Analog Computers

Reference / Paper · 1969

Introduction to Simulator (Part 2)

Read the PDF (11 pp) ↗

The second installment of Hitachi's Technical Information Series No. 10 surveys four analog-computer simulation applications: pipe network flow analysis (using electrical analogs of the Hazen-Williams and Williams-Hazen equations), voice synthesizers (spectrum-analysis and configuration-analog methods), nerve-cell and nervous-system models, and cardiovascular/respiratory system simulators. Each topic is supported by circuit diagrams and block diagrams illustrating how the Hitachi analog-hybrid computer is programmed to represent the corresponding physical or biological phenomena. The document serves as an introductory survey of biomedical and engineering simulation capabilities available on Hitachi analog-hybrid equipment.

Manufacturer
Hitachi
Year
1969
Type
Reference / Paper
Language
English
Learning track
specific applications
Pages
11
Credit
Hitachi, Ltd. Technical Information Series No. 10, 1969. Printed in Japan.
  • Hitachi
  • analog simulation
  • pipe network
  • biomedical simulation
  • voice synthesis

← Back to the Reference Library

Introduction to Simulator (Part 2)

HITACHI Analog-Hybrid Computer Technical Information Series No.10 INTRODUCTION TO SIMULATOR ‘Part 2) 1969 Hitachi, Ltd. rinted in Japan Pipe Network Simulator In order to study new planning, improvement and supplying method of water supply pipe and gas pipe, it is desirable that the relation among the pressure within the pipe, the amount of flow, etc. can be readily analysed. For such a purpose, the pipe network simulator is made. In the calculation of pressure within the pipe network and the flow, the numerical integration method based on repetitive calculations were made. This calculation which was done by the digital computer can also be called simulation, but in this article, an explana- tion on exclusive simulation in which direct simulation is employed will be given. As for the direct simulation simulator, an electrical circuit identical to the pipe network is prepared and the current is made to correspond to the flow and the voltage is made to cor- respond to the pressure difference. There are the following advantages in comparison with the numerical integration. (1) The calculation speed is fast. Calculations which obtain the stationary condition of the current and pressure can be obtained at once. The response of each portion in relation with the changes in load and supply conditions with the elapse of time can be found easily. (2) The simulated circuit network is connected in the same layout (phasewise) with the actual pipe network, so the setting of problems can be done intuitively. (3) Since the solutions appear in parallel on the meters, the overall flow and pressure dis- tribution can be seen at a glance. In case pilot lamps are used as the pipe elements, in place where the loss is great, the lamps will shine brightly. Therefore, such places can be identified at once. (4) It is economical (Of course this depends a great deal on the precision and the scale of the machine). From the various points mentioned above, such simulators are more suitable if the person who uses the simulators performs various trials while observing the results, then conducts the optimum designing, or makes the optimum supplying program. On the other hand, (1) Ifthe system which you wish to simulate becomes large, the simulator will also become large, proportionally. (2) It lacks flexibility, Needless to say, it cannot be used for purposes other than calculations for pipe networks. Furthermore, even in case it is applied for the calculation of pipe networks, if the accuracy or scale exceeds the limits even slightly, the problem can no songer be analvsed. In accition to tne above, even if the assumption is changed slightly, it will become very aiificult to anaivse, The above points are given as defects. In case of the digital computer, the advantages and disadvantages are reversed. Besides the abovementioned types, hybrid systems are also considered. 3-1. Principle of the Pipe Network The relation between the amount of flow Q and the pressure difference H of the flow pass- ing through the pipe can be expressed as follows: A oO (27) ‘Sometimes for H the difference between the squares of the pressures is taken) Although n varies according to the equations in most cases, 1.85 or 2 are used. k, is a constant which is deter- mined by experiments. For instance, in case n= 1.85 is used for the water pipes, William- Hagen's equation shown below is used. ky =10.666C7185 E487 g ee teen eee (28) where Flow Coefficient Diameter of Pipe Length of Pipe SOQ On the other hand, as it is well known, the following re.aticn exists between the current and the voltage in the electrical circuit network. Ez Ro ee ee ene (29) oon If the resistance changes with the amount of current, tne voltages at bpotn ends of the resist- fle ance will be non-linear function of the electrical current. [fa #2113)... "on-linear resistance ne is used, the following relation can be satisfied a. |< (30) in and this will be similar to the relation which exist difference (or the difference in the squares of the pre Biv Tum? of tiow, Furthermore, no matter which intersection of the ftanen, there is no acculula- m tion of the water or gas that the flowed in. Trereicr, Ra get rs amowill be zero. In electrical circuit networks, the same equation anulie itor tir oon fis Law. The difference that lies between the electrical circuits ang tie tite 2 wits is 11.41 ine Tormer requires a return route. However, as shownin Fiz. 2¢, _ Pipe resistance if the outside pressure (or standard dressur-! of the portion used is made to corresnons wit: the ground potential, and if the pressure -: [ oo— the fluid is made to correspond with the el tr - | motive force, one to one correspoingence can >: ‘ poN made completely with electrical circuits. \ IN N -.2.. Simulation of Pipe Network 3.2. Operational Element of Pipe Network 13 In the abovementioned way, in order to mainta~ ue * “+ 2s rrespondence, the matter which will become a problem from the practica. Sin intis ssatcts use for the abovementioned non-linear resistance. As shown below, various ~materia-.+ are uses. fl: (1) The Method Using Lamps According to Mclloy's method which showed the first oractical success for pipe network simulation, uses tungsten filament lamps for the Dive : s. In this method, the change in resistance owing to temperature rise is uti: cst of each element is low, and there is further advantage that a rough idea of the press an be obtained by observing the brightness intensity of the lamps, as mentioned above. : ‘sacvantage of this method is that the non-linearity is based on the physical property of the -amno. Thus, adjustments can not be made and the accuracy will be limited. Furthermore, is necessary to prepare beforehand, various kinds of lamps which will correspond to all sorts of pine network elements, (2) The Method in which Two Terminal Function Generators are Used Diode Function Generators in analog computers either combine the broken line character- istics of the diode, and give an arbitrary non-linear characteristic between the input and output, or based on the same concept, resistance of two terminal circuit can be made to be set as an arbitrary function of the voltage at both ends. mae Fic, 21 shows this principle. =e diode is biased in the normal direction by the battery, the state of continuity is maintained, and the circulating current flows in the direction of the diagram. Ifthe E for the battery of the reverse direction is determined adequately, there will be no voltage difference between terminal 1 and terminal 2, and they will act merely as a resistance in case of outside circuits. In case the voltage which is added to terminals 1 and 2 from the outside becomes large, each time it exceeds the battery voltage, the diode connected to that battery will enter a non-continuity state and the total resistance will become high. In actual practice, alternating currents are recti- fied and used in place of the battery. Thus, itis necessary to maintain high precision by employing a power source stabilized by feed back, Further- more, since this is a power sources which are insulated for direct current shall be used, Fig. 22 shows an example of two terminal function Fig. 21. Example of Two Terminal Function Generator generator using transistors. At this time, the saturation characteristics shown by the transistors as illustrated in Fig. 23 are utilized to maintain the relation shown in Equation (30). 6 ’ + _ Fig, 22. Two Terminal Function Generator Employing Transistor Circuit (3) In Case Function Generators of 4 Terminals are Used Current Voltage Fig. 23. Characteristics of Fig. 22 Fig. 24 shows a pipe network simulator used for alternating current. The current i which flows through the pipe network is detected by current transformer, rectified, and made into a voltage which is proportional to the current. A non-linear characteristic is formed from this by the function generator SC which can be used for normal analog computers. Subs equent- ly, it is modulated, shaped, then fed back to the original branch by transformer T. A A lI >Re p<] | mop : | El< y ex; Ay 2 wo Fy Fig. 24, Block Diagram of Pipe Element Using 4 Terminal Function Generator -3- Fig. 25 shows the circuit of such function generators. All diodes used are zener diodes. In this circuit, the voltage of D1 to Dn is too high, and in order to lower this Eb and Dp are connected in reverse. If suitable Zener volt- ages are obtained, this may be omitted. As for the method in which function generators of (2) and (3) are used, even if battery voltage or Zener voltage is constant, by making the re- sistance variable, the current and the voltage characteristics can be adjusted arbitrarily. Furthermore, by increasing the number of parallel circuits, the required accuracy can be obtained. However, there is the disadvan- tage of the circuits getting too large. Besides the pipe network elements, load, pump and tanks can also be connected like the actual Fig, 25. Circuit Drawing of the SC system. The load can be simulated by the Portion (Function Generator) resistance or constant current device. Snownin Fig, 24, Simulation made by resistances is directly analogues to the opening of the faucet being fixed, and the outflow will change in accordance with the changes in pressure at both ends. Simulation made by the constant current device is directly analogues to the user adjusting the opening of a faucet in accordance with the amount of flow. From the abovementioned standpoint, the latter is used a great deal. In addition to the above, by inserting function generators having the same pressure flow characteristics as the pump, tank, and valves at their respective positions, the pipe network containing such items can be analysed, The abovementioned simulators are normally used for simulation of constant state, and mathematically speaking, it is similar to obtaining the solution of a non-linear algebraic equa- tion, Even in case of what is called dynamic simulation, if the conditions change together with time, in other words, in case the coefficient and constant terms of the algebraic equation changes with time, in most cases calculations to see how the solutions follow, are made. Problems of transient conditions in their true sense can be analysed by combining capacitors with the abovementioned static operating elements. In such a case, good frequency character- istics are required for the pipe network elements, and the No.2 method is the preferable method. As for the simulators for this type that are manufactured in Japan, those having variable func- tional forms have several tens of pipe elements. The functional forms can be changed arbit- rarily according to the problem. This simulators using lamps have several thousands of lamps prepared, and suitable types are selected according to the problem. 4. Voice Simulator Needless to say, the most natural and convenient method of exchanging information among human beings is to emit voice and talk, and to listen to what is spoken. As a means to com- municate between distant people, the telephone is much more conveneient than a letter or telegrams. Not only is it faster but the content is much more abundant, The intention and information can be transferred much easier. Even between the computer and the human being if information can be exchanged directly by voice, the work will be much faster and the computer will be easier to use. The utilization value will not doubt rise con- siderably. Not only computers but also the other machines and equipment will increase efficiency con- siderably if they can understand what the human being say, and also respond to it. The con- venience in this case is beyond our imagination. Of course, there have been some machines which can talk. If the same sentence is to be repeated, the tape recorder can record and reproduce the voice. If you wish to change the sentence from time to time, you can break up the sentence into elements and combine them adequately. In case we are informed of wrong telephone numbers or given time service over the phone, it does not mean that a nice young lady is talking at the other end of the phone, "The time now is....,.. " ‘That number has changed .......... Please re-dial.'' The dotted line portion is changed, and it is a recorded tape that is giving the information. As for more complex usage, tne sentence is broken up into words, then the words are reassemb-ec again to form sentences and reproduced. In case of the stock excne in New York, if the security dealer who stocx value of a certain machine com- oa 2 taiking computer by dialling culine voice will answer is outside wishes to know pany, he can get the intor: a specified number. For instance a Vocal tract back "........ Machineries, Onen! Quotation 83, Highest Price 85, Lowest Price 82, Turnover 7000. In order to really utilize voice as a means for conveying intormation instead of the system Vocal chords .r, the voice emit- “, and complicated Trachea which is like cutting takes anc putting * ting mechanism should be controlled by voices should be synthesized. Fig. 26 In order to accomplish this, the mecranismr trem which voice is , emitted, and the elements whicn compos xe checked, Asa synthesizing method basec on ia: nvbric computer system in which the analogous avnin nvioe is contrecled by the computer output is considered to be the common sense concent, As for the analogous synthesizing device, there is the sp. otrum analog system (Up to quite recently, this was called the terminal analog svst vith. VOCODER system is included in this) which simulates the sound as a pnenomreron a? cantiguration analog system (This used to be called the vocal tract analog) which simulates +h: ization movement of the mouth by tracing back to the physiology and physics of v The spectrum characteristics of the voice wave nave been if necessary, it can be measured directly and confirmed. T possess is abundant, and a lot of quantitative analysis cata » hardware is also comparatively simple in case of tne spectrum an it is more economical. On the other hand, the research work on vocalizing organs anc vr chiefly based on the personal views and personal observations o: result, most descriptions are qualitative descriptions, and with sighted research work, objective and quantitative observations b2as movies have just begun. At present, we still lack knowledge anda more, as for the configuration analog method, it is very complicaty¢ as mentioned later, and there are still many technological difficulties. Therefore, the orice is very high. However, if we take up the problem of making laws for the synthesis, it will be more sub- stantial (i.e. The description of law will be more simple, and there will be less exceptions, ) to go back to the cause or vocal movement rather than to think in the dimension of spectrum characteristics or phenomenon, In addition to the above, judging from the ability of synthesiz- ing continuous vocal sounds, it may be said that the configuration analog system which has limitations necessarily imposed on vocal waves owing to the utterance of the human beings set into the operation of the synthesizing device, has a higher potential capacity. Voice emitting mechanism, Stem, <8 great Gial in the past and ore, the knowledge that we aren accumulated, The < svstem. Consequently, ~oca. specialists. Asa exception of a few fore- >on X ray or X ray itative data. Further- (a) Spectrum Analog System The research on synthesis based on the laws of spectrum analog system is being done at various places in Japan such as Tohoku University, NHK General Technological Labs, Oki Electric Co., and Tokyo University. Among such research works, the work of Ilolmes of Great Britain on the Joint Speech Research Unit has shown quite a success by concluding it in the form of a law, As a synthesizing equipment, parallel spectrum analog as that shown in Fig. 27 was used, and control was performed on the following nine parameters, FE : Basic frequency of voiced sound source. Ss : Selection of whether the sound is voiced or unvoiced. BR, EB, Fi : Formant frequency from No.1 to No. 3 Ai, Az, As : Intensity of each formant component. A: Intensity of high frequency energy. -5- As an example of synthesis, "A bird in the hand is worth two in the bush" was shown in the following three steps. (i) The values of all control parameters were extracted as a result of analys- ing the live human voice, and control and synthesis was done accordingly. Amplitude Frequency | 5 c/s >ontrol control (ii) The continuing time of each syllable i ! 100 ~ :. a ¥ 1000 which greatly influences the natural- P, eel : a, oH 700 ~ F, 2400 ness, and the pitch frequency F. of the vowel portion have been controlled by the results obtained from the analysis + of the live human voice. Only the con- trol of the formant characteristics was A; I \ done in accordance with the calculations | a4 thon oOut based on the law. A P ; + Fy in accordance with the law. Although _ it is quite natural that the quaity of the - | — BPE synthesized voice will drop in the order Nowe o + 0 of (i) (ii) and (iii), the meaning of (iii) was quite understandable. However, it was recognized that the naturalness was Fig. 27. Composition Drawing of lacking. Synthesizing Unit for JSRU (iii) The control of all parameters were done Pulse L of oo a4 3500 | | | F (b) Configuration Analog System The research work based on the configuration analog system is also being done in Japan at the Electro Technical Lab., Meiji University, ete. Tne recent results obtained at I are quite remarkable. ce block diagram is shown in The external view of the equipment is shownin Fig, 2:5, a Fig. 29. If we give a brief explanation of the principle, + The channel from the glottis to the lins is consiceres a length of about 17 cm. This is simulated by seventeen areas anda length of 1 cm each. According to the soi sound for each round tube can be expressed by the fo-lowin posed by an analog computing circuit shown in Fig. 30. Read ‘distribute MT . ; Fig. 29 Configuration Analo Fig. 28 Voice Synthesizing Unit ° System ° (Configuration Analog System) an na To Um-il-lm o 7 Am Pm 2 m m pe* mt m-1 where Volume speed of air inside of the cvlinder Sound pressure Length of the cylinder (cm) Density of air (g/cm) Speed of sound (cm/sec) Cross Section of Cylinder (cm?) povUrycd The cross section A will have to be controlled in accordance with place and time, (Actually, this is controlled by the area ratio.) The multiplier will perform D-A conversion at the same time it conducts multiplication. However, in order to prevent the flick caused by the digital control Fig. 30. Basic Block Diagram signals, the multiplier is designed for a smoother characteristic against control signals. 5. Simulator in Relation with Medicine It has been known from olden times that living bodies are gatherings of complex control systems which is composed of multi-feed back loop. However, the analvsis of the system was quite qualitative. Together with the development of the centro. theory, quantitative analysis was begun from very simple feedback loops. Since the control loop of the living bodies contain a number of ~ sinear elements, simula- tion is indispensable for their analysis. The development o: simulators is being advanced for nerve cells and certain functions of the organs. At first, the purpose of developing such simulators was for cbtaining an auxiliary means for studying physiology, but now some are being studied for the ceve_-opment of control equipments for living bodies, and for engineering means. As simulators [cr living bodies, there are vari- ous kinds such as nerve system, respiratory circulating svstem, metabolism are exactly the same as those used for chemistry, they will be omitted mere, anc explanations shall be given on nerve system and respiratory circulating system. d-1. Simulator of the Nerve System The nerve cell which composes the nerve system is composed of tne cell body, dendrite, and axon. The axon is also called the nerve fiber, and ths end of it i= branched. The branched portion is called synopsis and it connects with another cel body or dendrite of other nerve cells. The information transmission of such nerve cells can be detected Dendrite from the outside as electrical impulses. Thus signal input from other nerve cells to the synopsis, and its relation witn the ceil output, the signal transmission characteristics inside of nerve fibers, etc., have been made clear, and simulators have been proposed. In the following paragraphs examples of simulators are shown by discriminating the former as nerve cells and the latter as nerve fibers. 3-1-1. Cimulators for Nerve Cells The nerve cell receives the output of other cells as input, via the Synopsis. There are two types of inputs. One which will excite the cell and cause output (i.e. exciting input), and another .niehn is reverse and which has suppressive effects, Over ten uitferent properties between the output and input have been made Fig.31 Nerve Cell -7- clear, Since it is very difficult to develop a simulator which can satisfy all of the conditions simultaneously, simulators possessing the major properties have been developed. For instance, Harmon of the Bell Telephone Labs. assumed the following properties, and developed a simulator illustrated in Fig, 52. (i) In case a stimulus exceeding a certain limit is given within a limitec time, the cell will be excited anc sen¢ our pulsive output. No matter now long stimulus below a certain .evel is the cell will not become excite: limit value will change asa the elapsed excitation o: the cell. (ii) In case the cells become excited, no matter how large the stimulus is, tne cell will not become excited Tor about 2-3 m sec as if the threshold value has become infinite (This period if called the period of non-response. ) (iii) In case two or more stimuli less than the threshold value is given to one cell, those stimuli will be added, and if the (b) Model Circuit of Nerve Cell sum of the stimuli exceed the limit value, the cell will be excited, Fig. 32 Simulator for Nerve Cell. (iv) In case the suppressive synopsis is stimulated, even if the exciting synopsis is stimul- ated at the same time, the cell will not be excited. As shown in Fig. (a) of the block diagram, in order to obtain properties (i) and (iii) an incomplete integrator is used, and feedback is employed for obtaining (ii). Furthermore, suppressive input is given to obtain property (iv). 5-1-2. Nerve Fiber Simulator The nerve fibers which transmit the output of ceils have attracted the attention of people engaged in engineering since they possess an idealistic electrical pulse transmission system for forming waves. There is an excellent mathematical model of Hodgikin-Huxley for this transmission character- istic, and Nagumo et. al. have proposed a simulator using tunnel diode as shown in Fig. 33 by studying the above mentioned mathematical model. As shown in Fig. 34, this transmission system has an excellent wave forming property in which the amplitude and pulse width can be made identical. Besides the above, a large number of simulators have been proposed, but they will be omitted here. o- t t $ ° 5-2, Simulator of the Circulating System oq Kw © The circulating system which is composed 7 ) +} of the heart to send out blood, large and small arteries, capillary tubes and veins with the circulation of blood, metabolism of components Fig. 33 Simulator of Nerve Fiber within the blood, and their control system. For instance, as a Simulator for circulating dynamics, | f \ Mcleod developed one which is shown in Fig. 36. For instance the heart is composed of left and /~— fo——~ right arteria and ventricles of the heart, and the —~ blood circulates in the order of right atrium, —— 0 right ventricle, lung, left atrium, arteries, viens, o™ @ 4 we and right atrium. Thus, it can be expressed by the ~ I™. block diagram shown in Fig. 35(a). The valves ~ shown in the diagram are valves to prevent backflow, Fig. 34 Output of Simulator -8- &* .- omy weg wm BY, the and they are quite indispensable. However, on account of this, blood flow resistance will form. The thickness of the heart muscles for the left ventricle and the right ventricle are thicker than those of the arteria, and the compliance increase of decreases in proportion to Thus, an electrical equivalent circuit shown in Fig. (b) can be made. the volume. right atrium left atrium Pr Ppa Uright un, left ventricle c 8 P erericle C 7 1/c Trigeminal lung arteries Cowled : tv Er Ver pe ‘pa Valve Valve Valve Arteries (wv) (mv) Valve L/Rto 1/Rpa (PV) arteries (AV) Vr Vpa and Viens rr “Close for atrial defect ! pecs 08)— = a 1 Pa | Riu Row Rps Raw Rav | | i} / Cr ie T Cps Ct ~ 1/Cs (26) 1/ Sun 1/Ca Vel ; | V a? Rme AW LS Or: , d Py Cc + Ge rtih lo Fig. 35 Simulation of the Circulating System Fig. 36 Simulation Circuit Diagram of the Circulating System Fig. 36 shows the abovementioned equivalent circuit simulated on an analog computer. The responses of various circulating system diseases are being studied by employing this simulator. Therefore, various functons corresponding to the various diseases are added as shown by the dotted lines in Fig. 36. Although there are various proposals for the control system of the circulating system also, in this article, the writer would like to show a nerve control simulator versus heart beat which was reported by Gardner. In other words, heart beats originate at a portion called the trunk nodule located on the right atrium, and the rate of stimulus is controlled by the sympathetic nerve and the vagus which gives the input. In order to find this transmission characteristic, Gardner stimulated each nerve of a dog and investigated the response. As a result he found that the heart beat can be given by the following equations. HR, + GCs )f1 HR= ——————— G(s) = CAe"TS(1+t3s)*#I/C 1+ tys)C1t+ts5s) H(s) = BC1lt+t ys )*#*/CI+T2s8) where HR =: Heart beat at the time when there is no stimulus. f,, f2: Stimulus frequency of sympathetic nerve system and the vagus system. A, B: Non linear characteristic shown in Fig. 37. x This term will become effective only when 60 In case the abovementioned transmission characteristics are the stimulus frequency of the nerve system has increased. A (S beats/min ) simulated, relations shown in Fig. 38 can be obtained. Although the response of this simulator is not complete, it is very interesting because it suggests the research trend in the future, and by the development of such independent Simulators, the analysis of even more complex phenomenon can be made possible. -9- ovat f, =15% ou f=) “A”Right sympathetic i=] “B” Right vagus}* — ww So Ss 10 0 5 10 15 20 20 frorfs(%) Fig. 37 Static Characteristics of Heart Beat Control System 5.3, Simulator of the Respiratory System The simulation of the respiratory system also consists of the intake of oxygen into the blood, and discharging of carbon dioxide (i.e. gas exchange function), breathing dynamics of the lung, and the control system for the above. As an example, an ex- planation will be made on the respi- ratory dynamics system proposed by “-B Yoshimoto, et. al. Function generator In other words, the exchange of gas is done between the lungs and the fi outside air via the bronchial tube S P agus and branch. Consequently, the [ie | Sympa thetic ( ay “ A HR, +G(S)fy pressure of the lungs will balance vende Lead t Lead ts R= with the inside pressure of the chest fh yon . “A” n AA which surrounds the lungs and the A Pt 1 : : Delay Function , : compression force of the lungs owing tape [] generator to its elasticity. Furthermore, the amount of gas that will be exchanged will be determined by the inside pressure of the lungs and the imped- Fig. 38 Simulator of Heat Beat Control ance of the bronchial tube and branch. On the other hand, since it is clear that the compression force of the lungs based on their elasticity and the impedance of the bronchial tubes are non-linear. The simulator shown in Fig, 39 was developed under the assumption that the admittance will be proportional to the difference between the bronchial tube branch and the inside pressure of the chest. As shown in the drawing, this simulator consists of the left lung, right lung and the bronchial tube, In accordance with the change in chest pressure caused by the stimulus from the respiratory control center, the change of inside pressure in the lungs will occur. As a result, the gas L exchange flow amount can be d r [; sacha) tbe obtained as an output of the ad- \ Pressure (\——~. mittance determining portion of O, | <br ube Inside pressure of the chest left lung Inside pressure of the lung exchahge gas left lung <[} All exchange gas Insige pressure the bronchial branch which is othe bret determined by the Summer A and h—_—*] the Multiplier M. The integrated value of this exchange gas flow will change the lung volume and return the inside pressure of the lungs to the original level. By improving this simulator slightly, a simulator which com- bines the breathing number and control system are also developed. Fig. 39 Simulators for the Respiratory System Simulators for gas exchanging functions are also developed by Grodin but this matter will be omitted here. NP yy <J]] Right lung Right lung Inside pressure of the lung exchange gas Inside pressure of the chest. 9.4. Other Simulators Besides the above mentioned simulators, there are various other simulators such as body temperature controlling system, osmotic pressure adjustment svstem for the kidneys, and the blood sugar controlling system. Owing to the limit in space, such matters will be omitted. Since the functions to be simulated in the living bodies are non-linear, the simulator is becoming an even more important tool for the analysis of such phenomena. A great development can be anticipated for the future of simulators. -10-