Method and system for the decoupled control of a medical multi-variable pump system

Abstract

The invention relates to a method and a system for controlling a medical pump, for example a liquid pump or an insufflator, which is used for expanding a body cavity during a surgical operation. Such systems are often susceptible to pressure fluctuations, in particular during spontaneous suction processes or in the event of leakages. In order to ensure stable pressure controlling, the invention proposes a method in which the pump system is modelled as a multi-variable system (MIMO) having coupled inputs and outputs. Controlling is carried out by means of a state feedback the control rules of which are derived from a dynamically enhanced state space model. As a result of this approach, the controlled variables, such as the cavity pressure and the total outflow, are decoupled and linearised. This enables independent, precise and robust controlling of the pressure in the body cavity, even in the event of severe interferences, as a result of which the stability and safety of the medical operation are improved.

Description

[0001] Method and system for decoupled control of a medical multi-size pump system

[0002] Subject of the invention

[0003] The invention relates to a method and a system for controlling a medical pump, for example, a fluid pump or an insufflator, used to inflate a body cavity during a surgical procedure. To ensure stable pressure control, the invention proposes a method in which the pump system is modeled as a multivariable input / output (MIMO) system with coupled inputs and outputs. Control is achieved by means of state feedback, the control law of which is derived from a dynamically extended state-space model. This approach decouples and linearizes the controlled variables, such as cavity pressure and total outflow.

[0004] State of the art

[0005] The application of medical roller pumps for distending body cavities works as follows: A reservoir with a fluidic connection to the body cavity is attached to a bracket above the device. The occluding roller impeller allows the fluid to be pumped into the body cavity at a predetermined pressure. The tubing assembly thus serves to transport the fluid sterilely from the reservoir into the body cavity. This is achieved by a segment where the roller impeller, driven by peristalsis, propels the fluid (i.e., pushes partial volumes along a circular arc towards the body cavity). This segment is pre-tensioned to ensure secure occlusion. Behind the roller impeller, a pressure dome with one or more transmission diaphragms is fluidically connected to the tubing. One or more pressure sensors, positioned behind the transmission diaphragms, measure the pressure generated by the pump within the tubing.This measured pressure represents a pre-pressure, as the flow resistance reduces the pressure in the body cavity, which the device can compensate for. Measuring the pressure in the body cavity along the length of the tubing (typically 1.5 to 3 meters) from the roller wheel to the body cavity is typically not feasible, because the expansion relies on maintaining pressure in the body cavity, and there are many leaks in the body cavity through which fluid escapes and reduces the pressure. Until the pressure is equalized, a reduction in expansion is observed, which hinders the medical procedure by obstructing visibility or insufficient space for minimally invasive instruments.

[0006] The user typically sets a desired flow rate through the body cavity on the device so that the view inside the body cavity is clear while keeping fluid consumption as low as possible.

[0007] The fluid enters the body cavity via connections to so-called trocars, which are tubes containing a seal through which minimally invasive instruments and endoscopes can be inserted. Next to these tubes, there is usually a connection for fluid administration with a closable stopcock. In arthroscopy, so-called I / O shafts are used as trocars. These shafts have an inflow stopcock on one end and an outflow stopcock on the other, both of which connect to the channel through which the fluid enters the body cavity. Because the two stopcocks are not separate, a large portion of the irrigation fluid is drawn in through the inflow stopcock and immediately aspirated out through the outflow stopcock instead of being delivered into the body cavity.

[0008] In contrast, so-called 2-port configurations exist, in which the irrigation fluid is drawn into the body cavity via the inflow valve on one trocar / port and out of the body cavity via a second trocar / port.

[0009] The adverse behavior arises from the fact that the control algorithm, in its original calibration, assumes a 2-port configuration. Therefore, the estimation of body cavity pressure fails with I / O sheaths, and overpressure and oscillations can occur at high suction rates.

[0010] A typical countermeasure is to specifically calibrate the control algorithm. This allows it to handle I / O shafts effectively, but results in performance losses in standard 2-port configurations. Under normal circumstances, the control algorithm cannot determine the configuration of the trocar connections to the body cavity, and the developer must therefore assume the worst-case scenario, namely the I / O shaft configuration. A disadvantage of many medical (dual-roller) pumps is that the pump's outflow side operates independently of the inflow side. This means that pressure drops caused by the pump's own active suction cannot be quickly compensated for. Another disadvantage is that there is no regulation of the outflow side once larger external leaks occur. Consequently, the pressure in the body cavity cannot be kept constant.

[0011] In order to keep the pressure in the body cavity constant and to counteract the significant influence of leaks and suction instruments, such as shavers or high-frequency surgical devices, or also through variations in the normal flow rate via a so-called outflow line, the amount of fluid suctioned or leaking from the body cavity is included in the pumping quantity and pumped additionally before a pressure drop becomes visible at the detector.

[0012] The shaver / HF detection, which identifies whether a connected shaver or HF device is switched on or off, is originally intended to change the position of an exclusive pinch valve, which accordingly opens or closes the suction lines (see also DE102010047349 A1 or DE102018009537A1). However, this signal can also be used to determine whether fluid is being aspirated via a 2-port configuration. If an HF or shaver is active, there is clearly no I / O shaft configuration. With this information, the control algorithm can deviate from the worst-case calibration corresponding to the I / O shaft and switch to the original 2-port configuration. This enables more precise pressure control when a shaver or HF device is switched on. Currently, there is no peristaltic pumping system that integrates the information that suction is being carried out via a shaver or HF device into the control algorithm.The control algorithms of other pump systems on the market are designed to be independent of the suction source.

[0013] Inventive solution

[0014] The solution according to the invention for the aforementioned disadvantages consists of treating the complex, coupled pump system not as separate control loops, but as a multi-input multiple system (MIMO). This problem is solved by a control method that applies state feedback based on a MIMO state-space model of the system. According to the invention, the control law is derived from a dynamically extended version of the model, in which one of the control signal inputs is extended to a new state variable by an integrator. This control engineering approach decouples and linearizes the system outputs, in particular the pressure in the body cavity and the total outflow. This enables independent, precise, and robust control of the individual output variables, leading to a fundamental improvement in system stability.

[0015] The invention therefore relates to a method for controlling a medical pump, wherein a fluid is pumped into a body cavity by means of a supply line via a controlled pump device, wherein the supply line contains a pressure sensor which measures a pressure in the line, and wherein a controllable suction pump suctions the fluid out of the body cavity via a second line,

[0016] characterized in that the pumping device and the suction pump are operated according to a control law of a state feedback derived from a state-space model of a multivariable system (MIMO) of the pumping system, wherein the MIMO state-space model defines at least one control signal for the pumping device and one control signal for the suction pump as inputs and at least one signal representative of a pressure in the body cavity and one signal representative of a fluid outflow as outputs,

[0017] and wherein the state feedback control law is derived from a dynamically extended version of the MIMO state space model in which one of the control signal inputs is extended by an integrator to a new state variable, so that the inputs and outputs of the system are decoupled and linearized.

[0018] In an advantageous embodiment, information about the activation of an external suction device, such as a shaver or a high-frequency surgical device, is integrated into the control law. This allows the system to proactively counteract an expected pressure drop before it becomes measurable at the pressure sensor. Furthermore, the system can use this information to automatically select the appropriate calibration for the control law, for example, to switch between a 2-port configuration and an I / O shaft configuration, thus optimizing the control performance for the respective situation. In this further development of the invention, when a high-suction device (e.g., a shaver) is activated, the system automatically adjusts the control law accordingly.(Shaver) activates the pressure control even before the pressure drop occurs, since the pressure drop due to the long hose lines is only detectable later at the pressure sensor on the medical pump: The activation is detected, for example, analogously to the one in DE10233053 "Device for flushing a body cavity" or based on the power required for activation, as described in EP 2165720, US11 / 642457 or US15 / 077348.

[0019] In another advantageous embodiment, the amount of fluid escaping through leaks is estimated. Based on this estimate, the suction power of the outflow pump is reduced to maintain the total outflow at a desired level. This effectively minimizes the unnecessary consumption of rinsing fluid in the event of external leaks.

[0020] Leakage, as an unknown quantity, is preferably estimated using a disturbance observer. A Kalman filter is particularly preferred for this purpose, evaluating the pressure sensor readings, pump speeds, and the estimated pressure in the body cavity to precisely determine the leakage volume. The resulting total outflow, comprised of the active suction and the estimated leakage, serves as one of the decoupled output variables in the control law.

[0021] The goal of leakage estimation is to ensure that any additional leaks do not increase the total volume of fluid pumped (due to inflow control). This is achieved by reducing the suction power of the outflow impeller. An estimation of the additional leakage volumes is required, which corresponds to the amount of (additional) fluid (distension medium) in the cavity. The aim is to keep the volume of fluid in the cavity constant to ensure proper expansion. This can be determined via the pressure in the body cavity. The system is in a dynamic equilibrium between, on the one hand, inflow and outflow, both of which are known quantities based on the rotations of the impellers, and, on the other hand, the leakage as a disturbance, where the leakage is an unknown quantity that must be controlled. [Description of the figures]

[0022] Further features and advantages of the invention will become apparent from the following description of exemplary embodiments with reference to the accompanying figures. These show:

[0023] Figure 1 shows a basic block diagram of a multi-input multi-input (MIMO) system.

[0024] Figure 2 shows a block diagram of state feedback in an open control loop.

[0025] Figure 3 shows the principle of decoupling for a coupled system.

[0026] Figure 4 shows a block diagram of an integrated state feedback system in a closed control loop.

[0027] Figure 5 shows a schematic diagram of the pressure curve of a pump according to the state of the art during suction.

[0028] Figure 6 shows a schematic diagram of the stabilized pressure profile according to the invention during suction.

[0029] Figure 7 shows a block diagram of the integrated state feedback with dynamic feedback according to the invention.

[0030] Figure 8 shows a diagram with the results of the control behavior according to the state of the art in the real test.

[0031] Figure 9 shows a diagram with the results of the control behavior according to the invention in the real test.

[0032] Example 1: Reducing Outflow to Save Medium. Example 1 describes a simplified embodiment for a better understanding of the complete invention described later. The full solution according to the invention is described in 'Example 3', which implements complete decoupling and linearization.

[0033] The control algorithm according to the invention is designed to reduce the suction in the outflow when the pressure at the sensor drops AND the total outflow (i.e., the sum of suction and leakage) corresponds to the body cavity flow rate set by the user. To achieve this, the value of the total outflow (i.e., the sum of inflow and leakage) is determined and stored at regular intervals. This occurs as soon as a pressure value defined for the body cavity is exceeded. Once this condition is met, the current inflow is set to equal the total outflow. This also occurs if the pressure remains constant for a defined period. The method used to achieve the control algorithm is multi-variable control with the following input sensor variables: pressure downstream of the inflow impeller, inflow impeller speed, and outflow impeller speed, as well as the following output variables: inflow impeller speed and outflow speed.This is done using the following intermediate parameters: joint pressure estimator, 'total outflow' estimator, retention volume estimator = balancing ln / out. Additional boundary conditions determined may include the flow resistance on the inflow side, the hose volume, and others.

[0034] As a pragmatic approach, the flow setting in the user interface (GUI) is used, and the initial state is assumed to be inflow = outflow. The devices evaluate instrument detection, i.e., the determination of the flow resistance between the pressure measurement point and the distal end of the instrument, and thereby limit the outflow speed to prevent the body cavity from being emptied.

[0035] Inflow and outflow volumes are balanced by the pressure change coupled with the flow, which is determined by the rotational speed of the inflow impeller. At the initial stage of this balancing process, i.e., at the start of the medical procedure, the inflow volume is set to the total outflow. This is maintained after the control system starts until a predetermined value in the joint pressure estimator is exceeded. This value is the pressure in the body cavity determined from the pressure sensor, taking flow resistance into account. Once this value is exceeded, the outflow impeller starts rotating. This also occurs if the joint pressure value in the pressure estimator remains constant over a specific period.

[0036] If the leakage increases, the value in the pressure estimator for the joint pressure drops, and the pressure regulator's algorithm increases the inflow. This is added to the "total outflow" value. If the inflow volume exceeds a certain predefined value, the outflow speed is reduced to prevent the joint pressure from dropping. This reduction occurs "from above," meaning by initially reducing a fraction, e.g.,

[0037] The flow rate is reduced by 90% from the amount determined by the pressure estimator, then increased in small increments until the pressure estimator shows a slight reduction in pressure within the body cavity. Optionally, the inflow rate can then be increased again until the pressure estimator reaches its value prior to the leak.

[0038] For embodiment 1, this means that there is an initially independent inflow control and the outflow control is only activated when the 'total outflow', i.e., initially the inflow quantity at constant pressure, exceeds a predefined threshold.

[0039] Example 2: Coupling of Inflow Control and Outflow Control. Example 2 also describes a simplified embodiment for a better understanding of the complete invention described later. The full solution according to the invention is described in "Example 3", which implements complete decoupling and linearization.

[0040] In another embodiment, if the inflow volume, measured via the rotational speed of the inflow impeller, is increased, the outflow volume is increased in the same way, regardless of the pressure estimate. This increase in the outflow has an upper limit to prevent exceeding a user-defined maximum flow value.

[0041] In this system, the outflow rate is only controlled when the total outflow exceeds a threshold. A disturbance observer (e.g., a Kalman filter) is used to measure the total outflow by evaluating changes in pressure. This involves jointly evaluating the pressure sensor reading (hose pressure), the pressure estimator for the joint pressure, and the inflow and outflow rotational speeds to estimate the total outflow.

[0042] This assumes that a change in the outflow speed always results in a change in the 'total outflow,' the magnitude of which is not observable; a change to zero, i.e., no outflow, is also possible. Three sensor values ​​are required for this: inflow pressure, inflow impeller speed, and outflow impeller speed. Example 3: Controller design with feedback

[0043] The third embodiment takes into account the non-linear properties of the described pump systems. This means that these systems have non-linear mathematical characteristics and are therefore often mathematically more difficult to handle than linear systems. On the other hand, an analysis and design method based on the linear component can be determined for linear systems. However, it is rarely possible to transform non-linear systems into linear systems.

[0044] In contrast to linear system theory, where the mathematical description of the model is characterized and solved by differential equations and transfer functions, nonlinear system theory requires a higher mathematical effort than linear system theory.

[0045] The object of the invention is therefore to approximate a nonlinear model system into a linear system.

[0046] To describe the mathematical model of nonlinear systems, the state differential and initial equation are represented in the following form:

[0047] ẋ = f(x, u) (1.1)

[0048] y = g(x, u) (1.2)

[0049] The vector function and state vector x are n-dimensional, the input variable u is an m-dimensional input vector, and the output vector y and output vector function g are r-dimensional. This form is called a multiple-variable system or MIMO system (multiple input, multiple output). A single-variable system or SISO system (single-input, single-output) is a system in which the input variable u and the output variable y are one-dimensional. MIMO systems are multi-variable systems in which the system has multiple inputs and outputs. The pump-cavity system presented below is a MIMO system with the property that the number of input variables equals the number of output variables (see Figure 1). In most cases, changes to an input variable not only affect its corresponding output variable but can also affect other output variables (Figure 1).The coupling of the system increases with the number of input variables that an output variable influences. According to the invention, the coupled system is decoupled and linearized by a suitable method (e.g., input / output linearization). As shown in Figures 2 and 3, the decoupling is achieved via state feedback.

[0050] The advantage of this method is that the controller design can be carried out locally for each individual main path using control engineering methods. The coupling to the other main paths remains unconsidered (see Figure 4). Since the controller design (synthesis and dimensioning) can be carried out locally for each individual main path, high stability and robustness against parameter uncertainty can be achieved with respect to the overall system. Figure 5 shows the control behavior of a peristaltic pump according to the state of the art (assumption: SISO system [6]) when a desired suction is started at time ti (e.g., by the shaver) and stopped again at t2.

[0051] Figure 5 shows that at time ti the cavity pressure drops considerably and at t2 it overshoots as soon as internal suction (for example by a shaver) is started or stopped.

[0052] In contrast, the pump-cavity system according to the invention has the following properties due to the implementation of state feedback:

[0053] Minimizing the pressure drop as soon as suction is started by the double roller itself (e.g. shaver) (see Figure 6)

[0054] Minimizing overshoot as soon as a suction device (e.g., shaver) is stopped.

[0055] Improved accuracy in relation to the total output volume flow, especially when external leakage is present.

[0056] The solution according to the invention is based on the previously known principles of exact linearization. The differential equations of the pump-cavity system according to the invention are extended by an actuator (suction motor) in accordance with the master's thesis by Zeyßig, A. "Design and Testing of a Pressure Control Strategy for a Twin Roller Pump" (2012).

[0057] The nonlinear differential equations and the power of the pump system are given:

[0058]

[0059] q̇_Leak = 0

[0060] Assuming that the leakage q Leak which is not currently changing:

[0061] y₁ = p₂

[0062]

[0063] y₂ = q_Outflow = q_Suction + q_Leak = K₂n₂ + q_Leak

[0064] In this process, q SU ction is assumed to be a linear variable for the speed of the extraction motor n2 with a throttle constant K2:

[0065] q_Suction = K₂n₂,

[0066] Where K2 represents a throttle constant and n2 the speed of the extraction motor.

[0067] The initial non-linear system descriptions reveal that

[0068] ẋ = a(x) + B(x)u

[0069]

[0070] y = C(x) + Du

[0071] Here, the state vector x is n-dimensional, the input variable u is an m-dimensional input vector, and the output vector y and the output vector function g are r-dimensional. K₁ / C₁60

[0072] with a(x) =

[0073] C260

[0074]

[0075] 0

[0076] Q60

[0077] 0

[0078] C260

[0079] C(x) = [c₁; c₂] = [p₂; q_Outflow] x = [x₁; x₂; x₃] = [p₁; p₂; q_Leak]

[0080]

[0081] D = [d₁ᵀ; d₂ᵀ] = [0 0; 0 K₂]The variables have the meaning given in Table 1 (below).

[0082] Variable meaning of throttle constant for converting rotational speed into input volume flow [ml]

[0083] K2 Throttle constant for converting the rotational speed into the output volume flow [ml]

[0084] C₁ Hydraulic power of the supply hose C₂ Hydraulic connection power ζ₁, ζ₂ Resistance coefficients

[0085] n₁ Motor speed of the inlet pump [rpm]

[0086] n₂ Motor speed of the suction pump [rpm] p₁ Hydraulic pressure [mmHg]

[0087] P2 joint pressure [mmHg]

[0088] q_Leak Leakage [ml / min]

[0089] q_Suction Suction by the suction pump [ml / min]δ₁, δ₂ Relative proportion

[0090] B Input matrix

[0091] C Output matrix

[0092] D throughput matrix

[0093] u Input Variable

[0094] X state variable

[0095] V External Reference Variable

[0096]

[0097] Table 1: Parameters and description of the pump system. First, the relative proportion ö_i,i=1,2 for the pump system must be determined. The following equation applies:

[0098] f dLa ] 1 cx) “ 19» öj:= min -: - - B(x A 0 T ; j = 1,2...,n I 2 '

[0099] IC / > I

[0100] The first relative component of the system is δ₁ = 1

[0101] ∂c₁(x) / ∂x · B(x) = ∂p₂ / ∂x · B(x) =

[0010]

[0102]

[0103] The second relative part of the system δ₂ = 0, because the throughput term d₂ᵀ = [0 K₂] ≠ 0ᵀ.

[0104] The second step is to create a linear input-output dynamic of the system, as shown in the required Figure 2. The given formula for the linear input-output dynamic of the system is:

[0105] [u₁; ũ₂] = -D̂⁻¹([ĉ(x) + â(x)] - Λv(t))

[0106]

[0107] The derivation of this formula is already known.

[0108] The decoupling matrix for the system has the following structure:

[0109] D̂ := [d̂₁ᵀ; d̂₂ᵀ] ( 3 )

[0110]

[0111] The decoupling matrix of the pump-cavity system of (1) is:

[0112] D̂ := [d̂₁ᵀ; d̂₂ᵀ] = [∂c₁(x) / ∂x · B(x); d₂ᵀ] = [0 -K₂ / C₂60; 0K₂]

[0113]

[0114] d T 2

[0115] For the next steps, it is important that the decoupling matrix D is regular, as this matrix must be inverted. At first glance, it is apparent that the matrix D̂ is singular and therefore a different approach is required for the further procedure.

[0116] The singularity of the decoupling matrix can be avoided using the previously known method of dynamic extension. For this purpose, the input variable u2 is extended by an integrator and becomes a new state variable x4:= u2.

[0117] The extended system is

[0118] x = a(x) + B(x)uy = c(x)

[0119] 'Pi' P2 tfleak. X4.

[0120] y = (c₁(x); c₂(x)) = (p₂; q_Outflow) = (p₂; q_Suction + q_Leak) = (p₂; K₂x₄ + q_Leak)

[0121] With a(x) = [K₁ / C₁60(...); 1 / C₂60(...); 0; 0]

[0122] 1 / C₂60(-K₁ζ₂ / 2ζ₁ + K₁√((ζ₂ / 2ζ₁)² + 1 / ζ₁(p₁-p₂)) - (K₂x₄ + q_Leak))

[0123] 0

[0124]

[0125] 0

[0126]

[0127] With this approach, the second relative proportion ö2= 1 (based on formula ( 2 )):

[0128] ∂c₂(x) / ∂x · B(x) = ∂(K₂x₄ + q_Leak) / ∂x · B(x) = [0 0 1 K₂] · B(x) = [0 K₂] = d̂₂ᵀ ≠ 0ᵀ

[0129]

[0130] - 0 1- The first relative proportion is δ₁ = 2 (based on formula ( 2 )):

[0131]

[0132] The decoupling matrix ( 3 ) then has the following appearance:

[0133] ∂Lₐ¹c₁(x) / ∂x

[0134]

[0135] Therefore, D is regular and can be inverted:

[0136] D̂⁻¹ =

[0137] 0

[0138]

[0139] The previously known formula for calculating state feedback is as follows:

[0140] [ü2] = “^“VCcCx) + a(x)] - Av(t))

[0141]

[0142] The integrator at input u2 generates dynamic feedback in the state feedback (see Figure 7).

[0143] [u₁; u̇₂]|_{x₄=n₂} = -D̂⁻¹([ĉ(x) + â(x)] - Λv(t))

[0144] With ĉ(x) = [ĉ₁(x); ĉ₂(x)] = [Lₐ^δ₁c₁(x); Lₐ^δ₂c₁(x)] = [Lₐ²c₁(x); Lₐ¹c₂(x)]

[0145] ĉ₂(x) = Lₐ¹c₂(x)

[0146] 2C₂²ζ₁60²

[0147]

[0148] ĉ₂(x) = Lₐ¹c₂(x) = 0â(x) = [â₁(x); â₂(x)] = [Σ(k=0 to δ₁-1) α₁ₖLₐᵏc₁(x); Σ(k=0 to δ₂-1) α₂ₖLₐᵏc₂(x)] = [α₁₀c₁(x) + α₁₁Lₐc₁(x); α₂₀c₂(x)] = [α₁₀p₂ + α₁₁(...); α₂₀(K₂x₄ + q_Leak)]

[0149]

[0150] = K₁

[0151] The decoupled overall system now exhibits linear behavior and the transfer function is:

[0152] G(s) = y(t) / v(t) = [y₁(t) / v₁(t) 0; 0 y₂(t) / v₂(t)] = [G₁₁(s) 0; 0 G₂₂(s)]

[0153] G₁₁(s) = λ₁ / (s² + a₁₁s + a₁₀)

[0154]

[0155] ^2

[0156] G₂₂(s) = λ₂ / (s + a₂₀) The dynamics of the system have the degree of freedom to be shaped by the coefficients a₁₀, a₁₁, a₂₀, λ₁ and λ₂. In practice, however, the constraints of the setpoint limits restrict the free choice of the parameters a₁₀, a₁₁, a₂₀,

[0157]

[0158] and A2.

[0159] The aim is to verify whether the state feedback actually linearizes and decouples the input / output behavior of the overall system. For demonstration purposes, a simulation run based on the closed-loop control system is performed (see Figure 7). In the simulation, an external leakage of 200 ml / min begins at 250 seconds. The nominal cavity pressure and nominal flow rate are predefined in the simulation. The control behavior of the system according to the state of the art, i.e., without linearization and decoupling, is shown in Figure 8.

[0160] The results of the actual test (see Figure 8) show that as soon as a desired suction (in the middle curve) is started or stopped (at 30 seconds, 80 seconds, and 220 seconds), the cavity pressure (in the upper curve) briefly drops or overshoots. This effect is described in detail in Figure 5.

[0161] In this model, the multi-variable control according to the invention was not taken into account, i.e., the system has no decoupling between the output variables (cavity pressure and suction power) (see Figure 3).

[0162] In a structurally identical system with the control according to the invention, i.e. with decoupling and linearization, under otherwise identical conditions, the results look like those shown in Figure 9.

[0163] The measurement results from the actual test (see Figure 9) show that the pressure remains constant when only the intake is changed (see Figure 9, at 30 seconds, at 80 seconds, and at 220 seconds). Therefore, there are no undesirable pressure drops or overshoots.

[0164] Another desirable behavior is that, in the event of external leakage, the suction motor is throttled back to maintain the overall outflow (see Figure 9; an external leakage of 200 ml / min begins at 250 seconds). The decoupling and linearization according to the invention makes it possible to independently adjust and control parameters such as cavity pressure and suction. This allows, with respect to peristaltic pump systems used in medicine, suction through the shaver or another access point to be performed without the cavity pressure dropping or exceeding the limit (see Figure 9).

[0165] Furthermore, improved accuracy of pressure control with regard to extraction can be achieved through improved interference suppression.

Claims

Patent claims: 1.) Method for controlling a medical pump, wherein a fluid is pumped into a body cavity via a supply line using a controlled pumping device, the supply line contains a pressure sensor which measures the pressure in the line, and wherein an adjustable suction pump draws the fluid from the body cavity via a second line; characterized by the fact that The pumping device and the suction pump are operated according to a state feedback control law derived from a state-space model of a multivariable system (MIMO) of the pumping system, wherein the MIMO state-space model defines at least one control signal for the pumping device and one control signal for the suction pump as inputs and at least one signal representative of a pressure in the body cavity and one signal representative of a fluid outflow as outputs, and wherein the state feedback control law is derived from a dynamically extended version of the MIMO state-space model in which one of the control signal inputs is extended by an integrator to a new state variable, so that the inputs and outputs of the system are decoupled and linearized. 2.) The method of claim 1, characterized in that the method further comprises detecting an activation signal of an external suction device, in particular a shaver or a high-frequency surgical device, and that, in response to the detection of the activation signal, the pump device is proactively controlled to counteract an expected pressure drop in the body cavity before this is measured by the pressure sensor. The method of claim 2, characterized in that the proactive control comprises selecting a calibration for the state feedback control law that is appropriate to the detected suction situation, in particular switching from a calibration for an I / O shaft configuration to a calibration for a 2-port configuration. Method according to one of the preceding claims, characterized in that the method further comprises estimating a leakage volume, and that the rotational speed of the suction pump is reduced based on the estimated leakage volume in order to maintain a substantially constant total outflow. Method according to claim 4, characterized in that the leakage volume is estimated by means of a disturbance observer, in particular a Kalman filter, which evaluates changes in the pressure measured by the pressure sensor, the estimated pressure in the body cavity and the rotational speeds of the pumping device and the suction pump. Method according to one of the preceding claims, characterized in that the output signal of the MIMO state space model representative of a fluid outflow is an estimated total outflow composed of the volume pumped by the suction pump and an estimated leakage volume. Medical pump system, comprehensive: a controlled pumping device (inflow pump) for pumping a fluid into a body cavity; an adjustable suction pump (outflow pump) for suctioning the fluid from the body cavity; a pressure sensor; and a control unit configured to execute a method according to any one of claims 1 to 6.

Images