A digital control unit for beam line magnet power supply

Abstract

The invention discloses a digital control unit for a proton therapy beam line magnet power supply, comprising a magnet main power supply and a magnet slave power supply; the magnet main power supply digital control unit comprises a DDS communication / instruction module, a table lookup module, a current distribution and parameter forwarding module, and a power supply control algorithm module; the table of the magnet main power supply table lookup module is distinguished according to the type of magnets on the beam line, and magnets of the same type are in the same table, and the table stores multiple sets of control parameter information of different energies for a certain type of magnet, each energy corresponds to a set of magnet current control parameters, and each set of magnet current control parameters includes energy parameters, current parameters, PID parameters, feedforward parameters, and magnetic field response standard value parameters; the invention combines a DDS communication module, multiple sets of PID adjustment parameters, refined PID adjustment parameters, and a double-loop+feedforward closed-loop control method, and progresses from four aspects and five levels, and finally solves the problem of rapid dynamic response of the beam line magnetic field during proton therapy.

Description

Technical Field

[0001] The present invention belongs to the field of beam line magnet power supply control in proton therapy, and more specifically, relates to a digital control unit of a beam line magnet power supply. Background Art

[0002] In a cyclotron-based proton therapy system, the proton beam energy ranges from 70 to 240 MeV on the beam line. During the patient's treatment, the proton beam energy will change from large to small as the treatment thickness changes from deep to shallow.

[0003] The beam line magnets are arranged at intervals starting from the energy degrader exit and ending at the treatment head, playing the role of guiding and controlling the beam; the beam track in the beam line is unique:

[0004]

[0005] In the above formula, ρ is the radius of the deflection part of the beam orbit, B is the magnetic field value of the deflection magnet, E is the beam energy, and E r is equal to 938MeV; it can be seen from the formula that when the beam energy (E) changes, in order to ensure that the beam deflection orbit radius (ρ) remains unchanged, the magnetic field of each deflection magnet on the beam line must be changed; and the method of changing the magnetic field of each deflection magnet is to change the output current of each deflection magnet power supply.

[0006] The magnet power supply is a constant current source, which is used to provide excitation current to the beam line magnet, thereby generating a magnetic field and controlling the direction of the beam trajectory. When the proton beam energy changes with different treatment thicknesses, the output current of the beam line magnet power supply must also be adjusted accordingly to make the beam line magnetic field match the change in beam energy and control the beam to follow the correct trajectory.

[0007] The difficulty in controlling the output current of the beam line magnet power supply of the proton therapy system lies in the fact that the communication of the command to the power supply takes a long time, affects the synchronization of the responses between the power supplies, and cannot guarantee that the dynamic response speed in any section of the entire energy range can meet the standard.

[0008] The reason why the communication of the command sent to the power supply takes a long time and affects the synchronization of the responses between the power supplies is that: first, the control of the beam line energy is the energy control of the entire beam line. The energy control of the entire beam line requires dozens or even hundreds of magnets to be distributed from the starting point to the end point of the beam line. The very demanding requirement is that no matter how large the number of beam line magnets is, only when all magnet power supplies complete the dynamic process of output current change within the specified time can the entire beam line be guaranteed to complete energy reduction within the specified time; second, but the current proton therapy system controls the beam line magnet power supply in a serial command manner: commands are sent to multiple magnets one by one. When the number of power supplies to be controlled is large, the time consumed in the communication process is greatly increased, and the communication time consumption is superimposed on the dynamic response time, which slows down the dynamic response speed. At the same time, there is a long time interval between the first magnet power supply that receives the command and the last magnet power supply, resulting in asynchronous responses between the power supplies.

[0009] The reason why it cannot guarantee that the dynamic response speed in any section of the entire energy range can meet the standard is: At this stage, the control of the power supply is mainly based on a set of PID parameter control methods. Since the beam line magnet is a large inductive load, inductive energy storage and hysteresis will affect the dynamic response. Since the energy storage of the magnet is different under different currents, the degree of influence on the dynamic response is also different: in the case of high beam energy, the required magnetic field value is large, so the magnet energy storage is large, and the resistance caused by the magnet energy storage to the descent process is large, and the descent speed will be slower, and the expected energy reduction position is often not reached within the set time range; while in the case of low beam energy, the required magnetic field value is small, so the magnet energy storage is small, and the resistance caused by the magnet energy storage to the descent process is small, and the magnetic field response may exceed the expected energy reduction position and produce overshoot; if you choose PID parameters suitable for high beam energy conditions, speed up the high beam energy The magnetic field response changes at low beam energy. The magnetic field response will overshoot at low beam energy, and the overshoot recovery process is generally slow, resulting in the response time at low energy exceeding the requirement. If PID parameters suitable for low beam energy are selected to prevent overshoot at low beam energy, the magnetic field response speed at high beam energy will be slower, resulting in the response time at high energy exceeding the requirement. The above results in a set of PID parameters being difficult to apply to the wider energy range of proton therapy (70-240MeV), and it cannot guarantee that the dynamic response speed of any section in the entire energy range can meet the standard. Summary of the invention

[0010] In order to solve the problems existing in the prior art, the present invention proposes a digital control unit for a beam line magnet power supply, aiming to solve the problems of slow dynamic response speed from the starting point to the end point of the beam line magnet power supply of the proton therapy system in the prior art and failure to ensure that the dynamic response speed in any section of the entire energy range can meet the standard.

[0011] The present invention proposes the following technical solutions to solve the technical problems:

[0012] 1. A digital control unit for a proton therapy beam line magnet power supply, characterized in that: the digital control unit is respectively arranged in a magnet power supply control system of each of a plurality of magnets in the proton therapy beam line; the plurality of magnets are arranged at intervals from the starting point to the end point of the beam line of a proton therapy accelerator system, characterized in that:

[0013] The magnet power supply of each magnet includes a magnet main power supply and a magnet slave power supply; the digital control unit includes a magnet main power supply digital control unit and a magnet slave power supply digital control unit;

[0014] The magnet main power digital control unit includes a DDS communication / command module, a table lookup module, a current distribution and parameter forwarding module, and a power control algorithm module; the magnet slave power digital control unit includes a power control algorithm module, the input end of which is connected to the current distribution and parameter forwarding module of the magnet main power supply, and the output end is connected to the magnet power supply;

[0015] The DDS communication / command module of the magnet main power supply has an input end externally connected to the treatment head system and an output end internally connected to the table lookup module; the DDS communication / command module enables the digital control unit of each magnet power supply to simultaneously subscribe to the DDS data published by the treatment head when they are in the same communication domain as the treatment head, and simultaneously obtain the required beam energy value from the DDS data;

[0016] The table of the magnet main power supply lookup module is distinguished according to the type of magnets on the beam line. The same type of magnets is the same table. The table stores multiple sets of control parameter information of different energies for a certain type of magnet. Each energy corresponds to a set of magnet current control parameters. Each set of magnet current control parameters includes energy parameters, current parameters, PID parameters, feedforward parameters, and magnetic field response standard value parameters; wherein the energy parameters, PID parameters, feedforward parameters, and magnetic field response standard value parameters are used as common parameters of the magnet main power supply and the magnet slave power supply; the current parameters are allocated from the magnet main power supply to the magnet slave power supply;

[0017] The table lookup module of the magnet main power supply performs a table lookup according to the parameters of "energy-current-PID and feedforward parameters-magnetic field response standard value" based on the beam energy information sent by the DDS communication / command module of each magnet main power supply digital control unit, obtains the corresponding current, PID and feedforward parameter setting values, and magnetic field response standard value, and sends the table lookup result to the current distribution and parameter forwarding module;

[0018] The current distribution and parameter forwarding module of the magnet main power supply performs current distribution between the main power supply and the slave power supply, and then sends the current distribution result and the energy parameters, PID parameters, feedforward parameters, and magnetic field response standard value parameters obtained by table lookup to the magnet main power supply and the power supply control algorithm module of the magnet slave power supply, respectively, as the basis for each power supply control algorithm module to control the power supply.

[0019] Furthermore, the table is looked up according to the parameters of "energy-current-PID and feedforward parameter-magnetic field response standard value", the energy parameter is the beam energy value / MeV, the current parameter is the magnet power supply output current value / A, and the magnet magnetic field is controlled by the power supply output current value / A; the PID parameter is the proportional coefficient KP of the current control loop 1 , integration coefficient KI 1 and differential coefficient KD 1 , and the proportional coefficient KP of the voltage control loop 2 , integration coefficient KI 2 and differential coefficient KD 2 ; The feedforward parameters are the feedforward link parameters of the current control loop and the feedforward link parameters of the voltage control loop; the PID current loop proportional coefficient KP 1 , integration coefficient KI 1 and differential coefficient KD 1 It is used to control the dynamic energy reduction or energy increase process for the current energy within a specified time, and the specified time includes 80 milliseconds; the PID voltage loop proportional coefficient KP 2 , integration coefficient KI 2 and differential coefficient KD 2 It is used to control the dynamic energy reduction or energy increase process for the current energy within a specified time, and the specified time includes 80 milliseconds.

[0020] Furthermore, the power supply control algorithm module includes a current and voltage dual closed-loop PID algorithm, a feedforward algorithm, and a dynamic process fine PID parameter adjustment algorithm; the parameter configuration in the current and voltage dual closed-loop PID algorithm is obtained based on a table lookup, and the configured parameter values ​​are different under different energy steps; in this way, the PID parameters configured in the high-energy reduction step and the low-energy reduction step within the entire energy range are the most suitable, so that the energy reduction speed, stability, and accuracy of each energy reduction step can all reach the optimum; the dynamic process fine PID parameter adjustment algorithm further subdivides the dynamic process of magnetic field change corresponding to each energy step into multiple refined dynamic process segments, and adjusts the response speed and response state of each refined dynamic process segment, thereby further improving the response speed of each energy step.

[0021] Furthermore, the current-voltage dual closed-loop PID algorithm is provided with two closed-loop control loops: the first loop is a current loop, and the feedback quantity and control object of the current loop are the output current; the second loop is a voltage loop, and the feedback quantity and control object of the voltage loop are the load voltage, and the output quantity of the loop is a PWM wave with a variable duty cycle, which drives the H-bridge in the main circuit of the power supply, thereby controlling the output current of the power supply; the second loop is nested in the first loop, and the output signal of the first loop is used as the input quantity of the second loop; the first and second loops both contain PID control links, which function to improve the dynamic performance of the power supply and the system stability, and speed up the response speed; the feedforward algorithm refers to adding a feedforward link in each of the current loop and the voltage loop, which works together with the PID controller to achieve greatly accelerated response speed and compensate for the lag and damping of the controlled system on the basis of ensuring system stability and accuracy.

[0022] Furthermore, the feedforward link input of the current loop is taken from the current setting value of the input current loop; the feedforward link input of the voltage loop is taken from the voltage input of the voltage loop; and the feedforward link parameters are obtained from the "energy-current-PID and feedforward parameters-magnetic field response standard value" lookup table.

[0023] Furthermore, the dynamic process fine PID parameter adjustment algorithm refers to dividing the dynamic process of output current change into multiple sections, and automatically adjusting the PID parameters according to the magnetic field response of the load magnet; the specific method is: continuously monitoring the magnetic field value of the load magnet in each section, and automatically adjusting the PID parameters of the two control loops in each section according to the size of the gap between the magnetic field value and the standard value, so that the PID parameters can always match the dynamic process well, speed up the response speed and reduce overshoot, and eliminate the influence of hysteresis. The dynamic process is divided into multiple sections, at least three sections.

[0024] Furthermore, the output current control algorithm also includes a safety interlock logic algorithm; the safety interlock logic algorithm is divided into internal interlock and external interlock. The internal interlock logic refers to the controller continuously monitoring certain indicators inside the power supply. When they are abnormal, the digital controller can immediately detect and control the power supply to shut down or keep the current state unchanged. The external interlock logic is that when an indicator outside the power supply is abnormal, the digital controller can immediately detect and control the power supply to shut down or keep the current state unchanged. In addition to the necessary interfaces, the external interlock should also have additional interfaces for users to customize settings.

[0025] Furthermore, each magnet power supply control system includes an FPGA, a DSP, a power supply main circuit, a current acquisition module ADC, a signal amplification module, and an AD conversion module, and is characterized in that: the FPGA is a ZYNQ chip, the ZYNQ chip can be equipped with a digital control method, and the LINUX system can be integrated in the ZYNQ to better control and supervise the output process, communication process, problem diagnosis process, etc. of the power supply; the DSP is controlled by the ZYNQ chip and can generate a PWM wave for driving the H bridge in the main circuit of the power supply according to the output of the digital controller; the current acquisition module ADC, the signal amplification module and the AD conversion module are used to collect the output of the power supply on the load side, and amplify and convert the collected analog signal into a digital signal, and feed it back to the digital controller.

[0026] Furthermore, the DDS communication / instruction module includes a communication domain setting module, a communication domain subscription module, and an instruction acquisition module; the communication domain setting module is used to set its own magnet power supply and the current treatment head in the same communication domain; the communication domain subscription module is used to subscribe to the instruction message issued by the current communication domain; the instruction acquisition module is used to obtain the required beam energy value from the instruction.

[0027] Advantages and effects of the present invention

[0028] The present invention combines the DDS communication module, multiple sets of PID adjustment parameters, refined PID adjustment parameters, and a closed-loop control method of dual loops + feedforward, and progresses from the above four aspects and five levels, and finally solves the problem of rapid dynamic response of the beam line magnetic field during proton therapy. The DDS communication module enables dozens or even hundreds of magnet power supplies to receive commands at the same time and respond at the same time, avoiding the problem of increasing the total time of the dynamic process due to excessive communication time between magnet power supplies. This method solves the problem of rapid dynamic response of the proton therapy beam line magnetic field from the first aspect; multiple sets of PID adjustment parameters solve the problem of inconsistent dynamic processes of different energies (or energy reduction lag or energy reduction overshoot), and provide different PID adjustment parameters for high energy steps and low energy steps to adapt to the energy reduction characteristics of different energies. This method solves the problem of rapid dynamic response of the proton therapy beam line magnetic field from the second aspect. The problem of rapid dynamic response of streamline magnetic field; the refinement of PID adjustment parameters solves the problem of inconsistent dynamic process within an energy step (fast in the first half and slow in the second half), and by further subdividing each energy step into multiple segments, and adjusting the PID parameters in the middle and last segments respectively, the problem of inconsistent dynamic process within an energy step is solved. This method solves the problem of rapid dynamic response of proton therapy beam line magnetic field from the third aspect; the closed-loop control method of double loop + feedforward, in which the two links of current loop feedforward and voltage loop feedforward further shorten the dynamic response time from tens of milliseconds to a few seconds. This method solves the problem of rapid dynamic response of proton therapy beam line magnetic field from the fourth aspect; in which the current loop + voltage loop directly solves the voltage disturbance in the voltage loop, avoiding the problem of slow response of a single current control loop to voltage disturbance. This method solves the problem of rapid dynamic response of proton therapy beam line magnetic field from the fifth aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1-1 This is a schematic diagram of the magnet power supply layout for the proton therapy accelerator beam line;

[0030] Figure 1-2 A schematic diagram of a digital control unit for the main power supply of the accelerator beam line magnet of the present invention;

[0031] Figure 1-3 A schematic diagram of a digital control unit of a power supply for a beam line magnet of an accelerator of the present invention;

[0032] Figure 2 It is a schematic block diagram of the current control algorithm of the digital control unit of the present invention;

[0033] Figure 3 This is a schematic diagram of the hardware circuit structure of the beam line magnet power supply control system of the present invention;

[0034] Figure 4-1The dynamic process response effect diagram of the present invention before and after the improvement of the same type of magnet power supply for current treatment energy;

[0035] Figure 4-2 This is a dynamic process response effect diagram of the present invention after the improvement of different types of magnet power supplies for current treatment energy;

[0036] Figure 5 This is a flow chart of the control method of the digital controller based on the beam line magnet power supply of the present invention. DETAILED DESCRIPTION

[0037] The present invention will be further explained below in conjunction with the accompanying drawings:

[0038] Design principle of the present invention

[0039] 1. Problems with the prior art: First, when the beam line is longer and the number of magnet power supplies is more (dozens or even hundreds), it is actually very difficult to make the last magnet power supply on the beam line to receive the energy reduction instruction reach the same dynamic process response time as the first magnet power supply to receive the energy reduction instruction, for example, 80 milliseconds. The last magnet power supply to receive the instruction will have a response delay or even timeout. The reason is that according to the traditional method, the energy change command is transmitted one by one from the control system of the first magnet power supply. Although the dynamic change process time of the previous magnet power supply and the next magnet power supply for the current treatment energy is not connected end to end but there is a time superposition before and after, when the beam line is very long and the number of magnets from the starting point to the end point is as many as dozens or even hundreds, the problem of delayed response of the end magnet power supply will be highlighted; Second, the energy change range is 70-240MeV. When DDS When the instruction is to reduce energy from a high energy zone, for example, from 240MeV to 238MeV, the energy reduction per unit time is relatively slow, while when it is reduced from a low energy zone, for example, from 72MeV to 70MeV, the energy reduction per unit time is relatively fast. The traditional method uses a set of PID parameters regardless of high energy or low energy. The PID parameters are used to control the speed, accuracy, and stability of a certain energy change dynamic process. According to the traditional method, only one set of PID parameters is used. If the parameters are suitable for high beam energy, the magnetic field response at low beam energy will overshoot and drop below the magnetic field standard value. The overshoot recovery process is generally slow, resulting in a response time exceeding the requirement at low energy. If the parameters are suitable for the PID parameters of low beam energy, the magnetic field response at low beam energy will not overshoot, but the magnetic field response speed at high beam energy will be very slow, and the response time will exceed the requirement.

[0040] 2. Difficulties in the design of the present invention: Difficulty 1: The proton therapy process is continuous. The thickness and area of ​​tumors in different parts are different. First, when the thickness of the tumor changes from the previous thickness to the next thickness, the treatment energy is required to change accordingly. There is a regulation for the time of each level of treatment energy change, for example, 80 milliseconds to complete the change from the previous energy to the next energy; secondly, when the treatment energy reaches the requirement, the treatment time of this energy must also meet the requirement, because the tumor area of ​​the current thickness must be treated by scanning. The treatment time determines the area of ​​the tumor scanned by the treatment head. The energy reduction time and the treatment time complement each other. If the energy reduction time exceeds the specified requirement (for example, 80 milliseconds), the excess time will occupy the treatment time (because the treatment time of the tumor area of ​​this thickness is predetermined). The estimated 1 minute of scanning treatment time is less than 1 minute. In this way, the current tumor area of ​​this thickness has not been treated, and it will jump to the tumor position of the next thickness, so that a part of the tumor area is omitted and the opportunity for treatment is missed. The reason why the energy reduction time exceeds the expected time is that the time when the magnet power supply at the end receives the command is delayed compared to the time when the dozens of magnet power supplies in front receive the command. If the time of receiving the command is delayed, the energy reduction time of the entire beam line will be delayed. If the energy reduction time of the entire beam line is delayed, the delayed time will occupy the time for treatment with the energy. The second difficulty: Not only are the energy reduction speeds between the high energy ladder and the low energy ladder different (energy storage effect), but also the energy reduction rate within each energy ladder is inconsistent. For example, when the DDS instruction is reduced from 240MeV to 238MeV, the energy reduction rate of the same dynamic process is not uniform, starting fast, slowing down in the middle, and even slower at the end. For example, when the energy is reduced to 50%, the reduction rate will slow down. Therefore, if we only improve from using one set of PID parameters for multiple energies to using multiple sets of PID control parameters for multiple energies, we still cannot completely solve the problem, and there is still room for improvement in response speed; The third difficulty is that voltage disturbances in a closed-loop control circuit will affect the current. If the voltage fluctuates due to temperature changes somewhere in the circuit, it will further affect the output current. If only current loop feedback control is used, the voltage will still fluctuate, which is only a temporary solution and will lead to decreased system stability. Moreover, there are many steps in converting the voltage error into a current error and then controlling it based on the error value of the output current, which is also time-consuming. The fourth difficulty is that if a voltage loop is added after the current loop PID, the voltage loop can be used to correct the error caused by the voltage disturbance in a timely manner, reducing the delay time of the current loop PID to correct the current error. However, since feedback control is a hysteresis control, using only feedback control can only ensure that the dynamic response time is on the order of tens of milliseconds, making it difficult to further improve the response speed.

[0041] 3. The solution of the present invention: First, a DDS communication module is adopted to solve the problem of delay in receiving commands of the magnet power supply at the end of the beam line by synchronously receiving commands; second, a method of adjusting multiple sets of PID parameters is adopted to set the PID parameters corresponding to different treatment energies according to empirical values ​​and store them in a table. When in use, the PID parameters of the corresponding energy are obtained by looking up the table; third, a fine PID adjustment method within an energy step is adopted to divide an energy step into at least three sections. When the magnetic field value of the magnet exceeds 50% of the expected value, the PID control parameters are adjusted to solve the problem of slow speed. When the magnetic field of the magnet exceeds 80% of the expected value, the PID control parameters are adjusted again to further solve the problem of slow response; fourth, the current loop + current loop feedforward and voltage loop + voltage loop feedforward methods are adopted to greatly shorten the dynamic response time of the magnetic field through the feedforward method.

[0042] According to the above invention principle, the present invention designs a digital control unit for a proton therapy beam line magnet power supply. Figure 1-1 , 1-2 、1-3、 Figure 2 , Figure 3 As shown, the digital control unit is respectively arranged in the magnet power supply control system of each magnet of a plurality of magnets in a proton therapy beam line; the plurality of magnets are arranged at intervals from the starting point to the end point of the beam line of a proton therapy accelerator system, and is characterized in that:

[0043] The magnet power supply of each magnet includes a magnet main power supply and a magnet slave power supply such as Figure 1-2 , 1-3 As shown; the digital control unit includes a magnet main power digital control unit and a magnet slave power digital control unit;

[0044] The magnet main power digital control unit is as Figure 1-2 As shown, it includes a DDS communication / command module, a table lookup module, a current distribution and parameter forwarding module, and a power control algorithm module; the magnet is driven by a power digital control unit such as Figure 1-3 As shown, it includes a power control algorithm module, the input end of which is connected to the current distribution and parameter forwarding module of the magnet main power supply, and the output end is connected to the magnet power supply;

[0045] The DDS communication / command module of the magnet main power supply is as follows Figure 1-2 As shown, its input end is externally connected to the treatment head system, and its output end is internally connected to the table lookup module; the DDS communication / command module enables the digital control unit of each magnet power supply to simultaneously subscribe to the DDS data published by the treatment head when they are in the same communication domain as the treatment head, and simultaneously obtain the required beam energy value from the DDS data;

[0046] The table of the magnet main power supply lookup module is distinguished according to the type of magnets on the beam line. The same type of magnets is the same table. The table stores multiple sets of control parameter information of different energies for a certain type of magnet. Each energy corresponds to a set of magnet current control parameters. Each set of magnet current control parameters includes energy parameters, current parameters, PID parameters, feedforward parameters, and magnetic field response standard value parameters; wherein the energy parameters, PID parameters, feedforward parameters, and magnetic field response standard value parameters are used as common parameters of the magnet main power supply and the magnet slave power supply; the current parameters are allocated from the magnet main power supply to the magnet slave power supply;

[0047] The table lookup module of the magnet main power supply performs a table lookup according to the parameters of "energy-current-PID and feedforward parameters-magnetic field response standard value" based on the beam energy information sent by the DDS communication / command module of each magnet main power supply digital control unit, obtains the corresponding current, PID and feedforward parameter setting values, and magnetic field response standard value, and sends the table lookup result to the current distribution and parameter forwarding module;

[0048] The current distribution and parameter forwarding module of the magnet main power supply performs current distribution between the main power supply and the slave power supply, and then sends the current distribution result and the energy parameters, PID parameters, feedforward parameters, and magnetic field response standard value parameters obtained by table lookup to the magnet main power supply and the power supply control algorithm module of the magnet slave power supply, respectively, as the basis for each power supply control algorithm module to control the power supply.

[0049] Furthermore, the table is looked up according to the parameters of "energy-current-PID and feedforward parameter-magnetic field response standard value", the energy parameter is the beam energy value / MeV, the current parameter is the magnet power supply output current value / A, and the magnet magnetic field is controlled by the power supply output current value / A; the PID parameter is the proportional coefficient KP of the current control loop 1 , integration coefficient KI 1 and differential coefficient KD 1 , and the proportional coefficient KP of the voltage control loop 2 , integration coefficient KI 2 and differential coefficient KD 2 ; The feedforward parameters are the feedforward link parameters of the current control loop and the feedforward link parameters of the voltage control loop; the PID current loop proportional coefficient KP 1 , integration coefficient KI 1 and differential coefficient KD 1 It is used to control the dynamic energy reduction or energy increase process for the current energy within a specified time, and the specified time includes 80 milliseconds; the PID voltage loop proportional coefficient KP 2 , integration coefficient KI 2 and differential coefficient KD 2It is used to control the dynamic energy reduction or energy increase process for the current energy within a specified time, and the specified time includes 80 milliseconds.

[0050] Furthermore, the power control algorithm module is as follows: Figure 2 As shown, it includes a current and voltage dual closed-loop PID algorithm, a feedforward algorithm, and a dynamic process fine PID parameter adjustment algorithm; the parameter configuration in the current and voltage dual closed-loop PID algorithm is obtained based on a table lookup, and the configured parameter values ​​are different under different energy steps; in this way, the PID parameters configured in the high energy reduction step and the low energy reduction step in the entire energy range are the most suitable, so that the energy reduction speed, stability, and accuracy of each energy reduction step can reach the optimum; the dynamic process fine PID parameter adjustment algorithm subdivides the dynamic process of magnetic field change corresponding to each energy step into multiple refined dynamic process segments, and adjusts the response speed and response state of each refined dynamic process segment, thereby further improving the response speed of each energy step.

[0051] Furthermore, the current and voltage dual closed-loop PID algorithm and the power supply control algorithm module are as follows: Figure 2 As shown, there are two closed-loop control loops: the first loop is a current loop, and the feedback quantity and control object of the current loop are the output current; the second loop is a voltage loop, and the feedback quantity and control object of the voltage loop are the load voltage, and the output quantity of the loop is a PWM wave with a variable duty cycle, which drives the H bridge in the main circuit of the power supply, and then controls the output current of the power supply; the second loop is nested in the first loop, and the output signal of the first loop is used as the input quantity of the second loop; the first and second loops both contain PID control links, which are used to improve the dynamic performance of the power supply and the stability of the system and speed up the response speed; the feedforward algorithm refers to adding a feedforward link in each of the current loop and the voltage loop, and working together with the PID controller to achieve greatly accelerated response speed and compensate for the hysteresis and damping of the controlled system on the basis of ensuring system stability and accuracy.

[0052] Furthermore, if Figure 2 As shown, the feedforward link input of the current loop is taken from the current setting value of the input current loop; the feedforward link input of the voltage loop is taken from the voltage input of the voltage loop; and the feedforward link parameters are obtained from the "energy-current-PID and feedforward parameters-magnetic field response standard value" lookup table.

[0053] Furthermore, the dynamic process fine PID parameter adjustment algorithm refers to dividing the dynamic process of output current change into multiple sections, and automatically adjusting the PID parameters according to the magnetic field response of the load magnet; the specific method is: continuously monitoring the magnetic field value of the load magnet in each section, and automatically adjusting the PID parameters of the two control loops in each section according to the size of the gap between the magnetic field value and the standard value, so that the PID parameters can always match the dynamic process well, speed up the response speed and reduce overshoot, and eliminate the influence of hysteresis. The dynamic process is divided into multiple sections, at least three sections.

[0054] Furthermore, the output current control algorithm also includes a safety interlock logic algorithm; the safety interlock logic algorithm is divided into internal interlock and external interlock. The internal interlock logic refers to the controller continuously monitoring certain indicators inside the power supply. When they are abnormal, the digital controller can immediately detect and control the power supply to shut down or keep the current state unchanged. The external interlock logic is that when an indicator outside the power supply is abnormal, the digital controller can immediately detect and control the power supply to shut down or keep the current state unchanged. In addition to the necessary interfaces, the external interlock should also have additional interfaces for users to customize settings.

[0055] Furthermore, each magnet power supply control system is as follows Figure 3 As shown, it includes FPGA, DSP, a main circuit of a power supply, a current acquisition module ADC, a signal amplification module, and an AD conversion module, and is characterized in that: the FPGA is a ZYNQ chip, the ZYNQ chip can be equipped with a digital control method, and the LINUX system can be integrated in the ZYNQ to better control and supervise the output process, communication process, problem diagnosis process, etc. of the power supply; the DSP is controlled by the ZYNQ chip, and can generate a PWM wave for driving the H bridge in the main circuit of the power supply according to the output of the digital controller; the current acquisition module ADC, the signal amplification module and the AD conversion module are used to collect the output of the power supply on the load side, and amplify the collected analog signal and convert it into a digital signal, which is fed back to the digital controller.

[0056] Furthermore, the DDS communication / instruction module includes a communication domain setting module, a communication domain subscription module, and an instruction acquisition module; the communication domain setting module is used to set its own magnet power supply and the current treatment head in the same communication domain; the communication domain subscription module is used to subscribe to the instruction message issued by the current communication domain; the instruction acquisition module is used to obtain the required beam energy value from the instruction.

[0057] Based on the above digital control unit of the proton therapy beam line magnet power supply, the present invention also designs a digital control method for the proton therapy beam line magnet power supply. Figure 5 As shown,

[0058] A digital control method for a proton therapy beam line magnet power supply,

[0059] Its characteristics are: the digital control method comprises the following steps:

[0060] Step 1: Set multiple sets of magnet power control parameters for different proton therapy energy requirements;

[0061] The multiple sets of magnet power supply control parameters include multiple sets of current parameter values, multiple sets of PID adjustment parameter values, multiple sets of current and voltage feedforward parameter values, and multiple sets of magnetic field response standard values;

[0062] Step 2: Set multiple magnet power supplies on the proton therapy beam line and the current therapy head to be in the same communication domain;

[0063] Step 3: The digital control units of all magnet power supplies in the same communication domain receive the DDS instruction at the same time and obtain the beam energy value from the DDS instruction;

[0064] Step 4: The digital control unit of each magnet power supply obtains the relevant current parameter value, PID adjustment parameter value, current and voltage feedforward parameter value, and magnetic field response standard value by looking up the table according to the current beam energy value;

[0065] Step 5. The digital control unit of each magnet power supply configures the PID link and feedforward link of the current loop and the PID link and feedforward link of the voltage loop in the dual-loop current control loop according to the relevant magnet power supply control parameters obtained by table lookup, and controls the output current based on the current set value as the standard.

[0066] Furthermore, all magnet power supply controllers in the same communication domain of step 3 simultaneously receive the DDS instruction and obtain the beam energy value from the DDS instruction. The specific process is as follows:

[0067] 1) Multiple magnet power supplies are arranged along the proton therapy accelerator beam line from the starting point to the end point;

[0068] The multiple magnet power supplies are used to achieve synchronous changes in beam energy from the starting point to the end point of the beam line;

[0069] 2) In the DDS communication module of each magnet power supply, set the magnet power supply and the treatment head system as the same communication domain;

[0070] 3) The treatment head system sends a DDS data packet containing instructions to the digital control units of all magnet power supplies in the communication domain;

[0071] 4) The digital control units of all magnet power supplies in the same communication domain receive the DDS command at the same time;

[0072] 5) Obtain beam energy value from DDS command.

[0073] Furthermore, the digital control unit of each magnet power supply in step 5 configures the PID link, feedforward link of the current loop and the PID link and feedforward link of the voltage loop in the current control algorithm according to the relevant magnet power supply control parameters obtained by table lookup, and controls the output current based on the current setting value as the standard. The specific process is as follows:

[0074] (1) Obtain the current parameter, PID parameter, and feedforward parameter for the current beam energy by looking up the table; the current parameter is the output current value of the magnet power supply / A; the PID parameter is the proportional coefficient KP of the current control loop 1 , integration coefficient KI 1 and differential coefficient KD 1 , and the proportional coefficient KP of the voltage control loop 2 , integration coefficient KI 2 and differential coefficient KD 2 ; The feedforward parameter is a feedforward link parameter of a dual-loop current control loop and a feedforward link parameter of a voltage control loop; the current parameter is used as a current setting value in a dual-loop current control algorithm;

[0075] (2) The current set value signal is input into the current control loop and the collected output current feedback value is subtracted to obtain the current error value. The current error value generates an action result signal after being acted on by the current loop PID controller. At the same time, the current set value signal also obtains an action result signal through the current loop feedforward link. The action result signals of these two control links are added together as the input signal of the voltage control loop.

[0076] (3) The voltage control loop input signal is subtracted from the collected load voltage feedback value to obtain a voltage error value. The voltage error value generates an action result signal after being acted upon by the voltage loop PID controller. At the same time, the voltage control loop input signal also obtains an action result signal through the voltage loop feedforward link. The action result signals of these two control links are added together to become the final control signal to control the main circuit of the power supply. The final control signal is in the form of a PWM wave with a variable duty cycle. The larger the duty cycle, the greater the final output current.

[0077] Furthermore, the specific control method of the current loop PID controller or the voltage loop PID controller is as follows:

[0078] A. The dynamic process for the current treatment energy is further divided into at least three detailed dynamic processes, so as to achieve fine control of the dynamic process and improve the response speed and performance; the dynamic process for each treatment energy is a dynamic energy reduction or energy increase process for each treatment energy;

[0079] B. The current loop PID controller or the voltage loop PID controller is set to a parallel connection of a proportional link, an integral link and a differential link, the coefficient of the proportional link is KP, the coefficient of the integral link is KI, and the coefficient of the differential link is KD;

[0080] C. Look up the table to get the proportional coefficient KP of the current control loop for each treatment energy 1 , integration coefficient KI 1 and differential coefficient KD 1 , and the proportional coefficient KP of the voltage control loop 2 , integration coefficient KI 2 and differential coefficient KD 3 ;

[0081] D. Continuously monitor the magnetic field value of the load magnet in each detailed dynamic process, and automatically adjust and modify the PID parameters of the two control loops according to the difference between the magnetic field value and the standard value, so that the PID parameters can always match the dynamic process well, speed up the response speed, reduce overshoot, and eliminate the influence of hysteresis;

[0082] Furthermore, the method D automatically adjusts and modifies the PID parameters of the two control loops according to the difference between the magnetic field value and the standard value, and the specific steps are as follows:

[0083] 1) Using the magnetic field response standard value obtained by table lookup and the magnetic field value collected in real time as reference, calculate the degree of completion of the dynamic process; the calculation formula is:

[0084] Dynamic process completion degree (%) = (B 磁场响应达标值 -B 实时磁场值 ) / (B 磁场响应达标值 -B 磁场初始值 )

[0085] The initial value of the magnetic field in the formula is the magnetic field value before the dynamic response, and can be approximately replaced by the standard value of the magnetic field response in the previous dynamic process;

[0086] 2) When the dynamic process of the current treatment energy reaches 50% completion, the proportional coefficient KP in the two-loop PID controller parameters is 1 and KP 2 Each increased by 20%;

[0087] 3) When the dynamic process of the current treatment energy reaches 80% completion, the integral coefficient KI in the two-loop PID controller parameters is 1 and KI 2 Increase by 20%;

[0088] Furthermore, the power supply control algorithm module includes a current and voltage dual closed-loop PID algorithm, a feedforward algorithm, and a dynamic process fine PID parameter adjustment algorithm; the parameter configuration in the current and voltage dual closed-loop PID algorithm is obtained based on a table lookup, and the configured parameter values ​​are different under different energy steps; in this way, the PID parameters configured in the high-energy reduction step and the low-energy reduction step within the entire energy range are the most suitable, so that the energy reduction speed, stability, and accuracy of each energy reduction step can all reach the optimum; the dynamic process fine PID parameter adjustment algorithm further subdivides the dynamic process of magnetic field change corresponding to each energy step into multiple refined dynamic process segments, and adjusts the response speed and response state of each refined dynamic process segment, thereby further improving the response speed, stability, etc. of each energy step.

[0089] Furthermore, the current-voltage dual closed-loop PID algorithm is provided with two closed-loop control loops: the first loop is a current loop, and the feedback quantity and control object of the current loop are the output current; the second loop is a voltage loop, and the feedback quantity and control object of the voltage loop are the load voltage, and the output quantity of the loop is a PWM wave with a variable duty cycle, which drives the H-bridge in the main circuit of the power supply, thereby controlling the output current of the power supply; the second loop is nested in the first loop, and the output signal of the first loop is used as the input quantity of the second loop; the first and second loops both contain PID control links, which function to improve the dynamic performance of the power supply and the system stability, and speed up the response speed; the feedforward algorithm refers to adding a feedforward link in each of the current loop and the voltage loop, which works together with the PID controller to achieve greatly accelerated response speed and compensate for the lag and damping of the controlled system on the basis of ensuring system stability and accuracy.

[0090] Furthermore, the feedforward link input of the current loop is taken from the current setting value of the input current loop; the feedforward link input of the voltage loop is taken from the voltage input of the voltage loop; and the feedforward link parameters are obtained from the "energy-current-PID and feedforward parameters-magnetic field response standard value" lookup table.

[0091] Furthermore, the dynamic process fine PID parameter adjustment algorithm refers to dividing the dynamic process of output current change into multiple sections, and automatically adjusting the PID parameters according to the magnetic field response of the load magnet; the specific method is: continuously monitoring the magnetic field value of the load magnet in each section, and automatically adjusting the PID parameters of the two control loops in each section according to the size of the gap between the magnetic field value and the standard value, so that the PID parameters can always match the dynamic process well, speed up the response speed and reduce overshoot, and eliminate the influence of hysteresis. The dynamic process is divided into multiple sections, at least three sections.

[0092] Embodiment 1

[0093] Take the beam line magnet power controller of CYCIAE-230 as an example

[0094] The present invention provides a digital control system for a proton therapy beam line magnet power supply, including a DDS communication / command module, a table lookup module, and a power supply control algorithm module. Figure 1-1 The figure shows the power supply of the different types of beam line magnets. Figure 1-1 Only 2 types of 10 magnets are shown, but this embodiment actually uses 5 types of 34 beam line magnets that need to use the power supply of this system to control.

[0095] 1. Hardware connection: The digital controller of each of the 34 power supplies is connected to the treatment head through a network cable. Each power supply sets itself and the treatment head in the same communication domain through the DDS communication / command module in the digital control system.

[0096] 2. Take the 60° deflection magnet in the process of reducing the beam energy from 237.71MeV to 236.816MeV as an example: the 60° deflection magnet power supply consists of a distributed architecture of one master power supply and three slave power supplies, and their models, rated voltages and current output performances are exactly the same;

[0097] A DDS command signal is sent from the treatment head, which contains the information of 236.816 MeV. Because the magnet power supplies are in the same communication domain as the treatment head, they receive the DDS command signal at the same time and obtain the information of the new beam energy value from it.

[0098] Subsequently, the DDS communication / command module of the digital control unit of the main power supply of the magnet sends this 236.816MeV to the table lookup module. The table lookup module uses the "energy-current-PID and feedforward parameters-magnetic field response standard value" lookup table as the basis to query the current setting value corresponding to this energy to be 436.8A, and the PID and feedforward parameters corresponding to this dynamic process are: current loop: proportional coefficient KP1=2.75, integral coefficient KI1=17.4e-05 and differential coefficient KD1=0, and the proportional coefficient KP2=2.1e-05, integral coefficient KI2=3e-05 and differential coefficient KD2=0.0003 of the voltage control loop; the current loop feedforward parameter is 0.9, and the voltage loop feedforward parameter is 0.55; the magnetic field response standard value under 236.816MeV is 1.6488T.

[0099] For different magnet power supplies, such as 30° deflection magnet power supply, 75° deflection magnet power supply, quadrupole magnet power supply, thin quadrupole magnet power supply, the above parameters will be different.

[0100] The "energy-current-PID and feedforward parameters-magnetic field response target value" of the beam line magnet power supply for CYCIAE-230 is obtained through joint debugging experiments, edited and stored in the FPGA of the control circuit in advance for future use. In the lookup table of different types of magnet power supplies, the current value, PID and feedforward parameters corresponding to the same beam energy, that is, the corresponding target value of the magnetic field, are different.

[0101] The main power supply divides the set current of 438.6A into four equal parts, each part is 109.65A, which is used as the set current of one main power supply and three slave power supplies respectively; the main power supply distributes the current setting value of 109.65A and the other parameters mentioned above to the three slave power supplies through the parameter forwarding module, and the four power supplies configure the PID link and feedforward link of the current loop and the PID link and feedforward link of the voltage loop in their respective dual-loop current control loops according to the above parameters, and control the output current with the current setting value of 109.65A as the standard; finally, the main power supply and the three slave power supplies are connected in parallel for output, and a current of 436.8A can be obtained.

[0102] In the dynamic process, in order to make the response faster, the dynamic process fine PID adjustment algorithm is used; the specific steps are as follows:

[0103] 1) Using the magnetic field response standard value 1.6488T obtained by table lookup and the magnetic field value collected in real time as reference, calculate the degree of completion of the dynamic process; the calculation formula is:

[0104] Completion degree of dynamic process (%) = (B magnetic field response standard value - B real-time magnetic field value) / (B magnetic field response standard value - B magnetic field initial value)

[0105] The initial value of the magnetic field in the formula is the magnetic field value before the dynamic response, which can be approximately replaced by the standard value of the magnetic field response of the previous dynamic process, 1.6523T;

[0106] 2) When the degree of completion of the dynamic process reaches 50% (i.e., when the magnetic field reaches 1.6505T), the proportional coefficients KP1 and KP2 in the two-loop PID controller parameters are increased by 20% each (i.e., KP1 = 3.3, KP2 = 2.1e-05);

[0107] 3) When the completion degree of the dynamic process reaches 80% (i.e., when the magnetic field reaches 1.6516T), the integral coefficients KI1 and KI2 in the two-loop PID controller parameters are increased by 20% (i.e., KI1 = 20.88e-05, KI2 = 3.6e-05);

[0108] The safety interlock system in the power digital control system in this embodiment includes: internal interlocking includes: (1) the power temperature is too high (over 40°C), the power leakage current is too large (greater than 0.1A), the output exceeds the rated value, and the energy information in the DDS instruction is incorrect (for example, not within the proton therapy energy range). When the above situation occurs, the digital controller should control the power supply to shut down immediately or keep the current state unchanged. (2) The external interlocking has 5 interfaces, one of which is used to connect the magnetic relay that detects whether the load magnet is overheated, and the remaining 4 are reserved for subsequent development; if one of the safety interlocks is triggered, the power supply will immediately stop output and shut down.

[0109] like Figure 2 The block diagram of the current output control program in this embodiment is shown. The two PID parameters and feedforward parameters of the dual loop are determined by experimental tests to ensure a good adaptation to the dynamic response of each energy segment. These values ​​are stored in the "energy-current-PID and feedforward parameters-magnetic field response standard value" lookup table.

[0110] like Figure 3 The figure shows a schematic block diagram of the hardware circuit composition of the digital controller in this embodiment. The ZYNQ chip is based on the ARM core and integrates the LINUX system to better control and supervise the output process, communication process, problem diagnosis process, etc. of the power supply. This also facilitates the development of the visual interface of the power supply.

[0111] The ZYNQ chip integrates DDS command response program and digital control algorithm, and stores the "energy-current-PID and feedforward parameters-magnetic field response standard value" lookup table. After completing the control of the current, the DSP is controlled to generate PWM waves, thereby driving the main circuit of the power supply and generating output current.

[0112] In this embodiment, to ensure high precision, the current feedback module adopts DCCT, and the precision of the AD conversion module is 24 bits.

[0113] The lookup table of the present invention includes PID parameters and feedforward parameters, so that the parameters in the dynamic process of different energy steps (for example, 237.71MeV is reduced to 236.816MeV and 71.73MeV is reduced to 70MeV) are different; this is to make the PID parameters configured in the high energy reduction step and the low energy reduction step in the entire energy range the most suitable, so that the energy reduction speed, stability, and accuracy of each energy reduction step can reach the most suitable;

[0114] The differences in current parameters, PID and feedforward parameters between power supplies for different types of magnets are due to the different types of magnets, which is essentially different from the different PID and feedforward parameters between different energy steps described in the previous paragraph.

[0115] like Figure 4-1The magnetic field response comparison diagram provided for an embodiment of the present invention takes the magnetic field response of a 60° magnet under a current of 307A as an example to illustrate the superiority of the method of this article. In the CYCIAE-230 proton therapy system, the dynamic response time of the magnetic field is required to be less than 80ms. If the method of the present invention is not used, that is, when a set of PID parameters are used for different energy segments, the dynamic response at a current of 307A will overshoot, resulting in a response time of up to 273.5ms. The use of the digital control method proposed in this article to improve the response can eliminate the overshoot and reduce the response time from 273.5ms to 50.67ms, meeting the requirement of less than 80ms. The superiority of the method of this article is demonstrated through comparison.

[0116] like Figure 4-2 The magnetic field response time comparison diagram of four types of magnets provided in one embodiment of the present invention shows that the fastest response time of each magnet is 27ms and the slowest is 67ms, all of which meet the requirement of less than 80ms. Therefore, it is shown that the patent method is applicable to all types of magnets and magnet power supplies.

[0117] It should be emphasized that the above specific embodiments are merely explanations of the present invention, and they are not limitations of the present invention. After reading this specification, those skilled in the art can make modifications to the above embodiments without any creative contribution as needed, but as long as they are within the scope of the claims of the present invention, they are protected by patent law.

Claims

1. A digital control unit for a proton therapy beam line magnet power supply, the digital control unit being respectively arranged in a magnet power supply control system of each of a plurality of magnets in the proton therapy beam line; the plurality of magnets being arranged at intervals along the starting point to the end point of the beam line of a proton therapy accelerator system, characterized in that: The magnet power supply of each magnet includes a magnet main power supply and a magnet slave power supply; the digital control unit includes a magnet main power supply digital control unit and a magnet slave power supply digital control unit; The magnet main power digital control unit includes a DDS communication / command module, a table lookup module, a current distribution and parameter forwarding module, and a power control algorithm module; the magnet slave power digital control unit includes a power control algorithm module, the input end of which is connected to the current distribution and parameter forwarding module of the magnet main power supply, and the output end is connected to the magnet power supply; The DDS communication / command module of the magnet main power supply has an input end externally connected to the treatment head system and an output end internally connected to the table lookup module; the DDS communication / command module enables the digital control unit of each magnet power supply to simultaneously subscribe to the DDS data published by the treatment head when they are in the same communication domain as the treatment head, and simultaneously obtain the required beam energy value from the DDS data; The table of the magnet main power supply lookup module is distinguished according to the type of magnets on the beam line. The same type of magnets is the same table. The table stores multiple sets of control parameter information of different energies for a certain type of magnet. Each energy corresponds to a set of magnet current control parameters. Each set of magnet current control parameters includes energy parameters, current parameters, PID parameters, feedforward parameters, and magnetic field response standard value parameters; wherein the energy parameters, PID parameters, feedforward parameters, and magnetic field response standard value parameters are used as common parameters of the magnet main power supply and the magnet slave power supply; the current parameters are allocated from the magnet main power supply to the magnet slave power supply; The table lookup module of the magnet main power supply performs a table lookup according to the parameters of "energy-current-PID and feedforward parameters-magnetic field response standard value" according to the beam energy information sent by the DDS communication / command module of each magnet main power supply digital control unit, obtains the corresponding current, PID and feedforward parameter setting values, and magnetic field response standard value, and sends the table lookup result to the current distribution and parameter forwarding module; The current distribution and parameter forwarding module of the magnet main power supply performs current distribution between the main power supply and the slave power supply, and then sends the current distribution result and the energy parameters, PID parameters, feedforward parameters, and magnetic field response standard value parameters obtained by table lookup to the magnet main power supply and the power supply control algorithm module of the magnet slave power supply, respectively, as the basis for each power supply control algorithm module to control the power supply.

2. A digital control unit for a proton therapy beam line magnet power supply according to claim 1, characterized in that: The table lookup is performed according to the parameters of "energy-current-PID and feedforward parameters-magnetic field response standard value", the energy parameter is the beam energy value / MeV, the current parameter is the magnet power supply output current value / A, and the magnet magnetic field is controlled by the power supply output current value / A; the PID parameters are the proportional coefficient KP1, integral coefficient KI1 and differential coefficient KD1 of the current control loop, and the proportional coefficient KP2, integral coefficient KI2 and differential coefficient KD2 of the voltage control loop; the feedforward parameters are the feedforward link parameters of the current control loop and the feedforward link parameters of the voltage control loop; the PID current loop proportional coefficient KP1, integral coefficient KI1 and differential coefficient KD1 are used to control the dynamic energy reduction or energy increase process for the current energy within a specified time, and the specified time includes 80 milliseconds; the PID voltage loop proportional coefficient KP2, integral coefficient KI2 and differential coefficient KD2 are used to control the dynamic energy reduction or energy increase process for the current energy within a specified time, and the specified time includes 80 milliseconds.

3. A digital control unit for a proton therapy beam line magnet power supply according to claim 1, characterized in that: The power supply control algorithm module includes a current and voltage dual closed-loop PID algorithm, a feedforward algorithm, and a dynamic process fine PID parameter adjustment algorithm; the parameter configuration in the current and voltage dual closed-loop PID algorithm is obtained by table lookup, and the configured parameter values ​​are different under different energy steps; in this way, the PID parameters configured in the high energy reduction step and the low energy reduction step within the entire energy range are the most suitable, so that the energy reduction speed, stability, and accuracy of each energy reduction step can reach the optimum; the dynamic process fine PID parameter adjustment algorithm subdivides the dynamic process of magnetic field change corresponding to each energy step into multiple refined dynamic process segments, and adjusts the response speed and response state of each refined dynamic process segment, thereby further improving the response speed and stability of each energy step.

4. A digital control unit for a proton therapy beam line magnet power supply according to claim 3, characterized in that: The current-voltage dual closed-loop PID algorithm is provided with two closed-loop control loops: the first loop is a current loop, and the feedback quantity and control object of the current loop are the output current; the second loop is a voltage loop, and the feedback quantity and control object of the voltage loop are the load voltage, and the output quantity of the loop is a PWM wave with a variable duty cycle, which drives the H-bridge in the main circuit of the power supply, and then controls the output current of the power supply; the second loop is nested in the first loop, and the output signal of the first loop is used as the input quantity of the second loop; the first and second loops both contain PID control links, which function to improve the dynamic performance of the power supply and the system stability and speed up the response speed; the feedforward algorithm refers to adding a feedforward link in each of the current loop and the voltage loop, and working together with the PID controller to achieve greatly accelerated response speed and compensate for the hysteresis and damping of the controlled system on the basis of ensuring system stability and accuracy.

5. A digital control unit for a proton therapy beam line magnet power supply according to claim 4, characterized in that: The feedforward link input of the current loop is taken from the current setting value of the input current loop; the feedforward link input of the voltage loop is taken from the voltage input of the voltage loop; the feedforward link parameters are obtained from the "energy-current-PID and feedforward parameters-magnetic field response standard value" lookup table.

6. A digital control unit for a proton therapy beam line magnet power supply according to claim 3, characterized in that: The dynamic process fine PID parameter adjustment algorithm refers to dividing the dynamic process of output current change into multiple sections, and automatically adjusting the PID parameters according to the magnetic field response of the load magnet; the specific method is: continuously monitoring the magnetic field value of the load magnet in each section, and automatically adjusting the PID parameters of the two control loops in each section according to the size of the gap between the magnetic field value and the standard value, so that the PID parameters can always match the dynamic process well, speed up the response speed and reduce overshoot, and eliminate the influence of hysteresis. The dynamic process is divided into multiple sections, at least three sections.

7. A digital control unit for a proton therapy beam line magnet power supply according to claim 1, characterized in that: The power supply control algorithm module also includes a safety interlock logic algorithm; the safety interlock logic algorithm is divided into internal interlock and external interlock. The internal interlock logic means that the controller continuously monitors certain indicators inside the power supply. When they are abnormal, the digital controller can immediately detect and control the power supply to shut down or keep the current state unchanged. The external interlock logic means that when an indicator outside the power supply is abnormal, the digital controller can immediately detect and control the power supply to shut down or keep the current state unchanged.

8. A digital control unit for a proton therapy beam line magnet power supply according to claim 1, characterized in that: Each magnet power supply control system includes FPGA, DSP, power supply main circuit, current acquisition module ADC, signal amplification module, and AD conversion module. The digital control unit is arranged in FPGA. The FPGA is a ZYNQ chip. The ZYNQ chip can be equipped with a digital control method, and the LINUX system can be integrated in the ZYNQ to better control and supervise the output process, communication process, and problem diagnosis process of the power supply; the DSP is controlled by the ZYNQ chip and can generate a PWM wave for driving the H bridge in the main circuit of the power supply according to the output of the digital controller; the current acquisition module ADC, signal amplification module and AD conversion module are used to collect the output of the power supply on the load side, and amplify and convert the collected analog signal into a digital signal, and feed it back to the digital controller.

9. A digital control unit for a proton therapy beam line magnet power supply according to claim 1, characterized in that: The DDS communication / command module includes a communication domain setting module, a communication domain subscription module, and an instruction acquisition module; the communication domain setting module is used to set its own magnet power supply and the current treatment head in the same communication domain; the communication domain subscription module is used to subscribe to the instruction message issued by the current communication domain; the instruction acquisition module is used to obtain the required beam energy value from the instruction.

Images