JPS6129446B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to a method of predicting the thermal equilibrium temperature between a temperature sensing element and a measured object before thermal equilibrium and measuring the temperature of the measured object. Electronic thermometers that use temperature-sensitive elements such as thermistors are used in thermometers and the like. However, since a temperature sensing element generally has a thermal time constant, it is impossible to instantaneously measure the temperature of the object to be measured. Therefore,
Generally, after bringing the temperature sensing element into contact with the object to be measured,
The conventional method was to wait for both to reach thermal equilibrium and then measure the temperature, but it usually takes about 50 minutes to reach thermal equilibrium.
Since it takes about a minute, it has the disadvantage that it cannot be measured in a short period of time. Therefore, a method of measuring the temperature of the object by predicting the thermal equilibrium temperature between the temperature sensing element and the object to be measured before the thermal equilibrium is considered has been proposed. This can be done by adding a fixed value corresponding to the temperature difference between this temperature and the thermal equilibrium temperature to the temperature detected by the thermosensor at a specific time before thermal equilibrium, or by using data at multiple times before thermal equilibrium of the temperature detected by the thermosensor. There is a method of predicting the thermal equilibrium temperature by calculation. However, the former method is easily influenced by the initial temperature of the temperature sensing element, and moreover, the temperature difference varies depending on the individual, so accurate measurement cannot be expected at all. On the other hand, the latter method only requires that the thermal time constant of the thermosensing element be constant; however, in reality, this thermal time constant varies depending on the thermal time constant of the object to be measured, so it depends on the individual and on the measurement site. Therefore, there was a drawback that measurement errors occurred. This invention was made in view of the above points, and its purpose is to predict the thermal equilibrium temperature before thermal equilibrium, and to quickly and accurately measure body temperature, etc., without causing errors due to differences in the object to be measured or changes in the measurement site. It is an object of the present invention to provide a temperature measurement method that can measure temperatures and reduce the memory capacity for storing classification data of temperature rise curves. This invention will be explained in detail below. When a temperature sensing element such as a thermistor is brought into contact with an object to be measured, for example a human body, the detected temperature T of the temperature sensing element increases exponentially with time t, as shown in the figure.
Eventually the thermal equilibrium temperature Tu is reached. This temperature rise curve changes depending on the object to be measured, for example, from person to person, and even in the same object, it changes depending on the measurement site to which the temperature sensing element is brought into contact. Therefore, in the above-mentioned conventional adding method or method for determining the thermal equilibrium temperature by calculation, measurement errors occur due to differences in individuals or measurement sites. To solve this problem, store a large number of temperature rise curves of the temperature detected by the thermosensor in advance, calculate the temperature rise curve for each measurement, and find out which of the stored curves it is most similar to. , based on which the thermal equilibrium temperature may be predicted before thermal equilibrium. However, many examples of past temperature curves (e.g.
Memorizing 100 or several hundred cases would require an enormous amount of memory, making it completely impractical. Therefore, when storing temperature increase curves, a method can be considered in which the curves are classified and stored according to their characteristics, as is done in pattern recognition devices. In that case, unlike characters or figures, the problem is what features to extract from the temperature rise curve. In order to classify the temperature rise curve of the temperature detected by a thermosensor according to its characteristics, this invention divides the portion of the curve before reaching thermal equilibrium into multiple time segments, and calculates the amount of temperature change in each of the divided time segments. The combination is adopted as a feature. That is, the characteristics of various temperature increase curves are extracted by combining the amount of temperature change in each of the plurality of time periods, classified according to characteristics, and the classified data is stored. In this way, the capacity of the memory is much smaller than that required for storing the temperature rise curve itself. However, in reality, the thermal equilibrium temperature cannot be predicted even if only the characteristics of the temperature rise curve are known, so the temperature difference between the temperature detected by the thermosensor and the thermal equilibrium temperature at the end of the final time segment when the characteristics of the temperature rise curve are extracted is calculated. and store it in association with the classification data. Then, during actual measurement, the characteristics of the temperature rise curve of the temperature detected by the thermosensor are extracted in the same way, and it is determined which of the stored classification data of the temperature rise curve corresponds, and the characteristics are matched with that classification data. The thermal equilibrium temperature is predicted and the temperature of the object to be measured is measured by adding the temperature difference value to the temperature detected by the temperature sensing element at the end time of the final time segment when extracting the feature of the temperature increase curve. Note that when dividing the temperature rise curve into multiple sections, the time length of each time segment may be constant; however, since the slope of the temperature rise curve gradually becomes gentler, it is better to increase the above time length as time passes to increase the measurement time. It is desirable to shorten the curve and more accurately capture the characteristics of the curve. Furthermore, the classification data obtained for each characteristic of the temperature rise curve may be data showing the combination of each temperature change amount itself, but it may be data obtained by comparing and classifying the temperature change amount using a single or multiple threshold values. But that's fine. In this way, the memory capacity can be even smaller. Next, the present invention will be specifically explained using examples. As shown in the figure, the time after the temperature T detected by the temperature sensing element exceeds a certain reference value Ts is defined as _t0 , and the value of T at this time is defined as _T0 . And time △ from t ₀
The T value at time t ₁ after t ₁₁ (for example, 1 second) is T ₁
Similarly, after △t ₁₂ (for example, 2 seconds) from t ₁ ,
Let the value of T at t ₂ be T ₂ , let the value of T at t 3 after △t ₁₃ (for example, 3 seconds) after t ₂ be T ₃ , and let _△ from t ₃
The value of T at t 4 after t ₁₄ (for example, ₄ seconds) is T ₄ , and the value of T at t 5 after Δt ₁₅ (for example, 5 seconds) from t ₄ is T ₄ .
Let the value of be T ₅ . From this, each time segment △t ₁₁ ~ △
The amount of change in T at t ₁₅ from △t ₁₁ to △t ₁₅ is △T ₁₁ =T ₁ −T ₀ △T ₁₂ =T ₂ −T ₁ △T ₁₃ =T ₃ −T ₂ △T ₁₄ =T ₄ −T ₃ △T ₁₅ = T ₅ − _{T 4} . For example, θ ₁ , θ ₂ , θ
Classification is performed using _three threshold values (where θ ₁ >θ ₂ >θ ₃ ). Here, θ ₁ ≦A, θ ₂ ≦B
<θ ₁ , θ ₃ ≦C<θ ₂ , D<θ ₃ , that is, D
If <θ ₃ ≦C≦θ ₂ ≦B≦A, △T ₁₁ ~△
The combinations of T ₁₅ can be classified as combinations of A to D. For example, △T ₁₁ to △T ₁₅ are θ ₁ < △T ₁₁ , θ ₁ < △
T ₁₂ , θ ₂ <△T ₁₃ , θ ₂ <△T ₁₄ <θ ₁ , θ ₂ <
If ΔT ₁₅ <θ ₁ , the classification result is “AAABB”. A similar classification is performed for various temperature rise curves, and the classification data is individually stored in memory. On the other hand, the value of T at the end time t ₅ of the final time segment Δt ₁₅ when extracting the features of each temperature increase curve is T ₅
The temperature difference between the temperature and the thermal equilibrium temperature Tu is Tu−T ₅ =T _RX .
It is stored in association with the classification data of the characteristics of the temperature rise curve. Note that the measurement of Tu for determining T _RX may be performed over a long period of time (for example, 5 minutes) in the same manner as in a normal measurement method, and the highest value of T at that time may be taken as Tu. When classifying as described above, the number of possible classifications is 64.
becomes. Actually, the value of θ ₁ to θ ₃ is θ ₁ = 0.18℃
When θ ₂ =0.12°C and θ ₃ =0.06°C, the above processing was performed on the temperature increase curves of 213 examples, and the following results were obtained.

[Table] However, this table only shows items that appear relatively frequently. The deviation (standard deviation) of T _RX from the average value for each category was small, and it was confirmed that there was no problem even if T _RX was fixed for each category. Next, during measurement, △ ₁₁ to △ ₁₅ are similarly obtained to extract and classify features. Then, it is determined which of the stored classification data the extracted feature corresponds to, and the value of T _RX corresponding to that classification data is determined.
Tu can be found by adding to T ₅ . This Tu becomes the temperature of the object to be measured. Note that if a feature is extracted that does not correspond to any of the classification data stored in advance during measurement,
Either add the total average value of T _RX corresponding to each classification data to T ₅ , or indicate this with some kind of display, and actually measure Tu over a sufficiently long period of time, similar to measurement using a mercury thermometer, etc. do it. On the other hand, if a contact failure of the temperature sensing element with the object to be measured occurs, any one of ΔT ₁₁ to ΔT ₁₅ may become a negative value. In such a case, you can correct the contact condition and start the measurement again from the beginning, but if a contact failure inevitably occurs instantaneously, hold the temperature detected by the thermosensor at its peak. , assuming that the detected temperature when a contact failure occurs is the same as the previous peak value, △T ₁₁
~ΔT ₁₅ may be determined and features may be extracted. When implementing the method of the present invention, the characteristics of the temperature rise curve of the temperature detected by the thermosensor are extracted from T ₀ to _{T 5} after the value of T exceeds a certain reference value Ts, as shown in the figure. For accurate measurements, it is desirable to use the value of The first reason is T ₀
This is because if the value of ~ _T5 is too low, an error due to room temperature will occur, and from this point of view, Ts is preferably equal to or higher than room temperature. The second reason is that in a region where the values of T ₀ to _{T 5} are low, there is a possibility that the contact state of the temperature sensing element is poor. For these reasons, the appropriate value for Ts is, for example, 34 to 35 degrees Celsius when measuring body temperature.In this way, it will not be affected by room temperature under normal room temperature conditions, and the value of T will not exceed this value. If the temperature exceeds a certain temperature, the contact condition of the temperature sensing element is sufficiently good, so that features can be extracted accurately and temperature can be measured accurately. Further, in the above embodiment, ΔT ₁₁ to ΔT ₁₅ were compared and determined using a plurality of threshold values θ ₁ to _{θ 3} , but a single threshold value may be used for comparison and determination. In that case, it is desirable to increase the number of temperatures detected by the temperature sensing elements (T ₁ , T ₂ , . . . ) used for extracting features of the temperature rise curve. As described above, the present invention extracts and classifies the characteristics of the temperature rise curve of the temperature detected by the thermosensor for various objects to be measured and graduation parts, and then calculates the thermal equilibrium temperature based on the characteristics before thermal equilibrium. Since this is a method of measuring the temperature of the object to be measured, there is essentially little measurement error due to differences in the object to be measured or changes in the measurement site, and temperature can be measured accurately and quickly. In addition, when extracting and classifying the characteristics of the temperature rise curve, the amount of temperature change in multiple time periods is compared and determined using a single or multiple threshold values, which significantly reduces the memory capacity for storing the above classification data. be able to. Although the method of this invention is particularly suitable for measuring body temperature, it is generally applicable to measuring the temperature of a measured object by predicting the thermal equilibrium temperature before thermal equilibrium using a temperature sensing element such as a thermistor with a thermal time constant. is possible.

[Brief explanation of the drawing]

The figure is a diagram for explaining one embodiment of the present invention.

Claims (1)

[Claims] 1. In a method for measuring the temperature of an object to be measured using a temperature sensing element having a thermal time constant before the temperature sensing element and the object to be measured reach thermal parallelism, the temperature sensing element is attached to the object to be measured. Characteristics of various temperature rise curves of the temperature detected by the thermosensor when in contact are extracted by combining the results of comparing and determining the amount of temperature change in multiple time periods before reaching thermal equilibrium using a single or multiple threshold values. Then, the classification data is classified by characteristic and stored, and the temperature difference between the temperature detected by the temperature sensing element and the thermal equilibrium temperature between the temperature sensing element and the measuring object at the end time of the final time segment at the time of feature extraction is calculated. Find the value and store it in association with the above classification data,
At the time of measurement, a feature is similarly extracted from the temperature rise curve of the temperature detected by the thermosensor, it is determined which of the classification data the feature corresponds to, and the temperature difference value corresponding to the classification data is determined at the time of feature extraction. A temperature measuring method characterized in that the temperature of the object to be measured is measured by predicting the thermal equilibrium temperature in addition to the temperature detected by the temperature sensing element at the end time of the final time segment. 2. The temperature measuring method according to claim 1, wherein the time length of each of the plurality of time segments is increased sequentially as time passes. 3. The temperature measuring method according to claim 1, wherein the feature of the temperature increase curve is extracted after the temperature detected by the temperature sensing element exceeds a certain reference value.