Digital pid control apparatus

0001The present invention relates to a digital PID control apparatus used in various process instrumentation control systems.

0002The PID (P: proportional action, I: integral action, D: derivative action) control mode is considered as the main mode of the feedback control directly acting on plant , event at present time where the plant are driven at substantially the optimal efficiency with an advanced flexibility, and the PID control mode is widely used in many process controls.

0003The basic formula of the PID control is expressed in a sum of a proportional action (P action) for outputting a calculation result which is proportional to a deviation, an integral action (I action) for outputting a calculation result which is proportional to an integration of the deviation, and a derivative action (D action) for outputting a calculation result which is proportional to a derivative of the deviation. The basic formula can be expressed in the form of transfer function as the formula (1) below:<maths id="math0001" num="(1)"><math display="block"><mrow><mtext>C(s) = MV(s) / E(s) = KP {1 + 1 / TI·s)</mtext><mspace linebreak="newline" /><mtext> + (TD·s) / (1 + η·TD·s)}</mtext></mrow></math><img file="EP0750236A2_D0001.tif" /></maths>    where C(s): transfer function of PID, MV(s): operation amount, E(s): deviation, KP: proportional gain, TI: integrated time, TD: derivative time, s: Laplace operator, and 1 /η: derivative time.

0004The formula (1) can be expressed in the form of velocity-type digital operation formula by use of a control operation period Δ, as the following formulas (2a) and (2b):<maths id="math0002" num="(2a)"><math display="block"><mrow><mtext>ΔMVn = KP {(en - en-1) + (Δt / TI) ·en + ΔDn}</mtext></mrow></math><img file="EP0750236A2_D0002.tif" /></maths><maths id="math0003" num="(2b)"><math display="block"><mrow><mtext>MVn = MVn-1 + ΔMVn</mtext></mrow></math><img file="EP0750236A2_D0003.tif" /></maths>

0005Further, the derivative output Dn can be expressed using backwards difference by the formula (3a) and (3b):<maths id="math0004" num="(3a)"><math display="block"><mrow><mtext>ΔDn = {TD / (Δt + η·TD)}·(en - en-1)</mtext><mspace linebreak="newline" /><mtext> - (Δt / (Δt + η·TD)}·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0004.tif" /></maths><maths id="math0005" num="(3b)"><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn</mtext></mrow></math><img file="EP0750236A2_D0005.tif" /></maths>    where MVn: operation amount at present point, MVn-1: operation amount at previous control operation period, ΔMVn: a difference in operation amount from previous time to present time, en: the amount of deviation at the present point, en-1: the amount of deviation at the immediately previous control operation period, Dn: derivative output at present point; Dn-1: derivative output at previous control operation period; ΔDn: a difference in derivative output from previous time to present time, and Δt + η·TD: effective differential gain.

0006FIG. 1 is a diagram showing the structure of digital PID control apparatus operating on the basis of the formulas (2a) to (3b).

0007In this apparatus, a deviation "en" between a target value SVn and a control amount PVn is obtained by a deviation operating unit 51, and then the deviation en is supplied to a velocity type proportional control unit 52, a velocity type integral portion 53 and a velocity type derivative control unit 54. The velocity type proportional control unit 52 calculates ΔPn = en - en-1 by a proportional adjustment operation, the velocity type integral controlling unit 53 calculates ΔIn = (Δt / TI)·en by integral adjustment operation, and the velocity type derivative control unit 54 calculates ΔDn by the operations of the formulas (3a) and (3b). Further, an output of each calculation result is guided to an adding unit 55, where an adding synthesis is carried out to obtain ΔPn + ΔIn + ΔDn.

0008After a while, a proportional gain KP is multiplied with an addition output from the adding unit 55, at a proportional gain unit 56, thereby obtaining a velocity type signal, that is, ΔMVn = KP(ΔPn + ΔIn + ΔDn), which is further supplied to a velocity type/position type signal converting unit 57. In this signal converting unit 57, the velocity type signal is converted into a position type manipulation signal MVn by performing an operation of MVn = MVn-1 + ΔMVn, and the converted signal is supplied to a control object 58. Thus, the controlling amount PVn and the target value SVn are equalized to each other, in other words, the deviation en (= SVn - PVn) is made zero.

0009It should be noted that in order to calculate ΔDn and Dn of the formulas (3a) and (3b), it is required that the control operation period Δt be sufficiently small as compared to (η·TD). However, in the DDC (direct digital control) employed in an actual plant, the relationship in amount between the control operation period Δt and (η·TD) is not satisfied in many cases, and there are some cases in which the relationship in amount is completely reversed.

0010In order to solve this problem, there is proposed a bilinear transfer mode as one of the derivative calculation methods which can be operated even if the control operation period Δt is larger than those of the formulas (3a) and (3b).

0011In connection with the derivative operation, when the bilinear transfer method is employed, and the differentiation conversion is carried out, the velocity type derivative digital operation formulas (4a) and (4b) as below can be obtained.<maths id="math0006" num="(4a)"><math display="block"><mrow><mtext>ΔDn = {2·TD / (Δt + 2·η·TD)}·(en - en-1)</mtext><mspace linebreak="newline" /><mtext> - (2·Δt / (Δt + 2·η·TD)}·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0006.tif" /></maths><maths id="math0007" num="(4b)"><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn</mtext></mrow></math><img file="EP0750236A2_D0007.tif" /></maths>

0012As is clear from the formulas (4a) and (4b), when the bilinear transfer method is employed, the control operation period Δt can be doubled as compared to those of the formulas (3a) and (3b).

0013However, when the above bilinear transfer method is employed in the control system under the condition represented by the formula (5) below, an oscillation response which involves a resonance phenomenon shown in FIG. 2, as an example, is obtained.<maths id="math0008" num="(5)"><math display="block"><mrow><mtext>(2·Δt)/(Δt + 2η·TD) > 1</mtext></mrow></math><img file="EP0750236A2_D0008.tif" /></maths>

0014The simulation conditions regarding what is shown in FIG. 2 are: the control operation period Δt = 0.5 sec, η = 0.1, the proportional gain KP, which is a PID parameter, = 2.5, the integral time TI = 3.2 se and the derivative time TD = 0.8 sec, with respect to the control object G(s) = e^-2s / (1 + 5s).

0015Therefore, with regard to the digital control operation methods such as expressed by the formulas (4a) and (4b), the control operation period Δt can be set high by employing the bilinear transfer method; however at the same time, the control may become very unstable, depending upon the relationship between the control operation period Δt and the derivative time TD.

0016The object of the present invention is to provide a digital PID control apparatus capable of maintaining the stability of the control and the safety of the plant, even for a digital derivative action employing a bilinear transfer method.

0017According to an aspect of the present invention, there is provided a digital PID (P: proportional, I: integral, D: derivative) control apparatus, in which a deviation is obtained from a controlling amount of an object to be controlled, and a target value, a velocity type PI control operation is carried out on the deviation, a velocity type D control operation is carried out on the deviation or the controlling amount, these velocity type control operation outputs are synthesized, and then converted into a position type manipulation signal, and this signal is supplied to the object to be controlled, the digital PID control apparatus comprising: a derivation processing unit including a judgment unit for judging a control operating period is a predetermined value or less, and an operation unit which carries out a lagged derivative operation by a bilinear transfer method when the control operation period is at the predetermined value or less, or carries out an exact derivative operation when the control operating period is larger than the predetermined value.

0018According to another aspect of the present invention, there is provided a digital PID (P: proportional, I: integral, D: derivative) control apparatus, in which a deviation is obtained from a controlling amount of an object to be controlled, and a target value, a velocity type PI control operation is carried out on the deviation, a velocity type D control operation is carried out on the deviation or the controlling amount, these velocity type control operation outputs are synthesized, and then converted into a position type manipulation signal, and this signal is supplied to the object to be controlled, the digital PID control apparatus comprising: a derivation processing unit including a judgment unit for judging whether K = 2·Δt / (Δt + 2η·TD) ≦ 1, or K > 1 by using a control operation period Δt, a derivative time TD and a derivative gain 1 / η, and an operation unit which carries out a lagged derivative operation by a bilinear transfer method when K ≦ 1, or carries out an exact derivative operation when K > 1.

0019According to still another aspect of the present invention, there is provided a digital PID (P: proportional, I: integral, D: derivative) control apparatus, in which a deviation is obtained from a controlling amount of an object to be controlled, and a target value, a velocity type PI control operation is carried out on the deviation, a velocity type D control operation is carried out on the deviation or the controlling amount, these velocity type control operation outputs are synthesized, and then converted into a position type manipulation signal, and this signal is supplied to the object to be controlled, the digital PID control apparatus comprising: a derivation processing unit further comprising: a judgment unit for judging whether K = 2·Δt / (Δt + 2η·TD) ≦ 1, or K > 1 by using a control operation period Δt, a derivative time TD and a derivative gain 1 / η, and <ul id="ul0001" list-style="none" compact="compact"><li>means for carrying out the velocity type D control operation on the deviation, including a means for extracting velocity type D control operation results ΔDn, ΔDn' and a derivative output Dn by performing the following operation:    when K ≦ 1,<maths id="math0009" num=""><math display="block"><mrow><mtext>ΔDn = {(2·TD / (Δt + 2η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(en - en-1) - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0009.tif" /></maths><maths id="math0010" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn</mtext></mrow></math><img file="EP0750236A2_D0010.tif" /></maths>    when K > 1,<maths id="math0011" num=""><math display="block"><mrow><mtext>ΔDn' = {(2·TD / (Δt + 2 η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(en - 2en-1 + en-2)</mtext></mrow></math><img file="EP0750236A2_D0011.tif" /></maths><maths id="math0012" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn';</mtext></mrow></math><img file="EP0750236A2_D0012.tif" /></maths> and</li><li>means for carrying out the velocity type D control operation on the controlling amount, including a means for extracting velocity type D control operation results ΔDn, ΔDn' and a derivative output Dn by performing the following operation:    when K ≦ 1,<maths id="math0013" num=""><math display="block"><mrow><mtext>ΔDn = {(2·TD / (Δt + 2η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(PVn-1 - PVn) - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0013.tif" /></maths><maths id="math0014" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn</mtext></mrow></math><img file="EP0750236A2_D0014.tif" /></maths>    when K > 1,<maths id="math0015" num=""><math display="block"><mrow><mtext>ΔDn' = {(2·TD / (Δt + 2η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(2PVn-1 - PVn - PVn-2)</mtext></mrow></math><img file="EP0750236A2_D0015.tif" /></maths><maths id="math0016" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn',</mtext></mrow></math><img file="EP0750236A2_D0016.tif" /></maths></li></ul>    where en: the amount of deviation at the present time, en-1: the amount of deviation at the time of the immediately previous control operation period, en-2: the amount of deviation at the time of one before the immediately previous control operation period, PVn-1: the control amount at the time of the immediately previous control operation period, PVn-2: the control amount at the time of one before the immediately previous control operation period, Dn: the derivative output at present time, and Dn-l: the derivative output at the time of the immediately previous control operation period.

0020According to still another aspect of the present invention, there is provided a digital PID (P: proportional, I: integral, D: derivative) control apparatus, in which a deviation is obtained from a controlling amount of an object to be controlled, and a target value, a velocity type PI control operation is carried out on the deviation, a velocity type D control operation is carried out on the deviation or the controlling amount, these velocity type control operation outputs are synthesized, and then converted into a position type manipulation signal, and this signal is supplied to the object to be controlled, the digital PID control apparatus comprising: a derivation processing unit further comprising: a judgment unit for judging whether K = 2·Δt / (Δt + 2η·TD) ≦ 1, or K > 1 by using a control operation period Δt, a derivative time TD and a derivative gain 1 / η, and <ul id="ul0002" list-style="none" compact="compact"><li>means for carrying out the velocity type D control operation on the deviation, including a means for extracting velocity type D control operation results ΔDn, ΔDn' and a derivative output Dn by performing the following operation:    when K ≦ 1,<maths id="math0017" num=""><math display="block"><mrow><mtext>ΔDn = {(2·TD / (Δt + 2η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(en - en-1) - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0017.tif" /></maths><maths id="math0018" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn</mtext></mrow></math><img file="EP0750236A2_D0018.tif" /></maths>    when K > 1,<maths id="math0019" num=""><math display="block"><mrow><mtext>ΔDn' = (TD / Δt)·(en - 2en-1 + en-2)</mtext></mrow></math><img file="EP0750236A2_D0019.tif" /></maths><maths id="math0020" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn';</mtext></mrow></math><img file="EP0750236A2_D0020.tif" /></maths> and</li><li>means for carrying out the velocity type D control operation on the controlling amount, including a means for extracting velocity type D control operation results ΔDn, ΔDn' and a derivative output Dn by performing the following operation:    when K ≦ 1,<maths id="math0021" num=""><math display="block"><mrow><mtext>ΔDn = {(2·TD / (Δt + 2η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(PVn-1 - PVn) - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0021.tif" /></maths><maths id="math0022" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn</mtext></mrow></math><img file="EP0750236A2_D0022.tif" /></maths>    when K > 1,<maths id="math0023" num=""><math display="block"><mrow><mtext>ΔDn' = (TD / Δt)·(2PVn-1 - PVn - PVn-2)</mtext></mrow></math><img file="EP0750236A2_D0023.tif" /></maths><maths id="math0024" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn',</mtext></mrow></math><img file="EP0750236A2_D0024.tif" /></maths></li></ul>    where en: the amount of deviation at the present time, en-1: the amount of deviation at the time of the immediately previous control operation period, en-2: the amount of deviation at the time of one before the immediately previous control operation period, PVn: the control amount at present time, PVn-1: the control amount at the time of the immediately previous control operation period, PVn-2: the control amount at the time of one before the immediately previous control operation period, Dn: the derivative output at present time, and Dn-1: the derivative output at the time of the immediately previous control operation period.

0021According to still another aspect of the present invention, there is provided a digital PID (P: proportional, I: integral, D: derivative) control apparatus, in which a deviation is obtained from a controlling amount of an object to be controlled, and a target value, a velocity type PI control operation is carried out on the deviation, a velocity type D control operation is carried out on the deviation or the controlling amount, these velocity type control operation outputs are synthesized, and then converted into a position type manipulation signal, and this signal is supplied to the object to be controlled, the digital PID control apparatus comprising: <ul id="ul0003" list-style="none" compact="compact"><li>a derivation processing unit further comprising: a judgment unit for judging whether K = 2·Δt / (Δt + 2η·TD) ≦ 1, or K > 1 by using a control operation period Δt, a derivative time TD and a derivative gain 1 / η; and</li><li>means for carrying out the velocity type D control operation on the deviation, including a means for extracting velocity type D control operation results ΔDn, ΔDn' and a derivative output Dn by performing the following operation:    when K ≦ 1,<maths id="math0025" num=""><math display="block"><mrow><mtext>ΔDn = {(2·TD / (Δt + 2η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(en - en-1) - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0025.tif" /></maths><maths id="math0026" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn</mtext></mrow></math><img file="EP0750236A2_D0026.tif" /></maths>    when K > 1,<maths id="math0027" num=""><math display="block"><mrow><mtext>ΔDn' = (TD / Δt)·(en - en-1)</mtext><mspace linebreak="newline" /><mtext> - {(Δt + 2η·TD / (2·Δt)}·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0027.tif" /></maths><maths id="math0028" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn';</mtext></mrow></math><img file="EP0750236A2_D0028.tif" /></maths> and    means for carrying out the velocity type D control operation on the controlling amount, including a means for extracting velocity type D control operation results ΔDn, ΔDn' and a derivative output Dn by performing the following operation:    when K ≦ 1,<maths id="math0029" num=""><math display="block"><mrow><mtext>ΔDn = {(2·TD / (Δt + 2 η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(PVn-1 - PVn) - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0029.tif" /></maths><maths id="math0030" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn</mtext></mrow></math><img file="EP0750236A2_D0030.tif" /></maths>    when K > 1,<maths id="math0031" num=""><math display="block"><mrow><mtext>ΔDn' = (TD / Δt)·(PVn-1 - PVn)</mtext><mspace linebreak="newline" /><mtext> - {(Δt + 2η·TD / (2·Δt)}·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0031.tif" /></maths><maths id="math0032" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn',</mtext></mrow></math><img file="EP0750236A2_D0032.tif" /></maths></li></ul>    where en: the amount of deviation at the present time, en-1: the amount of deviation at the time of the immediately previous control operation period, PV: the control amount at present time, PVn-1: the control amount at the time of the immediately previous control operation period, Dn: the derivative output at present time, and Dn-1: the derivative output at the time of the immediately previous control operation period.

0022According to still another aspect of the present invention, there is provided a digital PID (P: proportional, I: integral, D: derivative) control apparatus, in which a deviation is obtained from a controlling amount of an object to be controlled, and a target value, a velocity type PI control operation is carried out on the deviation, a velocity type D control operation is carried out on the deviation or the controlling amount, these velocity type control operation outputs are synthesized, and then converted into a position type manipulation signal, and this signal is supplied to the object to be controlled, the digital PID control apparatus comprising: <ul id="ul0004" list-style="none" compact="compact"><li>a derivation processing unit further comprising: a judgment unit for judging whether K = 2·Δt/(Δt + 2η·TD) ≦ 1, or K > 1 by using a control operation period Δt, a derivative time TD and a derivative gain 1 / η; and</li><li>means for carrying out the velocity type D control operation on the deviation, including a means for extracting velocity type D control operation results ΔDn, ΔDn' and a derivative output Dn by performing the following operation:    when K ≦ 1,<maths id="math0033" num=""><math display="block"><mrow><mtext>ΔDn = {(2·TD / (Δt + 2η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(en - en-1) - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0033.tif" /></maths><maths id="math0034" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn</mtext></mrow></math><img file="EP0750236A2_D0034.tif" /></maths>    when K > 1,<maths id="math0035" num=""><math display="block"><mrow><mtext>ΔDn' = (TD / Δt)·(en - en-1)</mtext><mspace linebreak="newline" /><mtext> - {(Δt / (Δt + η·TD)}·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0035.tif" /></maths><maths id="math0036" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn'; and</mtext></mrow></math><img file="EP0750236A2_D0036.tif" /></maths></li><li>means for carrying out the velocity type D control operation on the controlling amount, including a means for extracting velocity type D control operation results ΔDn, ΔDn' and a derivative output Dn by performing the following operation:    when K ≦ 1,<maths id="math0037" num=""><math display="block"><mrow><mtext>ΔDn = {(2·TD / (Δt + 2η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(PVn-1 - PVn) - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0037.tif" /></maths><maths id="math0038" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn</mtext></mrow></math><img file="EP0750236A2_D0038.tif" /></maths>    when K > 1,<maths id="math0039" num=""><math display="block"><mrow><mtext>ΔDn' = (TD / Δt)·(PVn-1 - PVn)</mtext><mspace linebreak="newline" /><mtext> - {Δt / (Δt + η·TD)}·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0039.tif" /></maths><maths id="math0040" num=""><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn',</mtext></mrow></math><img file="EP0750236A2_D0040.tif" /></maths>    where en: the amount of deviation at the present time, en-1: the amount of deviation at the time of the immediately previous control operation period, PV: the control amount at present time, PVn-1: the control amount at the time of the immediately previous control operation period, Dn: the derivative output at present time, and Dn-1: the derivative output at the time of the immediately previous control operation period.</li></ul>

0023Embodiments of the invention will now be described by way of example, with reference to the accompanying drawings, in which: - <ul id="ul0005" list-style="none" compact="compact"><li>FIG. 1 is a block diagram showing the structure of a conventional digital PID control apparatus;</li><li>FIG. 2 is a diagram showing a simulation result of the conventional digital PID control apparatus, obtained in the case where the bilinear transfer method is employed and the control operation period is prolonged;</li><li>FIGS. 3A to 3D are diagrams showing the open loop response characteristics of the conventional digital PID control apparatus, in the case where various control operation period is varied in various ways by the bilinear transfer method;</li><li>FIG. 4 is a block diagram showing the structure of a PID control apparatus according to the present invention;</li><li>FIG. 5 is a block diagram showing another example of the structure of the velocity type differentiating process unit provided in the digital PID control apparatus shown in FIG. 4;</li><li>FIGS. 6A and 6B are diagrams showing the response characteristics in the case where derivative action according to a first embodiment of the present invention is employed;</li><li>FIGS. 7A and 7B are diagrams showing the response characteristics in the case where derivative action according to a second embodiment of the present invention is employed;</li><li>FIGS. 8A and 8B are diagrams showing the response characteristics in the case where the derivative action according to a third embodiment of the present invention is employed;</li><li>FIGS. 9A and 9B are diagrams showing the response characteristics in the case where the derivative action according to a fourth embodiment of the present invention is employed; and</li><li>FIG. 10 is a diagram showing a simulation result obtained by the derivative action of the second embodiment.</li><li>Prior to the detailed description of the embodiments of the present invention, the relationship between the control operation period Δt and the derivative time TD, which causes instability in the control when using the digital derivative action method employing the bilinear transfer method, will now be described for the reader to be able to more easily understand the contents of the present invention.</li></ul>

0024When the increment (en - en-1) in the deviation in the formula (4a) is rendered zero, and a difference ΔDn in the derivative output from the immediately previous time to the present time becomes larger than the derivative output Dn-l at the immediately previous control operation period, the first half term of the formula is zero since the increment in the deviation is zero, and therefore the following formula (22) is obtained from the last half term of the formula (4a):<maths id="math0041" num=""><math display="block"><mrow><mtext>ΔDn = - {(2·Δt / (Δt + 2·η·TD)}·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0041.tif" /></maths> that is,<maths id="math0042" num="(22)"><math display="block"><mrow><mtext>2·Δt / (Δt + 2·η·TD) > 1</mtext></mrow></math><img file="EP0750236A2_D0042.tif" /></maths>

0025When a coefficient K is assigned to the left side of the formula (22), the following formula is obtained:<maths id="math0043" num="(23)"><math display="block"><mrow><mtext>K = 2·Δt / (Δt + 2·η·TD)</mtext></mrow></math><img file="EP0750236A2_D0043.tif" /></maths>

0026Based on the formula (23), the relationship between the control operation period Δt and the derivative time TD, will now be described.

0027First, as described in connection with the conventional example, when the control operation period Δt is rendered larger than η·TD, an oscillation occurs. In order to specify the cause for the oscillation, the open loop characteristics obtained when the digital derivative action method employing the bilinear transfer is applied, are examined while fixing the derivative time TD = 2 sec and the derivative gain η = 0.1, and varying the control operation period Δt, as can be seen in FIGS. 3A to 3D, and the obtained values in terms of the response properties are compared.

0028FIG. 3A illustrates an example when the control operation period Δt = 0.01 sec. In this example, K &lt;&lt; 1 (Δt &lt;&lt; 2·η·TD). As can be seen in this figure, derivative response characteristics similar to those of the backward difference method or the conventionally used simplified operation method expressed by the formula (3a) and (3b), can be obtained.

0029FIG. 3B illustrates an example when the control operation period Δt = 0.1 sec. In this example, K &lt; 1. As can be seen in this figure, derivative response characteristics different from those of the backwards difference method expressed by the formula (3a) and (3b), gradually starts to appear.

0030FIG. 3C illustrates an example when the control operation period Δt = 0.4 sec. In this example, K = 1 (Δt = 2·η·TD). As can be seen in this figure, complete derivative type response characteristics can be obtained.

0031FIG. 3D illustrates an example when the control operation period Δt = 0.6 sec. In this example, K > 1 (Δt > 2·η·TD). As can be seen in this figure, a pull-back phenomenon, in which the derivative output is inverted (in a negative direction), occurs. More specifically, in the case of a bilinear transfer, a closer approximation to the continuous type differentiation (ideal differentiation) than that of the case of the backwards difference method; however, in a region where Δt > 2·η·TD, a pull-back phenomenon occurs.

0032Consequently, when an operation control unit involving the derivative action in which a pull-back phenomenon occurs as shown in FIG. 3D, is applied to an object to be controlled, the response becomes unstable, and when the control operation period Δt is further prolonged, an oscillation occurs. Therefore, in order to obtain a stable derivative output, it is required that the derivative action should be carried out under the judgment condition where a pull-back phenomenon, which is an oscillation factor in the derivative action, does not occur.

0033Regarding the apparatus of the present invention, the above-stated formula (23), that is,<maths id="math0044" num="(24)"><math display="block"><mrow><mtext>K = 2·Δt > (Δt + 2·η·TD)</mtext></mrow></math><img file="EP0750236A2_D0044.tif" /></maths> is considered as a judgment condition value, and the derivative action formula is changed depending upon whether K is larger than 1 or smaller than 1. In this manner, a PID operation control unit, which is stable regardless of the characteristics of the object to be controlled, can be obtained, and the safety of the plant can be improved.

0034Embodiments of the present invention will now be described with reference to accompanying drawings.

(First Embodiment)

0035FIG. 4 is a block diagram showing the structure of a PID control apparatus according to the present invention. The control apparatus includes a deviation operation unit 1 for calculating a deviation en between a target value SVn and a control amount PVn (en = SVn - PVn), a velocity type proportional control unit 2 for calculating a proportional operation unit ΔPn by carrying out the following operation:<maths id="math0045" num=""><math display="block"><mrow><mtext>ΔPn = en - en-1</mtext></mrow></math><img file="EP0750236A2_D0045.tif" /></maths> using the deviation en and the deviation at the immediately previous time, en-1, obtained from the deviation operation unit 1, and    a velocity type integral control unit 3 for calculating an integral operation output ΔIn by carrying out the following operation:<maths id="math0046" num=""><math display="block"><mrow><mtext>ΔIn = (Δt / TI)·en</mtext></mrow></math><img file="EP0750236A2_D0046.tif" /></maths> using the deviation en obtained from the deviation operation unit 1, the control operation period Δt and the integral time TI.

0036In the apparatus, the velocity differentiating process unit 4 for calculating an operation output which is proportional to the derivative of the deviation en, is provided.

0037The velocity derivation processing unit 4 includes a velocity type derivative control unit 11 for carrying out a velocity type derivative action based on the following operation formula:<maths id="math0047" num="(25)"><math display="block"><mrow><mtext>ΔDn = {(A·2·TD) / (Δt + 2·η·TD)}·(en - en-1) - B·Dn</mtext></mrow></math><img file="EP0750236A2_D0047.tif" /></maths> using the bilinear transfer as a derivative term of the deviation en, <ul id="ul0006" list-style="none" compact="compact"><li>and</li><li>condition judgment unit 12 for judging whether a judgment reference value K of the formula (24) is K ≦ 1 or K > 1, setting correction coefficients A and B to the formula (25) of the derivative control unit 11 in accordance with the following condition:    when K ≦ 1,<maths id="math0048" num=""><math display="block"><mrow><mtext>A = 1, B = K,</mtext></mrow></math><img file="EP0750236A2_D0048.tif" /></maths> or    when K > 1,<maths id="math0049" num=""><math display="block"><mrow><mtext>A = 1, B = 1</mtext></mrow></math><img file="EP0750236A2_D0049.tif" /></maths></li></ul> and allowing the apparatus to carry out a derivative action.

0038Further, the apparatus includes an adding unit 5 for adding and synthesizing operation outputs from the control units 2 and 3 and the process unit 4, a proportional gain unit 6 for multiplying the synthesized output with a proportional gain KP, a velocity type/ position type signal conversion unit 7 for converting a velocity type signal to a position signal, and an object 8 to be controlled.

0039A difference ΔDn in the derivative output obtained from the velocity type derivative control unit 11 of the velocity type derivation process unit 4, will be as follows, depending upon the level of the judgment reference value K:    When K ≦ 1,<maths id="math0050" num="(26)"><math display="block"><mrow><mtext>ΔDn = {(2·TD) / (Δt + 2·η·TD)}·(en - en-1)</mtext><mspace linebreak="newline" /><mtext> - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0050.tif" /></maths>    K > 1,<maths id="math0051" num="(27)"><math display="block"><mrow><mtext>ΔDn' = {(2·TD) / (Δt + 2·η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(en - 2en-1 + en-2)</mtext></mrow></math><img file="EP0750236A2_D0051.tif" /></maths><ul id="ul0007" list-style="none" compact="compact"><li>where Dn: the derivative output at the present time, en: the amount of deviation at the present time;</li><li>en-1: the amount of deviation at the time of the immediately previous control operation period;</li><li>en-2: the amount of deviation at the time of one before the immediately previous control operation period;</li><li>1 / η: derivative gain and K: coefficient.</li></ul>

0040When the velocity type signal is converted into a position type differentiation signal by use of a difference ΔDn in derivative output, the following relationships can be induced:    When K ≦ 1,<maths id="math0052" num="(28)"><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn of the formula (26)</mtext></mrow></math><img file="EP0750236A2_D0052.tif" /></maths>    When K > 1,<maths id="math0053" num="(29)"><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn' of the formula (27)</mtext></mrow></math><img file="EP0750236A2_D0053.tif" /></maths>

0041With the above-described structure of the embodiment, when the judgment condition value K is equal to 1 or smaller, the derivative output by the bilinear transfer such as shown in FIG. 6A is obtained. However, when the judgment condition value K is larger than 1, the pull-back of the derivative output is completely cut as shown in FIG. 6B by performing the derivative action of a complete derivative type. Therefore, an oscillation response along with a resonance phenomenon, is never generated, and therefore the stability of control can be maintained.

0042As another example of the velocity type derivation processing unit 4, the following structure can be considered. That is, as shown in FIG. 5, a first velocity type derivative control unit 11A and a second velocity type derivative control unit 11B for carrying out operations of differences ΔDn and ΔDn' in derivative output such as in the formulas (26) and (27) are prepared in advance, and each of the control units 11A and 11B are switched over by means of a switch unit 13 in accordance with the result of judgment by the condition judgment unit 12, that is, whether K ≦ 1 or K > 1.

0043Further, the above-described embodiment has the structure in which a feedback signal, which indicates the controlling amount PVn, is guided to the deviation operation unit 1 for calculating the target value and the deviation. For example, it is possible that in the case of a proceeding derivative action control apparatus for carrying out a derivative action on a control amount, a difference in derivative output is calculated by the following operation:    When K ≦ 1,<maths id="math0054" num="(30)"><math display="block"><mrow><mtext>ΔDn = {(2·TD) / (Δt + 2·η·TD)}·(PVn-1 - PVn)</mtext><mspace linebreak="newline" /><mtext> - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0054.tif" /></maths>    K > 1,<maths id="math0055" num="(31)"><math display="block"><mrow><mtext>ΔDn' = {(2·TD / (Δt + 2·η·TD)}</mtext><mspace linebreak="newline" /><mtext> ·(2PVn-1 - PVn + PVn-2)</mtext></mrow></math><img file="EP0750236A2_D0055.tif" /></maths>

(Second Embodiment)

0044This embodiment involves a velocity derivation processing unit 4 which has the same structure as that shown in FIG. 4 or 5, except that the following correction coefficients A and B are set when K ≦ 1 and K > 1.

0045More specifically, under the following conditions: that is,    when K ≦ 1,<maths id="math0056" num=""><math display="block"><mrow><mtext>A = 1, B = K,</mtext></mrow></math><img file="EP0750236A2_D0056.tif" /></maths> or    when K > 1,<maths id="math0057" num=""><math display="block"><mrow><mtext>A = 1 / K, B = 1,</mtext></mrow></math><img file="EP0750236A2_D0057.tif" /></maths> the condition judgment unit 12 sets the correction coefficients A and B in the derivative control unit 11.

0046As a result, differences in derivative output, namely, ΔDn and ΔDn' are calculated by the following operation:    When K ≦ 1,<maths id="math0058" num="(32)"><math display="block"><mrow><mtext>ΔDn = {(2·TD) / (Δt + 2·η·TD)}·(en - en-1)</mtext><mspace linebreak="newline" /><mtext> - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0058.tif" /></maths>    When K > 1,<maths id="math0059" num="(33)"><math display="block"><mrow><mtext>ΔDn' = (TD / Δt)·(en - 2en-1 + en-2)</mtext></mrow></math><img file="EP0750236A2_D0059.tif" /></maths>

0047When the velocity type signal is converted into a position type differentiation signal by use of a differences ΔDn and ΔDn' in derivative output, the following relationships can be induced:    When K ≦ 1,<maths id="math0060" num="(34)"><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn of the formula (32)</mtext></mrow></math><img file="EP0750236A2_D0060.tif" /></maths>    When K > 1,<maths id="math0061" num="(35)"><math display="block"><mrow><mtext>Dn = Dn-1 + ΔDn' of the formula (33)</mtext></mrow></math><img file="EP0750236A2_D0061.tif" /></maths>

0048With the above-described structure of the embodiment, when the judgment condition value K is equal to 1 or smaller, the derivative output by the bilinear transfer such as shown in FIG. 7A is obtained. However, when the judgment condition value K is larger than 1, a derivative output is adjusted by the variation of the control operation period Δt, and the height of the derivative output is decreased as shown in FIG. 7B. However, the area of a differential is always constant, and therefore a complete derivative type derivative action is carried out. Therefore, the control waveform area same as that of the analog differentiation can be obtained. Consequently, the stability of the control can be further improved, and an effect similar to that of the analog differentiation can be obtained.

0049The difference of the derivative action on the control amount is transformed as in the first embodiment, and can be obtained by the following equations:    When K ≦ 1,<maths id="math0062" num="(36)"><math display="block"><mrow><mtext>ΔDn = {(2·TD) / (Δt + 2·η·TD)}·(PVn-1 - PVn)</mtext><mspace linebreak="newline" /><mtext> - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0062.tif" /></maths>    When K > 1,<maths id="math0063" num="(37)"><math display="block"><mrow><mtext>ΔDn' = (TD / Δt)·(2PVn-1 - PVn - PVn-2)</mtext></mrow></math><img file="EP0750236A2_D0063.tif" /></maths>

(Third Embodiment)

0050This embodiment involves a velocity derivation processing unit 4 which has the same structure as that shown in FIG. 4 or 5, except that the following correction coefficients A and B are set when K ≦ 1 and K > 1.

0051More specifically, under the following conditions:    when K ≦ 1,<maths id="math0064" num=""><math display="block"><mrow><mtext>A = 1, B = K,</mtext></mrow></math><img file="EP0750236A2_D0064.tif" /></maths> or    when K > 1,<maths id="math0065" num=""><math display="block"><mrow><mtext>A = 1 / K, B = 1 / K,</mtext></mrow></math><img file="EP0750236A2_D0065.tif" /></maths> the condition judgment unit 12 sets the correction coefficients A and B in the derivative control unit 11.

0052As a result, differences in derivative output, namely, ΔDn and ΔDn' are calculated by the following operation:    When K ≦ 1,<maths id="math0066" num="(38)"><math display="block"><mrow><mtext>ΔDn = {(2·TD) / (Δt + 2·η·TD)}·(en - en-1)</mtext><mspace linebreak="newline" /><mtext> - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0066.tif" /></maths>    When K > 1,<maths id="math0067" num="(39)"><math display="block"><mrow><mtext>ΔDn' = (TD / Δt)·(en - en-1)</mtext><mspace linebreak="newline" /><mtext> - {(Δt + 2·η·TD) / (2·Δt)}·Dn-1)</mtext></mrow></math><img file="EP0750236A2_D0067.tif" /></maths>

0053The position type derivative signal Dn can be obtained by use of differences ΔDn and ΔDn' in derivative output, as mentioned before.

0054The difference in the value of derivative action on the control amount can be obtained by transforming it as in the first embodiment by use of:    When K ≦ 1,<maths id="math0068" num="(40)"><math display="block"><mrow><mtext>ΔDn = {(2·TD) / (Δt + 2·η·TD)}·(PVn-1 - PVn)</mtext><mspace linebreak="newline" /><mtext> - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0068.tif" /></maths>    When K > 1,<maths id="math0069" num="(41)"><math display="block"><mrow><mtext>ΔDn = (TD / Δt)·(PVn-1 - PVn)</mtext><mspace linebreak="newline" /><mtext> - {(Δt + 2·η·TD) / (2·Δt)}·Dn-1)</mtext></mrow></math><img file="EP0750236A2_D0069.tif" /></maths>

0055With the above-described structure of the embodiment, when the judgment condition value K is equal to 1 or smaller, the derivative output by the bilinear transfer such as shown in FIG. 8A is obtained. However, when the judgment condition value K is larger than 1, the derivative output is gradually and stepwise decreased. However, as shown in FIG. 8B, the portion corresponding to the pull-back portion in the conventional case appears in the positive direction, while keeping the area of the derivative output of each control operation period at constant. Consequently, the safety of the control can be maintained.

(Fourth Embodiment)

0056This embodiment involves a velocity derivation processing unit 4 which has the same structure as that shown in FIG. 4 or 5, except that the following correction coefficients A and B are set when K ≦ 1 and K > 1.

0057More specifically, under the following conditions:    when K ≦ 1,<maths id="math0070" num=""><math display="block"><mrow><mtext>A = 1, B = K,</mtext></mrow></math><img file="EP0750236A2_D0070.tif" /></maths> or    when K > 1,<maths id="math0071" num=""><math display="block"><mrow><mtext>A = 1 / K, B = Δt / (Δt + η·TD),</mtext></mrow></math><img file="EP0750236A2_D0071.tif" /></maths> the condition judgment unit 12 sets the correction coefficients A and B in the derivative control unit 11.

0058As a result, differences in derivative output, namely, ΔDn and ΔDn' are calculated by the following operation:    When K ≦ 1,<maths id="math0072" num="(42)"><math display="block"><mrow><mtext>ΔDn = {(2·TD) / (Δt + 2·η·TD)}·(en - en-1)</mtext><mspace linebreak="newline" /><mtext> - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0072.tif" /></maths>    When K > 1,<maths id="math0073" num="(43)"><math display="block"><mrow><mtext>ΔDn' = (TD / Δt)·(en - en-1)</mtext><mspace linebreak="newline" /><mtext> - {Δt / (Δt + η·TD)}·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0073.tif" /></maths>

0059The position type derivative signal Dn can be obtained by use of differences ΔDn and ΔDn' in derivative output, as mentioned before.

0060The difference in the value of derivative action on the control amount can be obtained by transforming it as in the first embodiment by use of:    When K ≦ 1,<maths id="math0074" num="(44)"><math display="block"><mrow><mtext>ΔDn = {(2·TD) / (Δt + 2·η·TD)}·(PVn-1 - PVn)</mtext><mspace linebreak="newline" /><mtext> - K·Dn-1</mtext></mrow></math><img file="EP0750236A2_D0074.tif" /></maths>    When K > 1,<maths id="math0075" num="(45)"><math display="block"><mrow><mtext>ΔDn = (TD / Δt)·(PVn-1 - PVn)</mtext><mspace linebreak="newline" /><mtext> - {Δt / (Δt + η·TD)}·Dn-1)</mtext></mrow></math><img file="EP0750236A2_D0075.tif" /></maths>

0061With the above-described structure of the embodiment, when the judgment condition value K is equal to 1 or smaller, the derivative output by the bilinear transfer such as shown in FIG. 9A is obtained. However, when the judgment condition value K is larger than 1, the derivative output is gradually and stepwise decreased. However, as shown in FIG. 9B, the portion corresponding to the pull-back portion in the conventional case appears in the positive direction, and response characteristics similar to those of the fourth embodiment can be obtained.

0062Therefore, in each of the embodiments of the present invention, even if the control operation period Δt is prolonged with regard to the derivative time TD, a stable derivative control can be obtained.

0063FIG. 10 is a diagram showing a simulation result in the case where the derivative action of the second embodiment is employed.

0064The simulation conditions of this case are the same as those of the case shown in FIG. 2 (prior art). More specifically, with respect to the control object G(s) = e^-2s / (1 + 5s), the following conditions are set: the control operation period Δt = 0.5 sec, η = 0.1, the proportional gain, which is a PID parameter, = 2.5, the integral time TI = 3.2 sec and the derivative time = 0.8 sec. With these conditions, K becomes larger than 1 (K > 1), and in the case of the bilinear transfer method, oscillation occurs in the prior art technique. However, with the apparatus of the present invention, the unstable region, which is created by the bilinear transfer method in the prior art, can be removed by setting the correction coefficients A and B and varying the derivative action method. Therefore, even in a DDC involving a long control operation period, a stable control can be realized by employing the bilinear transfer method.

0065According to the present invention, various effects described below can be achieved.

0066According to the first embodiment of the present invention, even in the case where the bilinear transfer method, if the judgment condition value K is larger than 1, a complete derivative type derivative action can be performed. Consequently, the pull-back of the derivative output can be completely cut, and the safety of the control can be maintained.

0067According to the second embodiment of the present invention, even in the case where the bilinear transfer method, if the judgment condition value K is larger than 1, a complete derivative type derivative action can be performed. Consequently, the area of the control waveform can be made the same as that of the analog derivative, and the safety of the control can be further improved.

0068According to the third embodiment of the present invention, even in the case where the bilinear transfer method, if the judgment condition value K is larger than 1, the derivative action method is varied. Thus, the derivative output is gradually and stepwise decreased, and the pull-back portion, which is created in the prior art technique, appears in the positive direction. Further, the area of the control waveform can be made the same as that of the analog derivative, and the safety of the control can be further improved.

0069According to the fourth embodiment of the present invention, a differential operation substantially the same as that of the third embodiment, can be achieved, and therefore the safety of the control can be further improved.

0070As described above, according to the present invention, even in the case of a long control operation period, a differentiation term operation which is appropriate for the control operation period, is performed. An oscillation which is caused by a change in control operation period, can be avoided, and it is possible to make an effective use of the derivative action.