rate, it was possible to estimate the broken rice rate with high accuracy (FIG. 2, correlation coefficient = 0.95, sample). Number = 84). From this result, there is a specific frequency in which the sound pressure level changes with the content of broken rice among the components of the sound generated during rice milling by the test mill, and the rate of broken rice can be estimated by measuring the sound pressure level during rice polishing. It can be concluded that it is possible.
[0018]
[Example 3]
Estimating the rice milling rate by analyzing the power of the rice milling sound of the practical rice milling machine The progress of the rice milling can be monitored in real time without interrupting the rice milling operation by analyzing the acoustic power spectrum at the time of rice milling with the rice milling machine. In the example, a rice milling weight of 60 kg is made using a small rice mill manufactured by Chiyoda Engineering Co., Ltd., and the power spectrum by fast Fourier transform is recorded in approximately 2% increments until the rice polishing rate reaches 40%. Obtained. In normal rice milling operation, the rotation speed of the Kongo roll is decreased with the progress of the milling, but the rotation speed of the Kongo roll is set constant (1,000 rpm) for the sake of accuracy. As can be seen in Fig. 3 (drawing substitute photo), the acoustic power spectrum during the milling of the practical rice mill shows a flat sound pressure level distribution as in the case of the milling by the test mill, and in a frequency region exceeding 11,000 Hz. Decreases rapidly. According to the correlation analysis between the sound pressure level of the frequency component and the rice milling rate, it was found that the sound pressure level in the low frequency region decreases with decreasing rice polishing rate, and conversely the sound pressure level in the high frequency region increases.
[0019]
Therefore, if several frequencies are extracted by multiple regression analysis and a regression equation is created, the rice yield can be estimated with high accuracy. When a multiple regression equation was created from the sound pressure levels of two wavelengths of 7,350 Hz and 8,650 Hz and the apparent rice polishing rate was calculated, the correlation coefficient with the measured value was as high as 0.96. In the normal rice milling operation, the rotation speed of the Kongo roll is first set to about 1,400 rpm, and the rotation speed of the Kongo roll is decreased to about 900 rpm as the milling progresses. Then, the influence on the sound power spectrum analysis can be ignored (Fig. 4).
[0020]
[Example 4]
Analysis of the structural structure or physicochemical properties of raw rice by using the polished rice power spectrum If the principal component analysis of the polished rice power spectrum is performed, the physicochemical properties of the raw rice by polished rice can be investigated. The principal component analysis of the power spectrum data of Example 3 was performed, two principal components were extracted, and the respective principal component scores for each rice polishing rate were plotted. FIG. In FIG. 5, it is clearly divided into two groups with a rice polishing rate of 72% or more and a rice polishing rate of 70% or less, and it can be seen that the physicochemical properties of rice greatly change around 70% of the rice polishing rate. This is consistent with the fact that the rice milling rate rapidly decreases around 70% of the rice milling ratio, and shows that the physicochemical properties of the raw rice rice change greatly with this value as a boundary. Therefore, the difference in the structure or physicochemical properties of the raw rice during the milling process can be examined by the change in the rice mill sound power spectrum, and this can be used as a new discrimination index for brewing aptitude.
[0021]
【The invention's effect】
The present invention makes it possible to control a rice brewing machine for sake brewing, which has been performed based on the experience and intuition of conventional experts, in a real-time state using a computer. It can be measured. In addition, it is possible to measure the percentage of true milled rice during the milling process, which was difficult to measure.In addition to determining the rate of crushed rice and other brewing aptitudes, it is possible to remotely determine the suitability of rice milling operations and the operating status of the rice milling machine. This has the effect of increasing the suitability of brewing white rice for sake brewing such as sake, and enabling efficient rice milling operations.
[Brief description of the drawings]
[Fig. 1] In the power spectrum obtained by preparing fast polished rice with different milling ratios from the same brown rice for sake brewing using a test mill and performing fast Fourier transform of the acoustic power generated when these are milled again with a test mill, It is the figure which showed that the frequency with a large correlation with a rice milling rate or a rice polishing rate exists. Both correlation coefficients with the sound pressure level of 900 Hz are 0.95.
[Fig. 2] Fast Fourier transform of the acoustic power generated when milled with a test mill using white rice containing crushed rice as a sample, and according to the sound pressure level of 4,750 Hz, the most correlated among the obtained power spectra. It is the figure which showed the type | formula which estimates a broken rice rate. The correlation coefficient between the broken rice rate and the sound pressure level of 4,750 Hz is 0.95.
[Fig. 3] Rice milling with a practical rice milling machine using 60kg of brown rice weight for sake brewing to a rice milling ratio of 40%, and a power spectrum obtained by performing a fast Fourier transform of the generated acoustic power of 94% rice milling ratio It is the figure which showed 40% of data and the data at the time of no load which moved only the motor (drawing substitute photograph).
[Figure 4] Obtained by the regression equation created by extracting the two frequencies by conducting multiple regression analysis of the power spectrum of the milled rice sound recorded in about 2% increments up to 40% of the milling rate in the milling process with a practical rice mill It is the figure which showed the deviation of the calculated value and the measured value of the obtained rice polishing rate. The correlation coefficient between the calculated value and the actually measured value is 0.96.
[Figure 5] Principal component analysis of the power spectrum of polished rice sound recorded in approximately 2% increments up to 40% of the rice milling rate in the rice milling process with a practical rice milling machine, and extracting two principal components It is the figure which plotted the principal component score.

Claims (6)

Spectral analysis of the polishing power at 70 to 40% of the rice brewing rate in the brewing rice is performed, and a frequency highly correlated with the rice polishing rate is identified, and the rice polishing rate is estimated based on this result, or A method of identifying a highly correlated frequency and estimating the rate of broken rice based on the result, including the process of recording acoustic data during rice milling, the process of acquiring the power spectrum of acoustic data by fast Fourier transform, and the sound pressure at a specific frequency A method for estimating a rice polishing rate or a broken rice rate of a brewing rice, characterized by comprising a level extraction step, a rice polishing rate or a broken rice rate calculating step by a multiple regression equation.   Software that samples the sound generated during milling using an ordinary sampling device such as a precision sound level meter or condenser microphone, records it with a low-noise recording device, and performs fast Fourier transform on the resulting sound data The method according to claim 1, wherein the acoustic data is decomposed into a power spectrum by processing using a computer comprising:   3. The method according to claim 1, wherein a frequency having a high correlation with the rice milling rate is in a range of 100 to 12,000 Hz.   The method according to any one of claims 1 to 2, wherein a frequency having a high correlation with the broken rice rate is in a range of 100 to 12,000 Hz.   The method which estimates according to the method of any one of Claims 1-4, and also controls a rice mill based on the result.   6. The method according to claim 5, wherein the control of the rice mill is at least one control selected from the rotational speed of the Kongo roll, the load resistance of the discharge port, and the opening degree of the discharge port.

Images