CN102621271A - Soft measuring method for ethylene combustion ratio in ethylene oxychlorination process of industrial device - Google Patents

Abstract

The invention provides a soft measurement method of ethylene combustion rate in the process of industrial ethylene oxychlorination reaction, which can be directly sent to the oxychlorination unit according to the flow rate and concentration of the inlet stream of the ethylene oxychlorination fluidized bed reactor Real-time data such as the stream flow and concentration of the unit, the average temperature in the reactor, and the volume of circulating air and exhaust air are used to calculate the side reaction that occurs in the oxychlorination process online, that is, the combustion rate of carbon dioxide generated by ethylene combustion. In this way, the online soft measurement of ethylene combustion rate in the process of ethylene oxychlorination reaction can be realized, which can provide guidance for the optimization of production operation parameters. The invention can realize the on-line soft measurement of the key material consumption in the ethylene oxychlorination reaction process under the condition of no tail gas analysis equipment, and has strong industrial applicability.

Description

Soft measurement method for ethylene combustion rate in ethylene oxychlorination process in industrial plant

technical field

The invention relates to an online soft measurement method for the combustion rate of side reaction ethylene in the fluidized bed reactor of the ethylene oxychlorination unit in the production process of polyvinyl chloride (PVC) monomer vinyl chloride (VCM) in an industrial device, and belongs to the key material consumption in the chemical production process In the field of online soft measurement. the

Background technique

The online calculation of key material consumption in the chemical production process is the basis for process adjustment and production operation parameter optimization. Its main task is to quantitatively understand the impact of the main relevant operating parameters on the key material consumption in the production process by examining the characteristics of the chemical production process, understand its laws and mechanisms, and establish a correlation model, so as to provide support for the optimization of production operation parameters and the transformation of production processes, etc. Therefore, it is of great significance to establish a model that can accurately describe the consumption of key materials in the chemical production process. the

Vinyl chloride is an important raw material for the synthesis of polyvinyl chloride (PVC), and most of the vinyl chloride is used to synthesize PVC. Industrialized production was first realized by German Farben in 1931. The early production process was mainly calcium carbide acetylene method. This process consumes a lot of energy, and the catalyst will pollute the environment. In the 1960s, the emergence of the ethylene balanced oxychlorination method greatly promoted the rapid development of the polyvinyl chloride industry. This process reduces the production cost and has less environmental pollution. Compared with the calcium carbide acetylene method, it has many advantages. Therefore, , The ethylene oxychlorination method quickly replaced the calcium carbide acetylene method as the main production method of vinyl chloride. the

At present, there are two main production processes of vinyl chloride in my country's vinyl chloride industry: calcium carbide acetylene method and ethylene oxychlorination method. The calcium carbide method produces PVC with high energy consumption and high pollution. The activity of the catalyst gradually decreases due to sublimation or poisoning, and must be replaced after deactivation. The deactivated waste mercury catalyst cannot be recycled, causing pollution to the environment. From the perspective of long-term development, the calcium carbide method will be at a disadvantage in market competition due to factors such as production costs and environmental protection. At the same time, with the continuous improvement of the ethylene method process, the development prospect of the ethylene method will be better. the

The ethylene oxychlorination process has been industrialized in my country for more than 30 years, but at present, my country has not yet formed a systematic independent research and development technology in this field, and there is still a big gap with foreign companies. Ethylene oxychlorination to dichloroethane is the key technology of ethylene oxychlorination process, which is only owned by several large companies at present, such as Japan Mitsui Topress Company, Germany Woodhurst Company, European Ethylene Company and so on. Therefore, the online calculation method of ethylene combustion rate in the ethylene oxychlorination process of this patent is established for the oxychlorination unit of the Mitsui Topress process. In the Mitsui Topress Oxychlorination Unit, the material circulation system is shown in Figure 1. the

The pure C2H4 raw material gas sent from outside the boundary area is first mixed with the circulating gas, preheated, then mixed with the HCl gas from the cracking unit, and then mixed with the heated HCl and O2, and enters the bottom of the oxychlorination reactor together. The hot material from the top of the reactor is sent to the dichloroethane quenching tower for cooling, and the cooling medium is dichloroethane/water circulating fluid. The function of the quenching tower is to remove a small amount of hydrogen chloride in the material and cool the material at the same time. A small amount of dichloroethane in the bottom material of the quenching tower is separated, and the remaining non-condensable gas is sent to the alkali washing tower. The tail gas from the direct chlorination unit is also sent to the alkali scrubber. The role of the alkali washing tower is to absorb the CO2 in the mixed gas with the circulating alkali solution. The bottom material of the alkali washing tower is sent to the dichloroethane mixed gas to neutralize the acidic dichloroethane, and the non-condensable gas is sent to the gas-liquid separator after passing through the condenser, and finally compressed by the compressor as a circulating gas and combined with fresh C2H4 raw material The gas is mixed into the oxychlorination reactor. The dichloroethane obtained from the condenser and the gas-liquid separator is sent to the dichloroethane mixer, and then sent to the decanter. After being washed and separated by pure water, it is sent to the dichloroethane storage tank. The dichloroethane is sent to the refining unit. the

The combustion of ethylene to generate carbon dioxide is carried out in the oxychlorination fluidized bed reactor. Since there is no measuring instrument for the flow rate and component concentration at the reactor outlet, the ethylene combustion rate needs to be calculated by measuring several indicators in the oxychlorination cycle system. There are two main calculations here:

(1) The amount of carbon dioxide produced by the reaction. After leaving the reactor, most of the carbon dioxide is absorbed when it passes through the alkali washing tower, and the rest is evacuated to maintain the carbon dioxide concentration in the circulation system. the

(2) The amount of ethylene consumed by the reaction. It involves the amount of fresh ethylene imported from the reactor, the amount of ethylene in the tail gas sent from the direct chlorination unit and the amount of ethylene contained in the exhaust gas. the

The above method calculates the offline ethylene combustion rate, which is used as a training sample for the online soft sensor. By correlating the obtained off-line ethylene combustion rate with the production load, the feed ratio of C2H4/HCl, O2/HCl in the fluidized bed reactor and the average temperature in the reactor, the online ethylene combustion rate soft-sensing model was obtained. the

The results of soft measurement play an important role in the actual control of industrial production, and the results are required to be fed back to production control in time. Therefore, the online soft measurement of the ethylene combustion rate in the ethylene oxychlorination process and the real-time online calculation of the ethylene combustion rate of the oxychlorination unit based on the measurable operating parameters are the basis for the adjustment of the production operation parameters and the optimization of the operation parameters. the

Contents of the invention

The technical problem to be solved by the invention is to provide a soft measurement method for online calculation of the ethylene combustion rate in the ethylene oxychlorination unit in the VCM production process of the industrial device. Adopt neural network technology to set up the correlation model of production load, reactor feed ratio C2H4/HCl, O2/HCl and the average temperature in the reactor and the ethylene combustion rate obtained off-line; The purpose of the present invention is also: according to the measurable production load , the reactor feed ratio C2H4/HCl, O2/HCl and the average temperature in the reactor, and calculate the ethylene combustion rate in real time, so as to provide guidance for the adjustment of the production process and the optimization of the operating parameters. the

The present invention adopts the following technical solutions: the flow meter and concentration analyzer at the inlet of the fluidized bed reactor in the process of ethylene oxychlorination reaction, the flow meter and the concentration analyzer that are directly sent from the chlorination unit to the oxychlorination unit and concentration analyzer, temperature gauge inside the reactor, circulating air flowmeter, exhaust air flowmeter and the data acquisition system (including DCS and real-time database, etc.) , through the real-time and continuous acquisition of the input variables of the above model, the trained weights and thresholds are substituted and calculated to obtain the predicted value of the ethylene combustion rate. The neural network model is a typical BP model. the

First, a predictive model of the ethylene combustion rate needs to be established. In the process of ethylene oxychlorination reaction, there are four main factors affecting the ethylene combustion rate: production load, reactor feed ratio C2H4/HCl, O2/HCl and average temperature in the reactor. the

(1) The production load is determined by the hydrogen chloride flow of the reactor feed, and the hydrogen chloride flow is 9123Nm3/hr at 100% load. The production load ranges from 0 to 1. the

(2) Reactor Feed Ratio The ethylene flow in C2H4/HCl includes two parts, fresh ethylene feed and the amount of ethylene sent by the direct chlorination unit. the

(3) The oxygen in the reactor feed ratio O2/HCl is fresh oxygen feed. the

(4) The average temperature in the reactor. Since there are 33 horizontal baffles and 7 temperature measuring points inside the Mitsui Topress fluidized bed reactor, the average temperature is used as the operating basis in industrial production. the

The influence of the above-mentioned process operation parameters on the ethylene combustion rate is relatively complicated, and there is an interaction between each process parameter, and the relationship with the ethylene combustion rate is highly nonlinear. Therefore, the activation function is the SIGMOID function here, and a three-layer neural network model is established. , and use the error feedback algorithm to train the network. By collecting historical data information and using neural network technology to extract the information on the change of ethylene combustion rate, a neural network model of the main operating parameters and ethylene combustion rate YC2H4 in the oxychlorination process is established. the

The input variables of this neural network model are: production load (x1), reactor feed ratio C2H4/HCl (x2), O2/HCl (x3) and the average temperature in the reactor (x4, ℃), and use formula (1 ) for normalization:

${\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}^{\mathrm{<span\; class="notranslate">\; i</span>}}}{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; max</span>}}^{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; min</span>}}^{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1,2,3,4</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

表示采集到第i个操作参数的变化范围，归一化后输入自变量的变化范围为[0，1]。  In formula (1), xi is the actual measured value of the i-th operating parameter (ie independent variable), and sxi represents the normalized value of the i-th operating parameter as the input value of the neural network,

Indicates the variation range of the i-th operation parameter collected, and the variation range of the input independent variable after normalization is [0, 1].

For the collected n1 sets of data, each set of data contains [x1, x2, x3, x4, YC2H4], which is normalized to [sx1, sx2, sx3, sx4, YC2H4] to form a training sample; for ethylene combustion The neural network model of the rate, with [sx1, sx2, sx3, sx4] as the input of the network, and the corresponding ethylene combustion rate as the target value, train the network. When a certain accuracy requirement is reached, the training is stopped and the ethylene burning rate is obtained. the

In the neural network model, the production load, the reactor feed ratio C2H4/HCl, O2/HCl and the average temperature in the reactor are taken as independent variables, and the ethylene combustion rate is used as the output variable. The number of nodes in the input layer is 4, the number of nodes in the hidden layer is 4, and the number of nodes in the output layer is 1.

The invention can solve the calculation of the ethylene combustion rate when there is no analyzer at the outlet of the fluidized bed reactor. At the same time, since the influence relationship of each operating parameter on the ethylene combustion rate is established, the process operation parameters can be further optimized, and has the advantages of Strong industrial applicability. Although those skilled in the art can

The present invention is modified or equivalently replaced, but all technical solutions and improvements that do not deviate from the spirit and scope of the present invention shall be covered by the claims of the present invention. the

Description of drawings

Figure 1 is a flow chart of the oxychlorination reaction process of the Mitsui Topress process. the

Fig. 2 is a block diagram of a neural network model of the combustion rate of ethylene as a side reaction in the oxychlorination reactor. the

Detailed ways

Below in conjunction with accompanying drawing and by example the present invention will be further described:

Taking the ethylene oxychlorination reaction process of Mitsui Topress Technology as an example, the flow chart is shown in Figure 1. After the fresh feed and circulating gas are mixed in the gas mixer, they enter the bottom of the oxychlorination reactor. The oxychlorination reactor operates within a certain temperature range, and hydrogen chloride, ethylene, oxygen, etc. react to form dichloroethane (EDC for short). At the same time, the side reaction ethylene is burned to generate carbon dioxide. The gas from the reactor goes through the quenching tower to remove the HCl, and then goes through the alkali washing tower to mainly remove the CO2. The non-condensable gas from the EDC decanter is vented in a small amount, and most of it re-enters the reactor as a cycle gas. the

Figure 2 is a block diagram of the neural network model of the ethylene combustion rate of the reactor, selecting the production load (x1), the reactor feed ratio C2H4/HCl (x2), O2/HCl (x3) and the average temperature in the reactor (x4, ℃) As the independent variable of the model, each independent variable is normalized to form the input independent variable [sx1, sx2, sx3, sx4] of the network, and the range of the normalized variable is [0, 1]; the network output is Calculated value of ethylene combustion rate; in the neural network model, the number of nodes in the input layer is 4, the number of nodes in the hidden layer is 1 (1=2～25), and the number of nodes in the output layer is 1. the

Collect 1123 groups of reactors in the production process under different production loads (x1), reactor feed ratios C2H4/HCl (x2), O2/HCl (x3) and average temperature in the reactor (x4, ℃), corresponding to ethylene Sample data for burn rate formation. the

Use the formula (1) to normalize the above variables: the range of change of x1 [0.401, 0.907], the range of change of x2 [0.467, 0.649], the range of change of x3 [0.243, 0.317], the range of change of x4 [202.3, 234.3], forming standardized sample data. The standardized calculation formulas are as follows:

${\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.401</span>}}{\mathrm{<span\; class="notranslate">\; 0.907</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.401</span>}}$

${\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.467</span>}}{\mathrm{<span\; class="notranslate">\; 0.649</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.467</span>}}$

${\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.243</span>}}{\mathrm{<span\; class="notranslate">\; 0.317</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.243</span>}}$

${\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 202.3</span>}}{\mathrm{<span\; class="notranslate">\; 234.3</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 202.3</span>}}$

The network structure is as follows: the number of nodes in the input layer is 4, the number of nodes in the hidden layer is 4, and the number of nodes in the output layer is 1. Taking the standardized 1123 sets of sample data as training samples, the network is trained using the BP algorithm; when the network converges, Get the following set of weights and bias values:

${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 11</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2.086</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 4.510</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 8.946</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 41</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 7.908</span>}$

${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 12</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1.835</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; two</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 5.710</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 32</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2.615</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 42</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0.158</span>}$

${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 13</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.090</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; three</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0.811</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 33</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3.842</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 43</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0.357</span>}$

${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 14</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3.040</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; four</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.668</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 34</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 15.521</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 44</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1.764</span>}$

${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 11</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.0021</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.0031</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.0134</span>}$ ${\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 41</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 0.0066</span>}$

${\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 3.503</span>}$ ${\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1.609</span>}$ ${\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1.893</span>}$ ${\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 11.087</span>}$

${\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0.0132</span>}$

为输入层第i个节点到隐含层第j个节点的权值； 

为隐含层第i个节点到输出层第j个节点的权值； 

为隐含层第i个节点偏置项的值； 

为输出层第i个节点偏置项的值。  in

is the weight of the i-th node in the input layer to the j-th node in the hidden layer;

is the weight of the i-th node in the hidden layer to the j-th node in the output layer;

is the value of the i-th node bias item in the hidden layer;

is the value of the i-th node bias item in the output layer.

The correlation model of ethylene combustion rate is:

$\mathrm{<span\; class="notranslate">\; net</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 11</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 41</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; net</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 12</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; two</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 32</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 42</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; net</span>}\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 13</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; three</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 33</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 43</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; net</span>}\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 14</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; four</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 34</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 3</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 44</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>{\mathrm{<span\; class="notranslate">\; sx</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; out</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; net</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; out</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; net</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; out</span>}\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; net</span>}\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; out</span>}\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; net</span>}\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 9</span>}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; net</span>}\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 11</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; out</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; out</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; out</span>}\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; w</span>}}_{\mathrm{<span\; class="notranslate">\; 41</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; out</span>}\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 10</span>}<span\; class="notranslate">\; )</span>$

${\mathrm{<span\; class="notranslate">\; Y</span>}}_{{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 4</span>}}}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; exp</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; net</span>}\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 11</span>}<span\; class="notranslate">\; )</span>$

Assuming that the reactor operation data is: production load (x1=0.871), reactor feed ratio C2H4/HCl (x2=0.507), O2/HCl (x3=0.285) and the average temperature in the reactor (x4=216.4 ° C), After normalization by formula (1), the standardized variables are substituted into formulas (2) to (11), and the ethylene combustion rate YC2H4=1.82% is calculated, and the relative error between the calculated value and the manually calculated value is within ±8%. . the

The above process describes the whole process of how to calculate the ethylene combustion rate inside the reactor online based on the real-time operation data of the reactor. the

Appendix

200 unit ethylene combustion rate calculation method

C2H4+O2-→CO+CO2

The design value of ethylene combustion rate is 1.5%

1. The amount of CO, CO2, C2H4 in the exhaust gas

1-1. Concentration of C2H4(A), CO(B), CO2(C) in circulating gas

1-2. Exhaust gas volume D Nm3/hr

1-3. Conversion of CO2 and CO into ethylene emissions (E)

E＝D×(B+C)/2/22.4kmol/hr

1-4. Ethylene emission F

F＝D×A/22.4kmol/hr

2. The amount of CO2 absorbed by lye

2-1. The total amount of material produced by the alkali washing tower G t/hr

Dichloroethane boiling point 0.1MPa 83.5℃

At 50°C, the specific gravity of dichloroethane and water are: pH2O=0.99, pC2H4Cl2=1.21

2-2. Alkaline washing tower wastewater volume H t/hr. the

H＝[X·pH2O/(X·pH2O+Y·pC2H4Cl2)]×G

X-height of water in the mixture of water and dichloroethane (mm)

The height of dichloroethane in Y-water and dichloroethane (mm)

The mixture of water and dichloroethane is sampled in a beaker to measure the respective heights. the

2-3. Calculation of the amount of Na2CO3 and NaHCO3 in the alkali washing tower

Na2CO3 concentration 1wt% molecular weight is 106

NaHCO concentration Jwt% molecular weight is 84

The amount of Na2CO3 converted into C2H4 K＝H·I×1000/106/2kmol/hr

The amount of NaHCO3 converted into C2H4 L=H·I×1000/84/2kmol/hr

3. The amount of ethylene added to the amount of ethylene in the oxychlorination unit

3-1. Feed amount of fresh ethylene FQIC-224Nm3/hr(M)

4. Exhaust gas volume from unit 100 to unit 200

4-1.100 C2H4(AA) concentration in unit tail gas

4-2.100 unit exhaust emission Q Nm3/hr

4-3.100 The amount of ethylene in the unit tail gas S

S＝Q×AA/22.4kmol/hr

Ethylene combustion rate = [(E+K+L)/(M/22.4+S-F)]×100%.

Claims (3)

1. the online soft sensor method of an industrial ethylene oxi-chlorination process ethylene combustion rate comprises following step:

Set up the ethylene combustion rate and produce load, fluidized-bed reactor charge ratio C ₂H ₄/ HCl, O ₂The artificial nerve network model of medial temperature in/HCl and the reactor;

Real time data in line computation Reactor inlet stream plume amount and concentration;

Deliver to the stream plume amount of oxychlorination unit and the real time data of concentration in line computation direct chlorination unit;

Medial temperature in the line computation reactor;

The real time data of online computation cycles tolerance and emptying tolerance;

The volume ratio of water and ethylene dichloride in the off-line measurement alkaline washing tower substrate material;

Calculated off-line ethylene combustion rate is brought in constant renewal in train samples.

In line computation ethylene combustion rate.

2. the online soft sensor method of industrial ethylene oxi-chlorination process ethylene combustion rate according to claim 1; It is characterized in that; The ethylene combustion rate of off-line can be obtained by the ethylene combustion rate computing method in the appendix, but the ethylene combustion rate that the method obtains has certain hysteresis quality.The ethylene combustion rate of utilizing off-line to obtain, C is compared in production load, reactor feed ₂H ₄/ HCl, O ₂The incidence relation of medial temperature is set up model with nerual network technique in/HCl and the reactor.The activation functions that the relational model of these process operation parameters and ethylene combustion rate adopts is the SIGMOID function, sets up three-layer neural network, and adopts the Error Feedback algorithm that network is trained.Obtaining neural network model is:

Y _C2H4＝f(x ₁，x ₂，x ₃，x ₄)

Through to above-mentioned model input variable in real time, continuous acquisition, with weights that train and threshold value substitution and calculate, obtain the predicted value of ethylene combustion rate.

3. the online soft sensor method of industrial ethylene oxi-chlorination process ethylene combustion rate according to claim 2 is characterized in that the input variable of said neural network model is: C is compared in production load, reactor feed ₂H ₄/ HCl, O ₂Medial temperature in/HCl and the reactor is carried out normalization with this neural network model and is handled.

${\mathrm{sx}}_i=\frac{x_i-x_{\min }^i}{x_{\max }^i-x_{\min }^i},i=\mathrm{1,2,3,4}$

Wherein, x _iBe the actual measured value of i operating parameter, sx _iRepresent after i the operating parameter normalization value as the neural network input,

Expression collects the variation range of i operating parameter, and the variation range of input independent variable is [0,1] after the normalization; To the n1 group data that collect, wherein every group of data comprise [x ₁, x ₂, x ₃, x ₄, Y _C2H4], after normalization [sx ₁, sx ₂, sx ₃, sx ₄, Y _C2H4], form training sample; To the neural network model of ethylene combustion rate, with [sx ₁, sx ₂, sx ₃, sx ₄] as the input of network, corresponding ethylene combustion rate is as desired value, training network.

Images