An adaptive dynamic compensation chlorine-hydrogen ratio control method

Abstract

An adaptive dynamic compensation chlorine-hydrogen ratio control method belongs to the field of automatic control technology, and includes the following steps: step S01: obtaining the operating characteristics of chlorine and hydrogen loop valves and establishing a mathematical model; step S02: using P intelligence The method designs PID parameters for the model; step S03: dynamic correction of closed-loop deviation of hydrogen and chlorine flow; step S04: coordinated adaptive optimization of chlorine and hydrogen flow ratio. 2 , Cl 2 The ratio of synthetic HCL is optimized and coordinated to ensure that the real-time ratio meets the requirements of trace hydrogen and protects safe production.

Description

Technical Field

[0001] The invention relates to the field of automatic control technology, and in particular to an adaptive dynamic compensation chlorine-hydrogen ratio control method. Background Art

[0002] The main raw materials for the synthesis of vinyl chloride monomer are acetylene and HCL, among which HCL is commonly used Cl 2 , H 2 According to the combustion synthesis of a specific molar ratio, an excess of Cl 2 , H 2 It has an adverse effect on production. 2 The free chlorine enters the mixer and reacts with acetylene to form a highly explosive substance, causing a serious safety accident. 2 It will cause the purity of hydrogen chloride to decrease and even affect the yield of vinyl chloride. If there is too much excess, it may form an explosive mixture, which is not conducive to safe production.

[0003] In the control of the chlorine-hydrogen ratio, due to the importance and particularity of the process, the control system is usually unable to be put into automatic control state, and there are many interference factors that make it impossible for the chlorine-hydrogen ratio to meet the ideal ratio constraint control in real time. The labor intensity of operators is high and it is not conducive to improving the enterprise's self-control rate. Summary of the invention

[0004] In order to improve the self-control rate of the chlorine-hydrogen ratio section, reduce the operation frequency, and improve the real-time ratio control accuracy of chlorine and hydrogen, the present invention provides an adaptive dynamic compensation chlorine-hydrogen ratio control method, which effectively improves the self-control rate and meets the expected real-time ratio of chlorine and hydrogen when the fluctuation is within a small range.

[0005] In order to achieve the above technical objectives, the technical solution provided by the present invention is as follows:

[0006] An adaptive dynamic compensation chlorine-hydrogen ratio control method comprises the following steps:

[0007] Step S01: Obtaining the operating characteristics of the chlorine and hydrogen loop valves and establishing a mathematical model;

[0008] Step S02: Using P intelligence Methods PID parameters are designed for the model;

[0009] Step S03: Dynamic correction of hydrogen flow closed loop deviation;

[0010] Step S04: Coordinated adaptive optimization of the flow ratio of chlorine and hydrogen.

[0011] Furthermore, the step S01 includes the following steps:

[0012] S101: Extract steady-state samples of chlorine and hydrogen loop operation data. Determine the steady-state operation of the data and determine the sampling length: The collected data set is the loop SV, PV, and MV information. A sample data set of appropriate length N (for example, N is 3600) is selected from the data set. The method of moving sample points in a fixed interval is used to select data points of normal operation of the loop. The fixed window calculation sample length is n, and its average value is calculated. Calculate its standard deviation n std , the stability threshold is defined as: δ is the expected threshold coefficient, when When , it is considered to be stable working condition characteristic data, the data length N is traversed, the data sequence information is screened and saved; the MV is determined to be unsaturated, and the PV responds dynamically with the MV: the data is normalized based on the MV data and the variance D is solved MV , define D yz is the variance threshold, when D MV ≥D yz The MV is considered to be unsaturated, and the data sequence information is filtered and saved.

[0013] S102: Data high-frequency noise filtering. Filtering condition judgment: Determine the data sampling period of the current loop as T sample , control period T control , take steady-state data, at 3T control If the PV value fluctuates frequently within the time interval, it is most likely caused by high-frequency noise and needs to be filtered. The high-frequency noise is filtered using a filtering algorithm to screen and save the data sequence information.

[0014] First-order inertial filtering:

[0015] S(k)=βS(k)+(1-β)S(k-1) (1-1)

[0016] β is based on the oscillation period T zd Given, the recommended β is:

[0017]

[0018] S103: Establish an approximate model. The ARX model is established for the processed data using the least square method and converted into formula (1-3) for parameter setting.

[0019]

[0020] Furthermore, the step S02 includes the following steps:

[0021] S201:P intelligence The simplified block diagram of the process is shown in Figure 1 , the closed-loop transfer function of the controlled loop is expressed as:

[0022]

[0023] In order to obtain an independent expression of the controller K(s), mathematical transformation is performed on equation (2-1):

[0024]

[0025] Simplified to:

[0026]

[0027] Ideal dynamic characteristics closed-loop transfer function:

[0028]

[0029] Substituting (2-4) into (2-3) we can get the desired controller expression corresponding to the desired ideal dynamic characteristics:

[0030]

[0031] S202: PID controller derivation. Derived from formula (2-3) P intelligence Controller setting method, substitute formula (2-3) into (2-5):

[0032]

[0033] For -λs Using mathematical methods Taylor approximation:

[0034]

[0035] Substitute formula (2-7) into (2-6):

[0036]

[0037] P intelligence The recommended controllers are:

[0038]

[0039] Recommended primary variable circuit P intelligence =3λ, recommended secondary variable loop P intelligence =λ.

[0040] Furthermore, the step S03 includes the following steps:

[0041] S301: PID parameter setting for closed-loop flow circuit. Taking the recommended values ​​of S202 as the reference PID controller parameters, the recommended parameters for the chlorine and hydrogen circuits are:

[0042]

[0043] S302: Hydrogen closed-loop flow circuit interference feature judgment. Definition: The operating setting value is: The operating measurements are: The operating actuator opening is: The scanning cycle is: The expected threshold value Dis is determined based on the scan cycle length. desire The average value of the sample points within the scanning period The relationship is divided into characteristic one and characteristic two:

[0044]

[0045] Feature one: It is believed that the PID controller has a weak forward suppression effect on disturbances.

[0046] Feature two: It is believed that the PID controller has a weak reverse suppression effect on disturbance.

[0047] Based on the hydrogen flow setting value, It is considered as normal fluctuation, and the fluctuation beyond that is considered as interference fluctuation. scan Count the number of interference fluctuation peaks and troughs N within the period disturbance , and the expected fluctuation number threshold N disturbance_desire contrast:

[0048] Feature 3: N disturbance ≥N disturbance_desire The interference is considered to be greater.

[0049] The three feature detections are analyzed using the time window shift method.

[0050] S303: Hydrogen closed-loop flow rate circuit interference suppression: Interference suppression is performed according to the interference characteristics of step S302.

[0051] Characteristic 1 Interference Suppression: Due to the hydrogen flow loop gain is positive, the hydrogen valve opening is by It is a periodic callback, which is repeated according to the time window shift rule until in:

[0052]

[0053] Characteristic 2 Interference Suppression: Due to the hydrogen flow loop gain is positive, the hydrogen valve opening is by It is a periodic increase, and it is repeated according to the time window shift rule until in:

[0054]

[0055] Characteristic 3 Interference suppression: This type of interference may be caused by temporary mechanical failure of the actuator, resulting in self-oscillation, fixing the valve opening to the commonly used opening with a period of 6T scan , the moving time window continues to judge until the anomaly is eliminated.

[0056] Furthermore, the step S04 includes the following steps:

[0057] S401: Get real-time operating measurement values ​​of the chlorine gas circuit Real-time operating measurements of the hydrogen circuit Define the desired hydrogen-chlorine ratio R_desire, the real-time operating hydrogen-chlorine ratio R_run, the desired operating hydrogen-chlorine ratio R_desire_run, and the ratio adjustment margin R_margin.

[0058]

[0059]

[0060] R_desire_run_down=R_desire-R_margin (4-3)

[0061] R_desire_run_up=R_desire+R_margin (4-4)

[0062] R_desire_run_down≤R_desire_run≤R_desire_run_up (4-5)

[0063] S402: Coordinated optimization control of chlorine and hydrogen. Optimization is divided into the following situations:

[0064] Case1: R_run<R_desire_run_down

[0065] Decrease by the mean until

[0066] R_run<R_desire_run_down.

[0067] Case2: R_run>R_desire_run_up

[0068] Increase by the mean until

[0069] R_run<R_desire_run_down.

[0070] Case 3:

[0071] It is considered to be in the stable range and no process intervention is required.

[0072] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows:

[0073] (1) The process is modeled using valve characteristics and corresponding output responses, and the preprocessed data of historical data is used as the modeling data set, which does not interfere with production.

[0074] (2) Use P based on the expected closed-loop characteristics to build the model intelligence The method designs PID parameters of the model and is reliable.

[0075] (3) The model gain is used to suppress interference in the hydrogen loop to ensure the stability of the main flow loop, which is beneficial to the coordinated control of the secondary flow.

[0076] (4) Coordinated optimization ensures that the real-time ratio of chlorine to hydrogen meets the desired operating ratio range. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 The system structure block diagram of step S201 of the present invention;

[0078] Figure 2 It is a schematic flow chart of the adaptive dynamic compensation chlorine-hydrogen ratio control method of the present invention;

[0079] Figure 3 A schematic diagram of the process of obtaining the operating characteristics of the chlorine and hydrogen loop valves and establishing a mathematical model for the present invention;

[0080] Figure 4 For the use of the present invention intelligence Methods: PID parameter design flow chart for the model;

[0081] Figure 5 It is a schematic diagram of a hydrogen flow closed-loop circuit deviation dynamic correction process of the present invention;

[0082] Figure 6 The figure is a schematic diagram of the coordinated adaptive optimization process of the chlorine and hydrogen flow ratios of the present invention. DETAILED DESCRIPTION

[0083] The specific implementation process of the present invention is described in detail below in conjunction with the accompanying drawings of the present invention. Obviously, the described implementation process is only a part of the examples of the present invention, rather than all the examples.

[0084] The present invention provides an adaptive dynamic compensation chlorine-hydrogen ratio control method, which uses pre-processed data of historical data as a data set for modeling, and uses a method based on expected closed-loop constants to design PID parameters for the model. The PID parameters can be set according to process object characteristics and expected response speed, and interference suppression is performed on the hydrogen loop to ensure the stability of the main flow loop, which is beneficial to the coordinated control of the secondary flow. At the same time, coordinated optimization is performed to make the real-time ratio of chlorine and hydrogen meet the expected operating ratio range.

[0085] Example:

[0086] An adaptive dynamic compensation chlorine-hydrogen ratio control method, such as Figure 2 As shown, proceed as follows.

[0087] Step S01: Obtain the operating characteristics of the chlorine and hydrogen loop valves and establish a mathematical model.

[0088] Furthermore, in step S01, the specific contents to be calculated include:

[0089] Step S101: Obtain historical SV, MV, and PV data of the chlorine and hydrogen loop in the DCS system and perform normalization, steady-state data judgment, valve non-saturation judgment, and extract the original data sequence. For example, obtain one hour of historical data of the hydrogen loop:

[0090] H 2_PV:[…129.74130 129.61 129.1 128.38 127.29 126.07 124.94 124.06122.34 121.01 120.31 119.66 118.98 118.6 118.1 117.81 117.97 118.5 118.94119.17 119.78 120.18 120.47 120.77 120.96 121.18 121.51 121.39 121.53 121.8121.79 121.88 121.74 122.06 122.11 121.5 121.29 120.66 120.67 120.51 120.38120.33 120.54 120.33 120.2 120.26 120.41 120.09 119.89 119.82 119.2 119.41119.1 119.3 119.1 119.23 119.45 119.47 119.66 119.7 119.89 119.71 119.58119.91 119.77 119.8 119.39 119.25 119.06 118.96 119.12 119.29 119.53 119.56119.7 119.5 119.44 119.74 119.92 119.96 119.61 119.62 119.66 119.5 119.46119.92 119.7 119.24 118.54 118.3 118.08 117.73 118.08 118.24 118.4 118.58118.04 118.59 118.75…]

[0091] H 2_MV: […1.92 2.76 3.19 3.60 4.04 4.28 4.61 4.80 4.69 4.34 4.04 3.883.47 3.18 2.97 2.74 2.59 2.42 2.17 2.21 2.09 1.87 1.84 1.75 1.80 1.55 1.481.84 1.95 2.34 2.31 2.39 2.45 2.46 2.30 2.41 2.47 2.41 2.29 2.48 2.59 2.623.02 2.87 3.06 2.91 3.04 2.94 2.78 2.75 2.61 2.57 2.42 2.53 2.60 2.36 2.44 2.40 2.65 2.73 2.85 2.91 2.79 2.66 2.50 2.46 2.35 2.46 2.49 2.28 2.14 2.09 2.31 2.29 2.24 2.34 2.35 2.03 2.16 2.44 2.89 3.05 3.19 3.43 3.20 3.09 2.98 2.86 3.21 2.84 2.73 2.78 3.02 2.96 2.58 2.45 2.30 2.30 2.09 2.26…]

[0092] After normalizing the data, the fixed interval sample point moving method is used to calculate H 2 _PV is used to determine the steady state of the data, where the parameters are set as: n = 10, δ = 0.03; the valve saturation is determined and the unsaturated data set is extracted, where the parameters are set as: D yz =0.01.

[0093] Step S102: Filter the data information of step S101 corresponding to the original data sequence and perform filtering: for example, T sample =1,T control =3,3T control =9, T zd =3.

[0094] but:

[0095] The processed data is:

[0096] H 2_PV_lv:[…129.75 129.74 129.72 129.67 129.57 129.43 129.25 129.04128.78 128.46 128.14 127.80 127.45 127.09 126.73 126.38 126.04 125.74 125.47125.21 125.00 124.80 124.63 124.48 124.34 124.21 124.10 123.99 123.89 123.81123.73 123.66 123.58 123.52 123.46 123.38 123.30 123.19 123.09 122.99 122.89122.78 122.69 122.60 122.50 122.41 122.33 122.24 122.15 122.06 121.94 121.84121.73 121.63 121.53 121.44 121.36 121.29 121.22 121.16 121.11 121.05 120.99120.95 120.90 120.86 120.80 120.74 120.67 120.60 120.54 120.49 120.45 120.42120.39 120.35 120.32 120.29 120.28 120.27 120.24 120.22 120.19 120.17 120.14120.13 120.11 120.08 120.02 119.95 119.87 119.79 119.72 119.66 119.61 119.57119.51 119.47 119.44 119.41…]

[0097] H 2_MV_lv: […1.92 2.76 3.19 3.60 4.04 4.28 4.61 4.80 4.69 4.34 4.043.88 3.47 3.18 2.97 2.74 2.59 2.42 2.17 2.21 2.09 1.87 1.84 1.75 1.80 1.551.48 1.84 1.95 2.34 2.31 2.39 2.45 2.46 2.30 2.41 2.47 2.41 2.29 2.48 2.592.62 3.02 2.87 3.06 2.91 3.04 2.94 2.78 2.75 2.61 2.57 2.42 2.53 2.60 2.36 2.44 2.40 2.65 2.73 2.85 2.91 2.79 2.66 2.50 2.46 2.35 2.46 2.49 2.28 2.14 2.09 2.31 2.29 2.24 2.34 2.35 2.03 2.16 2.44 2.89 3.05 3.19 3.43 3.20 3.09 2.98 2.86 3.21 2.84 2.73 2.78 3.02 2.96 2.58 2.45 2.30 2.30 2.09 2.26…]

[0098] Step S103: The data processed in step S102 is used to establish an ARX model using the least squares method and converted into formula (1-3) for parameter adjustment.

[0099] The first-order lag transfer function model of the hydrogen loop is obtained as follows:

[0100] The first-order lag transfer function model of the chlorine gas loop is obtained as follows:

[0101] Step S02: Using P intelligence Method to design PID parameters of the model

[0102] Furthermore, the specific contents to be calculated in step S02 include:

[0103] Step S201: Loop P intelligence Derivation. For the hydrogen circuit, according to formula (2-5) and

[0104] For the chlorine gas circuit, according to formula (2-5) and

[0105]

[0106] Step S202: PID controller derivation. For the hydrogen loop, according to formula (2-9):

[0107]

[0108] For the chlorine gas circuit, according to formula (2-9):

[0109]

[0110] Step S03: Dynamic correction of hydrogen flow closed-loop circuit deviation.

[0111] Furthermore, the specific contents to be calculated in step S03 include:

[0112] Step S301: Setting PID parameters of closed-loop flow circuit. Taking the recommended values ​​in S202 as the reference PID controller parameters, the recommended parameters for the chlorine and hydrogen circuits are:

[0113] Hydrogen circuit:

[0114] Chlorine gas circuit:

[0115] Step S302: Determine the interference characteristics of the hydrogen closed-loop flow loop. For example, set the parameters as: Dis desire_up =0.01. The measured value of the hydrogen loop operation during the scanning period is:

[0116]

[0117] but:

[0118] Step S303: Hydrogen closed-loop flow circuit interference suppression. Interference suppression is performed based on the interference characteristics of step 302. The hydrogen valve opening is by It is a periodic callback, which is repeated according to the time window shift rule until in:

[0119]

[0120] Step S04: Coordinated adaptive optimization of the flow ratio of chlorine and hydrogen.

[0121] Furthermore, the specific contents to be calculated in step S04 include:

[0122] Step S401: Optimize the chlorine loop to follow the hydrogen loop. For example, obtain the real-time operating measurement value of the chlorine loop Real-time operating measurements of the hydrogen circuit The desired hydrogen-chlorine ratio R_desire=1.2, the real-time operating hydrogen-chlorine ratio R_run=1.1, the desired operating hydrogen-chlorine ratio 1.15≤R_desire_run≤1.25, and the ratio adjustment margin R_margin=0.05. Among them:

[0123] R_desire_run_down=R_desire-R_margin=1.15

[0124] R_desire_run_up=R_desire+R_margin=1.25

[0125] Step S402: Coordinated optimization control of chlorine and hydrogen.

[0126] R_run=1.1<R_desire_run_down=1.15, then:

[0127] Decrease by the mean until:

[0128] 1.15≤R_desire_run≤1.25.

[0129] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may alter, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. An adaptive dynamic compensation chlorine-hydrogen ratio control method, characterized in that: The steps include: Step S01: Obtaining the operating characteristics of the chlorine and hydrogen loop valves and establishing a mathematical model; Step S02: Using P intelligence Methods PID parameters are designed for the model; The step S02 comprises the following steps: S201:P intelligence Method derivation; the closed-loop transfer function of the controlled loop is expressed as: In order to obtain an independent expression of the controller K(s), mathematical transformation is performed on equation (2-1): Simplified to: Ideal dynamic characteristics closed-loop transfer function: Substituting (2-4) into (2-3) we can get the desired controller expression corresponding to the desired ideal dynamic characteristics: S202: PID controller derivation; derive P from formula (2-3) intelligence Controller setting method, substitute formula (2-3) into (2-5): For -λs Using mathematical methods Taylor approximation: Substitute formula (2-7) into (2-6): P intelligence The controller is: Main variable loop P intelligence =3λ, sub-variable loop P intelligence =λ; Step S03: Dynamic correction of hydrogen flow closed loop deviation; Step S04: Coordinated adaptive optimization of the flow ratio of chlorine and hydrogen.

2. The adaptive dynamic compensation chlorine-hydrogen ratio control method according to claim 1 is characterized in that: The step S01 comprises the following steps: S101: Extract steady-state samples of chlorine and hydrogen loop operation data; extract steady-state operation data and determine sampling length: the collected data set is the loop SV, PV, and MV information, and a sample data set of length N is selected from the data set; the method of moving sample points in a fixed interval is used to select data points of normal operation of the loop, and the fixed window calculation sample length is n, and its average value is calculated Calculate its standard deviation n std , the stability threshold is defined as: δ is the expected threshold coefficient, when When , it is considered to be stable working condition characteristic data, the data length N is traversed, the data sequence information is screened and saved; the MV is determined to be unsaturated, and the PV responds dynamically with the MV: the data is normalized based on the MV data and the variance D is solved MV , define D yz is the variance threshold, when D MV ≥D yz The MV is considered to be unsaturated, and the data sequence information is filtered and saved; S102: Data high-frequency noise filtering; Filtering condition judgment: Determine the data sampling period of the current loop as T sample , control period T control , take steady-state data, at 3T control If the PV value fluctuates frequently within the time interval, filtering is required. The filtering algorithm is used to filter the high-frequency noise, screen and save the data sequence information; First-order inertial filtering: S(k)=βS(k)+(1-β)S(k-1) (1-1) β is based on the oscillation period T zd Given, β is: S103: Establish an approximate model; establish an ARX model using the least squares method for the processed data and convert it into formula (1-3) for parameter setting; 3. The adaptive dynamic compensation chlorine-hydrogen ratio control method according to claim 1 is characterized in that: The step S03 comprises the following steps: S301: PID parameter setting for closed-loop flow circuit; based on S202 value, PID controller parameters, chlorine and hydrogen circuit parameters are: S302: judging the interference characteristics of the hydrogen closed-loop flow circuit; Definition: The operating setpoints are: The operating measurements are: The operating actuator opening is: The scanning cycle is: The expected threshold value Dis is determined based on the scan cycle length. desire The relationship with the mean value Dis of the sample points within the scanning period is divided into characteristic 1 and characteristic 2: Feature one: It is believed that the PID controller has a weak forward suppression effect on disturbances; Feature two: It is believed that the PID controller has a weak reverse suppression effect on disturbance; Based on the hydrogen flow setting value, It is considered as normal fluctuation, and the fluctuation beyond that is considered as interference fluctuation. scan Count the number of interference fluctuation peaks and troughs N within the period disturbance , and the expected fluctuation number threshold N disturbance_desire contrast: Feature 3: N disturbance ≥N disturbance_desire It is considered that the interference is large; The three characteristic detections are analyzed using the time window shift method; S303: Hydrogen closed-loop flow circuit interference suppression: interference suppression according to the interference characteristics of step 302; Characteristic 1 Interference Suppression: Due to the hydrogen flow loop gain is positive, the hydrogen valve opening is by It is a periodic callback, which is repeated according to the time window shift rule until in: Characteristic 2 Interference Suppression: Due to the hydrogen flow loop gain is positive, the hydrogen valve opening is by It is a periodic increase, and it is repeated according to the time window shift rule until in: Characteristic 3 Interference suppression: When self-oscillation occurs, the valve opening is fixed to the commonly used opening with a period of 6T scan , the moving time window continues to judge until the anomaly is eliminated.

4. The adaptive dynamic compensation chlorine-hydrogen ratio control method according to claim 1 is characterized in that: The step S04 comprises the following steps: S401: Chlorine circuit optimization follows hydrogen circuit; obtain real-time operation measurement values ​​of chlorine circuit Real-time operating measurements of the hydrogen circuit Define the desired hydrogen-chlorine ratio R_desire, the real-time operating hydrogen-chlorine ratio R_run, the desired operating hydrogen-chlorine ratio R_desire_run, and the ratio adjustment margin R_margin; R_desire_run_down=R_desire-R_margin (4-3) R_desire_run_up=R_desire+R_margin (4-4) R_desire_run_down≤R_desire_run≤R_desire_run_up (4-5) S402: Coordinated optimization control of chlorine and hydrogen; optimization is divided into the following situations: Case1: R_run<R_desire_run_down Decrease by the mean until R_run<R_desire_run_down; Case2: R_run>R_desire_run_up Increase by the mean until R_run<R_desire_run_down; Case3: It is considered to be in the stable range and no process intervention is required.

Images