KR20030016370A - Process control for acetic acid manufacture - Google Patents

Abstract

The present application relates to a process for producing acetic acid from carbonylation of methanol by real time process control in a reaction system. Reaction system samples were transferred to reactor vessels 10 (120, 130, 140, 150, 160, 170, 180, 184, 186, 188, 190, 192, 194, 196, 197, 198, 199, 200, 210). , 220) collected downstream and / or from columns 30, 40, 50, 60, 70, 80, and the concentration of one or more components of this sample is measured by infrared analyzer 90. This concentration measurement is then performed, for example, in a particular column (30, 40, 50, 60, 70, 80) with a temperature profile, rate of flow of the solution into or out of the column (30, 40, 50, 60, 70, 80). , Component concentration in the reaction system, directly or indirectly, by adjusting the rate of exhaust gas out of the reactor 10 or columns 30, 40, 50, 60, 70, 80, or by adding components to the solution or adjusting component extraction from the solution. Used to adjust. For optimum process control, this measurement is transferred to the control unit for real time analysis, and this adjustment is made substantially immediately after infrared analysis.

Description

PROCESS CONTROL FOR ACETIC ACID MANUFACTURE

Cross-Reference to the Related Application

This application is filed on Jul. 6, 2000, which is part of a continuation of US patent application Ser. No. 09 / 216,330, now entitled Patent No. 6,103,934, filed on December 18, 1998, entitled "MANUFACTURING AND PROCESS CONTROL METHODS." The filed invention is a partial serial application of US patent application Ser. No. 09 / 611,067 entitled "MANUFACTURING AND PROCESS CONTROL METHODS."

A common method for preparing acetic acid involves the continuous reaction of methanol and carbon monoxide in a stirred reactor. The reaction mixture contains a soluble catalyst from Group 9, in particular iridium or rhodium, and a methyl iodide / hydrogen iodide accelerator to accelerate the reaction rate. Two main reactions that occur in the acetic acid process include the carbonylation of methanol to carbon monoxide to form acetic acid, and the water gas shift reaction to form carbon dioxide and hydrogen from carbon monoxide and water. When hydrogen is generated from the water gas displacement reaction, incidentally, propionic acid impurities are formed in the reactor solution.

As regards liquid phase acetic acid reactions, a complex equilibrium network thus exists in the reactor. Even the smallest changes in these equilibrium conditions can result in significantly poor catalyst stability and activity in the reactor. These changes may eventually result in a change of composition in the liquid stream entering the refinery of the acetic acid plant that implements the methanol carbonylation technique.

The use of on-line infrared analysis in controlling reactor liquid compositions is described in US Pat. Nos. 6,103,934 and 6,2000, entitled "MANUFACTURING AND PROCESS CONTROL METHODS". The invention is described in US Patent Application No. 09 / 611,067 entitled "MANUFACTURING AND PROCESS CONTROL METHODS", the entire contents of each of which are referenced herein. In order to achieve optimal reactor performance, the reactor solution can be analyzed in real time and immediate adjustments can be made through the process control loop. In the acetic acid preparation and purification process, of course, it is necessary to remove other components from the acetic acid product and, if necessary, these other components are returned to the reactor or other parts of the process via recycle loops. The composition of these purification / recycle streams is partly a function of reactor composition / performance and partly a function of recycle / purification column performance.

Thus, there is a need to perform process control through on-line infrared analysis of the purification and recycle sections of the acetic acid reaction system as in manufacturing facilities.

Summary of the Invention

The present invention provides a process for real time process control of component concentrations in a reaction system for producing acetic acid from carbonylation of methanol. To this end and in accordance with the present invention, a reaction system solution sample is collected from a delivery line downstream of the reactor vessel and / or from a column, and the concentration of one or more components in the sample is measured by an infrared analyzer. Concentration measurements are used to make adjustments in this process. The concentration of one or more components is adjusted directly or indirectly at one or more locations in the reaction system in accordance with downstream measurements. For example, the flow rate of the solution stream in the delivery line is increased or decreased to proceed into or out of the column, changing the concentration of one or more components in the column or another vessel in the reaction system. Alternatively, the temperature of the solution in the column or stream, or the temperature profile or slope in the column, is increased or decreased, affecting the concentration of one or more components in the reaction system solution. In addition, the concentration of the reaction system component can be adjusted directly by adding this component directly into the solution or extracting it directly from the solution. For example, the concentration of water in the reaction system can be adjusted directly by increasing or decreasing the water supplied into the reactor vessel, and indirectly by increasing or decreasing the hydrous recycle stream to the reactor. The gas velocity exiting the reactor vessel or column can also be increased or decreased. Thus, the reaction system component concentration can be adjusted directly or indirectly by changing any number of process variables in the reaction system. In addition, the adjustment at one location of the reaction system may result in a change in concentration at that location or either upstream or downstream of that location. For optimum process control, this measurement is passed to the control unit for real time analysis, the adjustment being made substantially immediately after the infrared analysis. Thus, there is provided a method of optimizing the preparation and purification of acetic acid products by continuously updating the state of the reaction system to improve real time process control of the overall process.

The present invention relates to a method for improving process control in acetic acid production and purification and to a process for producing acetic acid using this improved process control.

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and illustrate the principles of the invention, together with the foregoing summary and the following detailed description.

1 is a schematic diagram of an acetic acid production facility.

FIG. 2 is a correlation of the actual and predicted concentration values for acetaldehyde (CH ₃ CHO), indicating the effectiveness of a laboratory calibration model in the decanter heavy phase. FIG.

Figure 3 shows schematically one way of on-line analysis of the present invention in a decanter of a light ends recovery section.

FIG. 4A schematically illustrates one manner of on-line analysis of the present invention applicable to light oil column bottoms, drying column feeds, drying columns, heavy oil column feeds and product tank feeds. One drawing.

FIG. 4B schematically illustrates another manner of on-line analysis of the present invention applicable to light oil column bottoms, dry column feeds, dry columns, heavy ends column feeds and product tank feeds. FIG.

Figure 5 schematically illustrates another way of on-line analysis of the present invention in the decanter of the diesel oil recovery portion.

FIG. 6 shows a multi-component trend file containing 30 minutes run time data for five medium phase decanter solution components and solution density.

FIG. 7 shows a multicomponent trend file containing 60 minutes run time data for five medium phase decanter solution components and solution density. FIG.

FIG. 8 illustrates a multicomponent trend file containing 140 minutes of run time data for six medium phase decanter solution components. FIG.

FIG. 9 shows a multicomponent trend file containing four minute decanter solution components and 15 minutes run time data for solution density. FIG.

FIG. 10 shows a multicomponent trend file containing 80 minutes of run time data for four ordinary decanter solution components. FIG.

FIG. 11 shows a multicomponent trend file containing 125 minutes of run time data for five ordinary decanter solution components. FIG.

FIG. 12 shows a multicomponent trend file containing 45 minutes of run time data for five current decanter solution components and solution density. FIG.

FIG. 13 shows a multicomponent trend file containing 50 minutes of run time data for three light oil column bottom solution components. FIG.

FIG. 14 illustrates a multicomponent trend file containing 65 minutes of run time data for two product stream solution components. FIG.

FIG. 15 illustrates a multicomponent trend file containing 180 minutes of run time data for two heavy oil column or product stream solution components. FIG.

FIG. 16 illustrates a multicomponent trend file containing 75 minutes of run time data for three dry column feed components.

FIG. 17 illustrates a multicomponent trend file containing 85 minutes of run time data for two dry column solution components. FIG.

FIG. 18 illustrates a multicomponent trend file containing 110 minutes of run time data for two dry column components. FIG.

A. Acetic Acid Reaction System

As outlined in FIG. 1, the acetic acid manufacturing plant implementing the methanol carbonylation technique can be conveniently divided into three functional areas: reaction, light ends recovery and purification. For example, as described in U.S. Pat. The variety may vary in type, and such variations known to those skilled in the art are included within the scope of the present invention. In general, the reaction part consists of a reactor 10 and a flash tank 20. The diesel fuel recovery unit is composed of a diesel fuel column 30 and a phase separation vessel 40 (decanter). The purification section also consists of a dry oil column 30 as well as a light oil column 30 and optionally a heavy oil column 60. The various columns and vessels are connected by a conveying line, such as a pipe (reaction system solution typically flows through this pipe by a pump). For simplicity of illustration and description, the conveying line and the stream within the conveying line are all referred to herein as "streams".

It is fed to the reactor 10 via a stream 12 consisting of methanol, dimethyl ether, methyl acetate or mixtures thereof. Water may also be provided to stream 12. A mixture with carbon monoxide or an inert gas is also fed to this process via stream 12. During normal reactor 10 operation, the reactor contents are continuously recovered as a liquid. This is accomplished by flashing the reactor solution across a valve (not shown) to produce a vapor-liquid stream 100, which is delivered to a flash tank 20 where steam is separated from the liquid. The liquid containing the catalyst accumulates in the lower portion 22 of the flash tank 20, pumped by the catalyst recycle pump P1 and returned to the reactor 10 through the stream 110. The vapor stream 120 remaining in the top 24 of the flash tank 20 contains acetic acid product, water, methyl iodide (MeI), hydrogen iodide (HI) and low levels of impurities, ie propionic acid and acetaldehyde. This steam is fed to the diesel oil column 30. Therefore, the main purpose of the flash tank 20 is to separate the catalyst from the crude product and return this catalyst to the reactor 10.

The diesel oil column 30 is important in the overall process because it serves two purposes: to purify crude acetic acid and return the iodide to the reaction section. The gas oil column 30 is supplied from the flash tank 20 from the overhead vapor stream 120. This separates the high boiling acetic acid from the low boiling components such as MeI and methyl acetate (MeOAc). Three streams 130, 140, 150 are removed from the diesel fuel column 30. The overhead stream 130 consists mainly of MeI, but also contains some water, MeOAc and acetic acid and is passed to a phase separator, ie decanter 40, described below. The sidedraw stream 140 from the gas oil column 30 consists of wet acetic acid, which is delivered to the drying column 50 by a pump P5. The diesel oil bottoms stream 150 consists of water, HI and acetic acid, which is recycled to the reaction section including the reactor 10 and the flash tank 20.

The second part of the diesel oil recovery portion is a phase separator 40, more generally a decanter. A heavy phase 42 consisting primarily of MeI and an immiscible light phase 44 consisting primarily of aqueous acetic acid are separated in the decanter 40. The primary purpose of separating the light phase 44 and the middle phase 42 is to recycle MeI to the reactor 10. The MeI to be recycled is collected in a small boot 46 of the decanter 40, the volume of which is much smaller than the volume of the remaining upper portion 48 of the decanter 40 containing the mirror 44. MeI in boot 46 is recycled to reactor 10 via bottom stream 160 by pump P2. The secondary purpose of separating the phases is for returning through stream 170 by pump P4, a portion of which phase 44 is refluxed to gas oil column 30 and the remainder of this phase 44 Is recycled to reactor 10 via stream 180 by pump P3. The amount of incline directed to stream 170 as compared to stream 180 may be adjusted by one or both of pumps P3 and P4.

Many acetic acid processes include the additional treatment of the middle phase 42 and the minor phase 44. In processes such as those described in US Pat. Nos. 4,102,922, 5,371,286 and 5,599,976, the middle phase 42 is additionally treated to remove alkanes. One such process is shown in FIG. 1, in which the hydrocarbon removal column 70 receives a portion of the middle phase 42 through stream 210. Overhead of column 70 is returned to flash tank 20 via stream 220. The alkane rich bottoms are delivered to the waste via stream 230. Many acetic acid processes further process the middle phase 42 and the mirror phase 44 to remove impurities such as acetaldehyde and its condensation products. Examples of such acetaldehyde removal systems are described in US Pat. Nos. 5,599,976, 5,723,660, 5,625,095 and 5,783,731, EP Patents 487,284 and PCT Publication No. WO9817619. All acetic acid process streams can be analyzed using the present invention to improve overall process control as well as control of individual process steps.

The refining unit includes a dry oil column 30 and a heavy oil column 60 as well as the light oil column 30 as described above. Many acetic acid processes include an additional column 80 to strip acetic acid from high boiling impurities. Drying column 50 is a large distillation column that feeds itself and receives wet acetic acid stream 140 from gas oil column 30 through pump P5. As the name of this column implies, the main purpose of the drying column 50 is to remove water from the acetic acid product. Water is removed and fed to the column overhead drum 56 through the overhead stream 184 and returned to the reaction section through the stream 186 by a pump P8. A portion of the condensed solution in drum 56 is returned as column reflux through stream 188 by pump P9. In certain acetic acid processes, such as those described in US Pat. Nos. 5,599,976 and 5,723,660, a portion of the overhead stream 184 condensed in the drum 56 is passed through a delivery line (not shown) to the diesel oil column 30 ) Or to a dry column feed stream 140. The present invention improves the process control of the processes described in the patent. Dry acetic acid is removed from the bottom 52 of the drying column 50 and pumped directly into a product tank (not shown) or through the stream 190 by pump P6 through the heavy oil column 60. Pumped to further remove impurities such as propionic acid. Additional acetic acid is recovered from stream 197 and fed to waste acid stripper column 80 by pump P7. The recovered acetic acid is returned to the heavy oil column 60 via stream 198. Waste propionic acid and higher acid are passed through stream 199 for disposal. The overhead of the heavy oil column 60 is transferred to the drum 64 via the stream 192 to condense and return to the heavy oil column 60 through the stream 196 as reflux by the pump P10 or to the stream. Passed through stream 194 to mix with 140.

The purpose of the optional heavy oil column 60 is to remove high boiling propionic acid impurities from acetic acid. The lower portion 62 of the heavy oil column 60 consists mainly of propionic acid, while the side draw stream 200 consists of pure acetic acid delivered for storage in the product tank.

B. Process Control

The overall facility control method that is generally accepted is the concept of inventory control (also known as level control). This is mainly achieved in two zones: the reaction section and the diesel oil recovery section. In the reaction section, the flashing rate of the reactor solution should be sufficient to match the reactor feed rate and thus maintain the reaction temperature as well as the desired reactor solution level. Reactor coolers (not shown) are frequently used to support reaction temperature control because the flashing ratio changes. In the diesel recovery section, the inventory concept has two main functions. The first function is to return MeI back to reactor 10 with minimal inventory shift. The level of control of the serious injury 42 at boot 46 accomplishes this goal. The second function is to remove the correct amount of acetic acid from stream 140 with the light oil column sided and feed it to the drying column 50. The level controlling the side draw tray (not shown) of the gas oil column 30 and the flow controlling the decanter phase recycle stream 180 to the reactor 10 achieve this goal. These two control actions are performed together because the flow of the recirculated stream 180 to the reactor 10 is reduced when a greater amount of crude acetic acid needs to be fed through the stream 140. to be. This refluxs more light phase through the stream 170 into the gas oil column 30 and then pushes more acetic acid below the column 30 into the tray with a level controlled sided. Conversely, when smaller amounts of crude acid have to be removed from the system, the flow of recycling the bed phase to reactor 10 through stream 180 of decanter 40 is increased. This results in lower column reflux through stream 170, more acetic acid within the overhead of column 30, and less acid on the side drawer tray to be fed through stream 140. Another major control concept is that it contains HI at the bottom 34 of the gas oil column 30. The minimum required water concentration in this bottom 34 ensures that HI is kept below the sided tray and returned to the reaction section rather than proceeding to the drying column 50.

Although the concept of inventory control plays an appropriate role in ensuring quality control, compositional changes can occur which can lead to undesirable conditions in the process. Since some of these compositional changes cannot be detected by inventory control (or at least not in an appropriate manner), continuously updating the component concentrations of the columns and the various streams may result in the operation of the diesel fuel recovery and purification sections. It is recognized by those skilled in the art of acetic acid manufacturing that it provides significant advantages in optimizing. This is especially true in the case of low water (in this technique, the concentration of water in the various columns is important for product purity as described below).

Typically, downstream analysis is done by off-line sampling prior to laboratory analysis on some of these streams. These samples are taken several times daily, and in general, there is a delay of several hours between sampling and stream composition data available to facility operators. In addition, this off-line data is useful for determining average column performance, but does not provide any information about any possible circulation in all of these vessels. Circulation of the refinery can infiltrate impurities, such as iodide, into the product stream 200 (these impurities eventually reach the product storage tank). This circulating action cannot be discerned by sparse daily sampling, but will be readily discerned from the continuous on-line data according to the present invention.

The ability to analyze various process streams and columns of a conventional acetic acid plant, such as the acetic acid plant shown in FIG. 1, allows for various levels of process control. Each vessel in the purification section can be controlled individually or in combination with each other. Each vessel can be controlled using the mass balance principle, ie the mass flow into and out of the vessel is adjusted based on the composition in and out of the stream, so that the monitored component or components can be controlled alone or in combination. Depending on the desired level of control required, it may be necessary to vary the temperature or temperature distribution in the distillation column. Typically, this can be accomplished by adjusting only the steam flow to the column reboiler or by adjusting the steam flow along with the column pressure. In addition, it is sometimes necessary to additionally control the columns before and after the column controlled by the cascaded control scheme. In addition to monitoring the various streams possible by the present invention, analytical probes are inserted at key points throughout the individual purification columns to provide information on component gradients within the columns. This type of profile information can be used to minimize the upset by controlling the rate of change of the column state. The process control of the present invention can provide significant advantages in low process processes, for example, as described in US Pat. Nos. 5,001,259, 5,026,908, 5,144,068, 5,750,007, 5,817,869 and 6,031,129. .

Although not explicitly shown in FIG. 1, as described, for example, in US Pat. No. 3,772,156, acetic acid equipment typically uses reactive distillation techniques in some columns to inject potassium acetate or methanol to remove iodide. The present invention controls the feed rate of the components injected into the column based on the analysis of the column profile, feed or outlet.

1. Process control in the gas oil decanter

The state in which the continuous update of the component concentration in the gas oil decanter vessel 40 becomes useful for process control is outlined below.

If reactor 10 is to be operated at low and / or high MeOAc concentrations, a condition in which MeI cannot be separated as a separate phase in decanter 40 should be avoided. When the water concentration in the mirror phase 44 drops to a low level, the mirror phase 44 mainly becomes acetic acid, and an increasing amount of MeI can be dissolved in the mirror phase 44. In extreme cases, only one phase is present in decanter 40, which dilutes the MeI concentration to low levels over the entire decanter 40. Thus, a pump designed to recycle a high MeI concentration phase cannot recycle enough MeI to reactor 10 to maintain a reaction. Thus, the reactor 10 is "quenched" and the operating parameters need to be adjusted to resume the desired reaction. In addition, the high concentration of MeOAc in the phase 44 dissolves MeI in the phase 44 such that only one phase is present in the decanter 40. For example, if subsequent analysis of stream 160 indicates that the MeI concentration is reduced, the flow may be increased in pump P2 of FIG. 1 so that a constant amount of MeI is returned to reactor 10.

The portion of the mirror bed 44 returned as reflux to the gas oil column 30 serves to balance the water load between the reactor 10 and the downstream drying column 50. If too much water is returned to the diesel oil column 30 by increasing the reflux through stream 170 and reducing the recycle to reactor 10 through stream 180, more water is sidedraws the product. It is eventually removed at the product takeoff and then passed to the drying column 50 via stream 140. The water load in the drying column 50 may be increased to such an extent that this drying column overflows to perform the water removal function. Conversely, if too much water is delivered to reactor 10 via recycle stream 180, drying column 50 is not critically loaded, resulting in impurities such as hexyl iodide, resulting in product stream 200. Can penetrate into). Another embodiment of the analytical performance of the present invention is to control the diesel oil column 30 by maintaining acetic acid and water concentrations in the side draw stream 140. One embodiment of the control scheme is to adjust only the gas oil column reboiler temperature based on acetic acid and water analysis or adjust this temperature with column reflux 170.

Since the concentrations of water and / or MeOAc can vary in the reactor and downstream regions, to increase the production rate or to change the quality of the product, monitoring the composition of both the phase 44 and the medium phase 42 is necessary. Importantly, it is particularly important to avoid the case where the decanter 40 contains only one phase. On-line analysis of the two phases 42, 44 may inform the operator of composition changes that may cause the state of one phase in the decanter 40 if left inaccurate. In particular, the water concentration in phase 44 and the concentration of MeOAc in both phase 44 and medium phase 42 can be used as guidelines to avoid undesirable conditions in dry column 50. Parameter. In addition to avoiding the state of one phase in the decanter 40, awareness of the water concentration may also be used to maintain proper equilibrium of water between the reactor 10 and the drying column 50. One embodiment of an integrated control scheme based on multiple successive analyzes of serious injury 42 and minor injury 44 is provided below. Referring to FIG. 1, when the MeI concentration decreases in the middle phase 42, the flow by the pumps P2 and P3 is adjusted to maintain a constant amount of MeI returned to the reactor 10. Since changes in the flow of the stag 44 affect the water equilibrium in the reactor 10, the water analysis of the stag 44 adjusts the make-up water (not shown) supplied to the reactor 10, It is used to maintain a constant water concentration.

Another embodiment of the present invention is to apply the process control of the present invention to the acetic acid processes described in US Pat. Nos. 5,599,976 and 5,723,660. If the analysis of the medium phase 42 and the minor phase 44 immediately indicates the presence of a single phase, a portion of the stream 186 containing the high water concentration is directed towards the stream 120 or decanter 40 (not shown). ) Phase separation can be improved. In addition, the equilibrium of water in the decanter 40 can be maintained by adjusting the flow to the pumps P3 and P4.

Hydrocarbons can sometimes be enhanced in the middle phase with MeI. Carbon monoxide (CO) used in methanol carbonylation generally results from incomplete combustion of natural gas or oil residues. Thus, the CO stream may contain traces of hydrocarbons, generally alkanes. Thus, these hydrocarbons, which generally have a low boiling point, tend to concentrate in the decanter mesophase 42, where these hydrocarbons have high solubility. Since the lower alkane density is about 0.7 g / mL, not only dilute MeI in the middle phase 42, but also the lower alkane tends to reduce the medium phase density. Since medium phase separation is a function of both solute immiscibility and density, increasing hydrocarbon concentrations can adversely affect decanter performance. As shown in FIG. 1, an alkane (hydrocarbon) removal column 70 receives the flow of the middle phase 42 including hydrocarbons and MeI from the boot 46 of the decanter 40 through the stream 210. MeI is separated and fed back to the reaction section via stream 220, while hydrocarbons are removed via stream 230. Alkanes removal columns are known to those skilled in the art of acetic acid production and can be operated batchwise or continuously. On-line analysis of the hydrocarbon allows to determine the proper flow of decanter middle phase 42 to the alkane removal column 70.

Acetaldehyde (CH ₃ CHO) is an undesirable by-product in reactor 10. This undergoes hydrogenation in reactor 10 to form ethanol and then carbonylated to form propionic acid, which is difficult to separate from acetic acid in the purification section. Acetaldehyde can also be condensed to provide alcohols such as C ₄ , C ₆ , C ₈ , which form the corresponding iodides. C ₆ iodide, ie hexyl iodide, is particularly poor because it is difficult to separate from acetic acid in the drying column 50 and can lead to iodide contaminants of the acetic acid product as described, for example, in EP Publication No. 985,653. It is difficult to implement acetic acid production technology.

Acetaldehyde has a boiling point of only 21 ° C. and therefore tends to accumulate above the gas oil column 30 and concentrate in the decanter 40. Thus, on-line analysis of this component on both decanter beds 42 and 44 is useful for identifying the optimal reactor conditions that minimize by-product formation in combination with on-line reactor analysis for other components. Various modifications to the process are known in the art for removing acetaldehyde, and the analytical methods described herein improve process control for any of these known processes. Although the concentration of acetaldehyde in the reactor solution is so low that it cannot be analyzed on-line, any change in acetaldehyde concentration in the decanter 40 (this change can be monitored in the decanter) is appropriately predicted so that the reactor 10 Can be correlated with changes within.

2. Process control at the bottom of the gas oil column

Hydrogen iodide (HI) forms a high boiling point azeotrope in acetic acid solution with more than about 5% by weight of water. If the water concentration falls below about 5% by weight, the azeotrope will begin to breakdown and HI will begin to volatilize. This volatilization can adversely affect the reactor iodide inventory by reducing HI in the bottom stream 150 and returning it to the reaction section. From now on, the volatilized HI will primarily be part of stream 140 as a wet acetic acid sided feed to dry column 50. In general, the process equipment used to manufacture HOAc does not actually work for these components, but if the HI concentration level in the system reaches an extremely high level, it may be corroded or otherwise adversely affected. Thus, the presence of significant concentrations of HI in this feed is of importance with regard to corrosion of the purification vessels and iodide contaminants of the final acetic acid product. Therefore, it is important to maintain the water level at a minimum concentration of about 5% by weight (about 3 moles) at the bottom 34 of the gas oil column 30 in relation to both reactor performance and refinery performance.

HI and water can be measured accurately by either extended mid-infrared (4000-7000 cm ⁻¹ ) or mid-infrared (400-4000 cm ⁻¹ ) spectroscopy. Thus, on-line analysis of the gas oil bottom 34 assists to avoid column upsets by acting to indicate when the water concentration reaches a critical low level.

3. Process Control in Dry Column Feed and Dry Column

The dry column feed is a side stream 140 from the gas oil column 30. It is mainly acetic acid and water with small amounts of MeI and MeOAc and trace amounts of HI. The main purpose of this column 50 is to produce dry acetic acid. Water is removed through overhead stream 184. However, the drying column 50 is not designed to produce a very pure overhead stream. The water composition in overhead stream 184 should optimize the additional cost of recycling acetic acid and the cost of separating acetic acid from water. A portion of this overhead stream 184 is recycled to the reactor 10 via stream 186 by pump P8 to maintain system water inventory or to pass water through the gas oil column 30 and It can be used to feed the decanter 40.

The drying column 50 is operated based on the concept of temperature gradient and has a tray (not shown) loaded with a plurality of liquids. On-line analysis of the water in the dry column feed 140 and in these trays provides an instantaneous and continuously updated water profile of this column 50. Such profiles are used for many purposes. When too much water is passed from the stream 140 to the drying column 50 in the light oil column sided, the water load in the drying column 50 may increase to such an extent that the drying column overflows and cannot perform the water removal function. Can be. Conversely, if too much water is passed through the stream 150 to the reactor 10 to recycle, the drying column 50 may not be loaded. Combining the on-line analysis of the gas oil bottom 34, the gas oil side draw stream 140 and the dry column tray allows any undesirable decrease or increase in column loading to be quickly identified and processed.

A second advantage of the water profile of the drying column 50 relates to the low water technique. Hexyl iodide, i.e., a small amount of byproduct, forms azeotropic mixtures having a boiling point with acetic acid and water, and are difficult to remove by distillation. In addition, the volatility of this azeotrope is reduced as the water concentration decreases due to the potential for increased iodide contaminants at the bottom 52 of the drying column 50. This water profile concept applied to the drying column 50 is described for example in EP Publication No. 985,653, and the present invention provides an improved control method for the water profile of the drying column 50 useful in this regard. It is. Hexyl iodide is provided at a billion level (ppb) level that is too low to be detected or quantified by FTIR, but the perception of column water content versus azeotrope breakdown water concentration is that an on-line analysis of water has shown that iodine in acetic acid product To be used as a means of controlling cargo contaminants. This dynamic control eliminates the need for a dedicated iodide removal column, thereby greatly reducing costs.

4. Process Control in Heavy Oil Columns

The main purpose of the heavy oil column 60 is to remove propionic acid impurities from acetic acid. The lower portion 62 of this column is a low volume propionic acid stream, while pure acetic acid is withdrawn from the side stream 200. Analysis of extended mid-infrared quantifies propionic acid down to about 200 ppm by either transmission cell or fiber coupled probes. Typically, the feed from stream 190 to heavy oil column 60 may comprise about 200-1000 ppm of propionic acid. Thus, analysis on feed stream 190 and acetic acid side draw stream 200 monitors column performance and continuously updates the purity of the acetic acid product with respect to propionic acid content. The assay also quantifies the water concentration down to about 100 ppm. Pure acetic acid products typically contain about 100-500 ppm of water.

Another embodiment of process control is the amount of condensed stream 192 versus the amount of column reflux stream 196 recycled from overhead drum 64 through stream 194 to stream 140 by pump P10. Is controlled based on the propionic acid concentration of stream 200. In addition to the flow of the two streams 194, 196, the column reboiler temperature may be adjusted.

Propionic acid and higher acid are provided to the bottom of the heavy oil column 62 with acetic acid at a substantial concentration. Additional acetic acid is recovered by the feed stream 197 from the bottom 62 to the waste acid stripper column 80. Analysis of waste stream 199 can be used to optimize heavy oil column temperature and flow by pump P7 to maximize acetic acid recovery and minimize waste for disposal.

C, calibration modeling

Infrared calibration models are obtained using chemometric techniques as described in US Pat. No. 6,103,934 and US Patent Application 09 / 611,067, the entire contents of which are incorporated herein by reference. Chemometrics are a form of chemical analysis that uses statistical data (in this statistical algorithm, algorithmic relations and mathematical logic are combined to obtain a calibration model including multivariate analysis). The term multivariate analysis refers to the relationship of component concentration in solution to many infrared wavelengths or frequencies. Get software products for applying chemometric technology. A typical example is PIROUETTE ^™ from Infometrix, Seattle, WA. The general steps involved in developing a chemometric calibration model are well known to those skilled in the art. In addition, the American Society for Testing and Materials (ASTM) published a document entitled "Standard Practices for Infrared Multivariate Analyses (No. E1655-94)", which is hereby incorporated by reference in its entirety. Reference is made.

In order to obtain a good chemometric calibration model, it is important to choose the calibration standard appropriately. If the signal of a component of interest that is superimposed with a signal from another component is significantly weak, a large number of calibration standards may need to be prepared and analyzed. This number can range from 30 to 300. In order to produce an accurate calibration model, a number of calibration standards are prepared, each of which typically includes all components provided in the reaction system solution. Some or all of these components must eventually be analyzed by infrared spectroscopy. Components with individual standards are changed independently by concentration, randomizing any bias or interference that one component may have over another. The predicted maximum and minimum concentration values in the reaction system solution serve as boundary limits for each component concentration. After the standards have been prepared, they are sequentially injected into the infrared analyzer and the spectroscopic signals are collected. In general, each spectrum for a calibration standard is first converted into a digitalized format and then set up in a spreadsheet with the concentration of the corresponding component to be measured. Then, the Partial Least Squares regression method is used to fit the data. Finally, the accuracy of the calibration model is compared with the concentration of the reaction system components obtained from the on-line analyzer during the actual process run and the concentrations obtained by actually sampling the various streams and columns, and using an independent off-line analytical method for It is checked by analyzing.

The spectral regions used to obtain calibration models for each component in the extended mid-infrared from 7000 cm ⁻¹ to 4000 cm ⁻¹ and in mid-infrared from 4000 cm ⁻¹ to 400 cm ⁻¹ are shown in Table 1.

TABLE 1

Attenuated Total Reflectance (ATR), mentioned in the heading of Table 1, is well known to those skilled in the art for spectroscopy and is incident from crystals made from high refractive index infrared transmissive materials from infrared sources. It includes light. This light is reflected internally and extends beyond the surface of the crystal to a sample in contact with the crystal. Part of the infrared energy is absorbed and the rest of the energy is passed to the detector. ATR generates very short path lengths for infrared light in this sample.

The term transmission, as used in Table 1, refers to the case where infrared energy passes directly through a sample, which absorbs a portion of the infrared energy. Similar to the ATR technique, the rest of the energy reaches the detector.

Infrared spectroscopy is well known to those skilled in the art so that the fundamental absorption bands in the mid-infrared have corresponding overtone bands or have lower intensity "echo" in the extended mid- / near-infrared region. The near-infrared region, which combines with the extended mid-infrared region, extends from approximately 4000 cm ⁻¹ to 12500 cm ⁻¹ . The first overtone band, which occurs in the extended mid-infrared region, is on the order of magnitude less than the basic absorbance. Similarly, the second overtone band, which occurs at much higher near-infrared frequencies, is again on the order of magnitude lower in intensity. Thus, any overtone band in near-infrared can be used for quantitative analysis by modifying the cell or probe window path length. All bands of interest for component measurements in accordance with the present invention are in the infrared region.

The infrared equipment, transmission cell, ATR cell and transmission probe to be depicted in the laboratory simulation are substantially the same as the equipment used in the manufacturing facility. Process-established versions of laboratory infrared spectrometers are provided by several vendors, such as Analect Instruments, Pomona, Canada, and ABB Bomem Inc., Quebec, Canada. The electronics and process spectrometers in the lab are essentially the same, but considering the safety and the environment, the process spectrometers need to be packaged in explosion proof purged cabinets. Similarly, infrared cells are typically kept at a constant temperature in a conditioned sample cabinet because of the changing climate conditions in the process environment compared to the laboratory.

If a particular component is more concentrated at one location in the reaction system compared to the concentration at another location, the appropriate spectral region used for quantitative analysis of the particular component is determined by the stream or vessel (this spectral region in this stream or vessel). It can be changed according to the analysis. This There know from Table 1, for example, MeI is the intermediate expansion of 6261-5351cm ^-1 in serious injury decanter-extended intermediate of 6211-5505cm ^-1 in the infrared region and a decanter current - is measured in the infrared region.

For each stream or vessel to be simulated, a multicomponent solution is prepared to obtain the spectroscopic data needed to generate the calibration model. The concentration ranges of these solutions match the ranges examined in the individual on-line examples described in detail. For example, the concentration ranges used in the preparation of decanter thin and medium phase calibration standards are listed below.

As outlined in the above application, the calibration model for all components in all containers / streams prepared solutions of known component concentrations to obtain infrared spectra for these components and to obtain these spectra using appropriate calibration models. Validated by quantitative analysis.

The effectiveness of the decanter mesophase calibration model is provided in table form in Table 2 and in the form of a graph for the CH ₃ CHO in FIG. 2, where eight solutions of known composition are prepared and the concentration is predicted using the calibration model. A correlation coefficient (R factor) of 0.997 or greater, associated with the values in Table 2, indicates that the accuracy is high in analyzing all components and also in quantifying the density. Similar validity is implemented for the decanter current calibration model. The data for this validity is shown in Table 3. Similarly, Tables 4, 5, and 6 report valid data for light oil column bottom solutions, heavy oil column / product tank solutions, and dry column feed solutions, respectively.

TABLE 2

TABLE 3

TABLE 4

TABLE 5

TABLE 6

The compatibility of the ATR method and the extended-medium transmission infrared analysis method is shown in Table 4 as two sets of data. This table compares the accuracy of the ATR method versus the extended medium permeation method for the analysis of light oil column bottom solutions. This table shows that the accuracy of both techniques is excellent over all ranges of HI, H ₂ O and HOAc tested.

The accuracy and precision of infrared measurements is primarily a function of the signal-to-noise ratio for the spectral response of the component of interest. Noise is typically referred to as a small, random change in the performance of equipment associated with an electronic device. It is well known to those skilled in the art with respect to spectroscopy that when the component to be measured is provided at a sufficiently high concentration, the signal to noise ratio is very large so that the noise does not adversely affect quantification. When the infrared signal is very small, the signal-to-noise ratio will be low enough so that the noise affects quantification, as in the ppp level measurement of acetaldehyde on two decanters and the ppp level measurement of propionic acid in a heavy oil column or product stream solution. Can be. This random contribution of noise can be averaged by

(a) Increase the length of time to capture the spectrum. Typical capture time is around 30 seconds. For example, increasing this acquisition time by 1-3 minutes will reduce noise while keeping the signal unaffected. or

(b) Report this mean by averaging a number of spectral results.

These two averaging techniques are used to obtain quantitative data for acetaldehyde, HI, water and propionic acid at ppm levels. The correlation in FIG. 2 shows the scatter involved in five replicate measurements of the decanter mesophase acetaldehyde in each of the eight effective solutions. At concentrations below 1000 ppm, the individually predicted values can vary by +/- 300 ppm from the actual values. However, the average of five measurements, the values reported in Table 2, improves the accuracy to about +/- 100 ppm. Even with this averaging, the signal becomes too low to allow reliable quantification at acetaldehyde at concentrations below 200 ppm. Similarly, the ppm values of propionic acid and water in Table 5 and the ppm values of HI in Table 6 are averages of five measurements. However, infrared measurements of individual components will vary in accuracy and precision as a function of the specific component mixture in the sample. Thus, if the specific spectral response is strong enough to be distinguished from the response due to other components in the sample, a good signal-to-noise ratio is obtained, even with low component concentrations. Thus, even if the concentration of water as reported in Table 5 has a low ppm value, water tends to have a stronger and more pronounced signal, resulting in greater reliability, actual and predicted compared to other low ppm concentrations of components. Make the deviation between the calculated values smaller.

In all embodiments involving the decanter 40, equipment as shown in FIG. 3 is used to simulate process equipment. A glass decanter 40 made from a glass reducing tee is used to study the appearance to measure the composition of the hard phase 44 and the medium phase 42. The base 48 of the decanter 40 is 3 inches (inner diameter) x 14 inches (length), and the boot 46 is 2 inches (inner diameter) x 5 inches (length). Typical light and medium phase compositions are prepared and added to the decanter 40. The serious injury 42 is collected at the bottom of the boot 46 with an interface about 2 inches below the bottom level of the base 48 of the decanter 40. In order to minimize the material for these studies, the minor 44 and medium 42 are independently and continuously through each of the return lines 300 and 310 by a pump P that returns to the decanter base 48. Recycled. The two streams from the return lines 300, 310 are mixed by a static mixer 41 just before entering the decanter 40, which is typically experienced in a commercial or pilot plant. Simulate a mixture of ingredients. In this way, the equilibrium of the composition can be set between the minor phase 44 and the medium phase 42. The composition is changed by adding additional water, acetic acid, methyl acetate, hydrocarbons, MeI, CH ₃ CHO or a mixture thereof to the light phase 44. These additions are made by a three-way valve (not shown) connected to the light pump inlet (not shown).

In all embodiments involving light oil column 30, drying column 50, heavy oil column 60 and product tank, equipment as shown in FIGS. 4A and 4B is used. A three-liter, rounded bottom 1 liter flask 92 is filled with 300 mL of the appropriate solution 94 and agitated. The composition is changed by adding the desired ingredients directly to the flask 92 via a syringe.

In all examples, the main methods used to manually analyze the obtained samples were gas chromatography (GC), Karl Fischer (KF) water titration and silver titration iodide. Argentometric iodide titration. These are widely established analytical methods in acetic acid production because these methods have an accuracy greater than or equal to +/- 5%. GC analysis is performed using Varina Instruments' CA Model 3400 instrument in Walnut Creek, which is equipped with a NUKOL®60 meter capillary column and flame ionization detector. Karl Fischer water analysis is performed using EM Sciences' AQUASTAR® Model V1B titrator, Gibbston, NJ. Iodide (eg hydroiodic acid) analysis is performed by titration with silver nitrate, where the end-point is determined by the use of the Eosin Y marker. Solution density is determined by weighing 1 mL of solution.

The infrared spectrometer used in all examples is a DIAMOND®20 model obtained from Analect Instruments, Pomona, Canada. Two types of detectors are used. In Examples 1-12, including the use of flow through permeate cells installed in the cell compartment of the spectrometer, a Deutero triglycine sulfide (DTGS) detector is used. In all other embodiments, including the use of optical fiber coupled transmission probes, InAs (Indium Arsenide) detectors are used. Signal loss when transmitted over fiber optic cables is offset by using a more sensitive InAs detector.

Several methods are used to obtain on-line infrared analysis for inspection purposes. In the extended-medium infrared region for the decanter mirror 44 and midphase 42, infrared analysis involves the return of one of the streams 300, 310 to zinc zinc selenide disposed in the cell compartment of the FTIR spectrometer 90. selenide) by connecting to the inlet of a 0.5 mm permeable cell equipped with a window, the outlet being connected to the pump P, which is then continuously recycled to the decanter 40 as shown in FIG. Let's do it. In addition, a 0.5 mm path length optical fiber coupled transmission probe (available from Axiom Analytical, Irvine, Canada) equipped with a sapphire window is provided with a base 48 and a boot of the decanter itself as shown in FIG. 46, or into return lines 300 and 310 (not shown). On-line analysis of the solution in the 1 liter flask 92 similarly continued the solution 94 back into the flask 92 through line 320 through the infrared cell 90 as shown in FIG. 4A. Or by inserting the optical fiber coupled transmission probe 96 directly into the 1 liter flask 92 as shown in FIG. 4B.

Infrared analysis in the mid-infrared region similarly replaces the return line of one of the decanter streams 300, 310 or the outlet line 320 of the flask 92 with a zinc selenide disposed in a compartment of the FTIR spectrometer 90. This is done by connecting to the inlet of the crystal-mounted ATR tunnel cell (available from Xiom Analytical), which outlet is connected to the pump as described above.

Example 1:

In this embodiment, the return line 310 of the decanter mesophase solution is connected to a 0.5 mm transmission cell equipped with a zinc selenide window, ie the mesophase 42 is continuously monitored by an extended mid-infrared analysis. This transmission cell is placed in a cell compartment of an FTIR spectrometer equipped with a DTGS detector. While the on-line data points are recorded about every 35 seconds as shown in FIG. 6, the decanter solution of the composition as shown in the first data column of Table 7 is allowed to circulate for about 18 minutes. The trend line during this 18 minute period indicates that measurement accuracy better than +/- 0.06 moles is achieved for all components. At this time, MeOAc is added to the decanter. The trend line of FIG. 6 shows the result of the addition of this MeOAc to the composition of the medium phase as measured by infrared light. As expected, MeOAc shows a significant increase in concentration from 2 moles to 2.65 moles, and MeI, ie, bulk solvent in the middle phase, shows a decrease in concentration from 11.7 moles to 10.7 moles. The presence of higher concentrations of MeOAc in the medium phase also allows greater solubility for both HOAc and water, as shown by the trend line.

The medium phase is sampled manually before and after the addition of MeOAc. These samples are analyzed by off-line independent analytical methods, in particular GC and Karl Fisher. The data in Table 7 show good agreement between the values predicted by the on-line infrared and off-line techniques, and the on-line infrared analysis performance results in small component concentrations resulting from increasing the concentration of MeOAc in the mid-phase. Even changes are detected.

TABLE 7

Example 2:

Conditions similar to those of Example 1 are used. The decanter mesophase solution of the composition as shown in the first data row of Table 8 is monitored for a period of time during which aliquots of hydrocarbon solution are added at approximately 12 minute intervals. The trendline of FIG. 7 is predicted to increase from about 0.25 moles to 0.70 moles of hydrocarbon concentration, and is associated with a decrease in the concentration of other heavy phase components when diluted. Solution density also decreases as a function of very light hydrocarbons (approximately 0.7 g / mL) replacing very heavy MeI (approximately 2.3 g / mL). This solution is sampled manually before each hydrocarbon addition and this sample is analyzed by off-line independent analytical methods, in particular by weighing the GC and density for concentration. The data in Table 8 indicate that the correspondence between on-line and off-line techniques is close, and that on-line infrared performance accurately quantifies changes in hydrocarbon concentration.

TABLE 8

Example 3:

Except where noted, similar conditions as in Example 1 were used. This example illustrates on-line monitoring of this acetaldehyde because the concentration of acetaldehyde increases in the decanter phase. The neutral phase solution is monitored for a two hour period during which data points are obtained every two minutes. During this two hour period, nine aliquots of acetaldehyde are added. The trend line in FIG. 8 shows that acetaldehyde increases from a starting value of zero to about 12,000 ppm. The solution is sampled manually before each acetaldehyde addition and this sample is analyzed by GC. Since there is residual noise in the infrared data, the five infrared data points are averaged to provide infrared data in each row of Table 9. As shown in this table, when infrared values are compared to GC values, data point averaging allows for an accuracy of approximately +/- 150 ppm. Thus, this embodiment effectively monitors acetaldehyde at a level normally predicted in decanter slander operation using, for example, a rolling point of data points over a suitable brief time period of approximately 10 minutes. Indicates.

TABLE 9

Example 4:

In this embodiment, the return line 300 of the decanter mirror 44 is connected to an infrared cell. The infrared equipment is the same as the infrared equipment described in Example 1. While the online data points are recorded about every 35 seconds as shown in Figure 9, the decanter solution of the composition as shown in Table 10 is allowed to circulate for about 9 minutes. During this period, the trend line indicates that a precision of +/- 0.05 moles is achieved for all components. At this time, water is added to the decanter. The trend in FIG. 9 shows the result of the addition of this water to the mirror composition as measured by infrared light. As expected, water indicates a significant change in concentration from 25.41 moles to 27.92 moles, and MeI, which cannot be mixed with water, indicates a maximum percentage reduction in concentration. Injuries are manually sampled before and after the addition of water and the samples are analyzed by GC and Karl Fischer titration. The data in Table 10 show that the correlation between on-line infrared and off-line infrared techniques is good, and that on-line infrared performance accurately quantifies changes in the current solution component in response to changes in water concentration.

TABLE 10

Example 5:

This example is similar to Example 4 except that HOAc is added to the decanter instead of water. The trend line in Fig. 10 shows the result of the addition of this HOAc to the current composition. As expected, the bulk solvent in water, i.e., phase 44, shows a significant decrease in concentration from 28.02 moles to 25.56 moles. The low water concentration in the phase 44 increases the MeI concentration at this phase, ie redistributes MeI from the medium phase 42 to the phase 42. The mirror 44 is manually sampled before and after HOAc is added. The data in Table 11 show good correlation between the values predicted by on-line infrared and independent off-line GC and Karl Fischer technology, and this performance is quantified by detecting changes as low as 0.05 moles at all component concentrations. It is shown.

TABLE 11

Example 6:

Conditions similar to those in Example 4 are used except as specifically mentioned. This embodiment shows on-line monitoring of this acetaldehyde because the concentration of acetaldehyde increases in the decanter phase 44. The current solution is monitored for a two hour period during which data points are obtained every two minutes. During this two hour period, ten aliquots of acetaldehyde are added. The trend line in FIG. 11 shows that acetaldehyde increases from a starting value of zero to about 21,000 ppm. The solution is sampled manually before and after each acetaldehyde is added and this sample is analyzed by GC. As in Example 3, five infrared data points are averaged to provide infrared data in each row of Table 12. An accuracy of approximately +/- 150 ppm is obtained, which indicates that the rolling average of the data points is used over a period of about 10 minutes to efficiently monitor acetaldehyde at levels normally predicted in decanter current operation.

TABLE 12

Example 7:

In this embodiment, the return line 310 from the decanter mesophase 42 is connected to an ATR tunnel cell equipped with zinc selenide crystals. Analysis is performed in the mid-infrared region using a DTGS detector. About 45 minutes of on-line data is obtained from the specific prepared decanter compositions described in Table 13. A data point acquisition of about 35 seconds is used. Samples obtained at the midpoint of the experiment are analyzed off-line by GC or Karl Fischer and the data is compared with the on-line data in Table 13. Close correspondence between off-line and on-line values indicates the accuracy of the ATR method in on-line analysis. The trend line in Figure 12 represents the precision of this technique of composition monitoring.

TABLE 13

Example 8:

Using equipment as shown in FIG. 4 and infrared equipment as described in Example 1 (except that sapphire is used rather than zinc selenide window), the light oil column bottom 34 type solution is 100 minutes. While circulating through the infrared transmission cell. The initial composition of this solution is described in the first row of Table 14. Data points are obtained approximately every 35 seconds. Six aliquots of HI are added to the flask by syringe over the first 60 minutes. This flask is sampled manually before adding each HI for off-line analysis by titration, GC and Karl Fischer. The trend line in FIG. 13 shows that HI increases stepwise after each addition. Water also indicates that the commercially available HI used in this experiment is added in calculation because it is a 55% aqueous solution. Three aliquots of water are added during the 60-90 minute monitoring period to indicate that the HI calibration model is unbiased or unaffected by the increase in water. The trend line indicates that the water increases stepwise during this period without increasing HI at all. HOAc dilutes with HI / water, indicating a constant decrease in concentration over the monitoring period. The good correlation between off-line analysis and on-line infrared analysis is evident from Table 14. Thus, this experiment shows that the on-line infrared measurement performance provides a complete and accurate profile of the light oil bottom 34 type solution. The measurement accuracy demonstrated in this example indicates that process control relating to the prevention of azeotrope breakdown can be made using on-line infrared analysis.

TABLE 14

Example 9:

Using equipment such as Example 8 and infrared equipment such as Example 1, the product stream 200 type solution is continuously circulated through the infrared transmission cell for about 1 hour. The starting solution is 300 mL of HOAc sold. It should be noted that all commercially available HOAc contains ppm levels of H ₂ O and propionic acid. Karl Fischer and GC analysis show that this starting solution contains 486 ppm H ₂ O and 105 ppm propionic acid. During monitoring, 5 aliquots of H ₂ O are added to the flask by syringe. The trend line shown in FIG. 14 indicates that the H ₂ O concentration is increased as measured by on-line infrared light. Passive samples obtained before the addition of each H ₂ O aliquot are analyzed by Karl Fischer. Comparison of on-line and off-line data in Table 15 shows that H ₂ O can be measured with an accuracy of approximately +/- 25 ppm. As in Examples 3 and 6, the five infrared data points are averaged to obtain the data reported in Table 15. The trend line shown in FIG. 14 also includes data for propionic acid. This figure shows that the variable for the actual value (approximately +/- 100 ppm accuracy) is not affected by the increase in H ₂ O concentration, ie the calibration model for propionic acid is damaged by the variable H ₂ O in solution. It does not appear. The results from this experiment indicate that product quality control with respect to H ₂ O and propionic acid concentration can be achieved through on-line infrared analysis due to the rolling average of data points approximately every 10 minutes.

TABLE 15

Example 10

In this case, an experiment similar to Example 9 is carried out except that an aliquot of propionic acid is added to HOAc in the flask rather than H ₂ O. The purpose of this experiment is to show that on-line infrared monitoring of heavy oil column 190 solution or product stream 200 solution can be used to efficiently quantify ppm level changes in propionic acid concentration. The solution is continuously circulated through the infrared cell for 3 hours with a data point frequency of 2 minutes. During this period, seven aliquots of propionic acid are added to the flask by syringe. The trend line shown in FIG. 15 shows an increase in propionic acid concentration as measured by on-line analysis. Passive samples obtained before all propionic acid addition are analyzed by GC. The comparison of on-line and off-line data in Table 16 (in this table, each reported infrared value is an average of 5 data points) shows that the propionic acid has an accuracy of approximately +/- 100 ppm by on-line infrared analysis. It can be measured as. This result indicates that the heavy oil column performance can be monitored and optimized based on the propionic acid concentration in the feed 190 versus the propionic acid concentration in the heavy oil stream to the product tank, ie the product stream 200.

TABLE 16

Example 11:

Using equipment as shown in FIG. 4 and infrared equipment as described in Example 9, the dry column feed 140 type solution is continuously circulated through the infrared cell for 75 minutes at a data point frequency of 1 minute. The initial composition of this solution is 3.0 moles of H ₂ O, 16.6 moles of HOAc and 0 ppm HI. For a period of 75 minutes, 15 aliquots of the HI solution are added to the flask via syringe. The trend line for HI in FIG. 16 shows that the concentration increases as measured by on-line infrared analysis. Passive samples obtained before each addition of HI are analyzed by titration. The comparison of on-line and off-line data in Table 17 (in this table, each reported infrared value is an average of 5 data points) shows that the HI has an accuracy of approximately +/- 100 ppm by on-line infrared analysis. It can be measured. Trend lines for H ₂ O concentration and HOAc concentration are also included in FIG. 16. The starting and ending concentrations as measured by off-line analysis are within 0.05 moles of the on-line infrared value. These results indicate that dry column feed 140, monitored by on-line infrared analysis, optimizes gas oil column performance to minimize HI concentration in this feed. Similarly, H ₂ O concentration measurements in this feed can be used to determine the optimal reflux 170 from the decanter mirror 44 to the gas oil column 30.

TABLE 17

Example 12:

Using the equipment as shown in Figure 4 and the infrared equipment as described in Example 8, the dry columnar solution is continuously circulated through the infrared cell for 85 minutes at a data point frequency of 35 seconds. The initial composition of this solution is 49.3 moles of H ₂ O and 2 moles of HOAc. During the 85-minute period of monitoring, 15 aliquots of HOAc are added to the flask via a syringe, allowing the composition to vary from almost 100% water on a molar basis to nearly 100% HOAc on a molar basis. The trend line in FIG. 17 shows the incremental decrease in H ₂ O concentration and the corresponding increase in HOAc concentration associated with the addition of HOAc aliquots. The purpose of this experiment was to determine the H ₂ O: HOAc molar ratio (which can vary from 0: 100 to 100: 0 from dry column bottom 52 to dry column overhead 56) by on-line infrared light. It can be measured as. The results in Table 18 show the correlation between off-line analyzes by GC and Karl Fischer for passive samples with corresponding on-line infrared values. Both HOAc and H ₂ O can be measured with an accuracy of +/- 0.1 mole over the entire concentration range.

TABLE 18

Example 13:

An experiment similar to Example 12 is performed except that an optical fiber coupled transmission probe as shown in FIG. 4B is used rather than a transmission cell. Two 1-meter diameter, 0.5mm path length probes with sapphire windows are available from two Ximer Analytical, Irvine, Canada, and two 5-meter low lows from CeramOptec Industries, East Long Meadows, Massachusetts. It is coupled to the FTIR through an OH silica optical fiber cable. InAs detectors are used. The purpose of this experiment is to show that permeation probes allow a level of accuracy and precision similar to that of permeation cells. The trend lines in FIG. 18 and the correlation data in Table 19 indicate that sample collection of these two methods yields similar precision and accuracy.

TABLE 19

The above discussion indicates that process control over component concentrations to maximize the production and purification of acetic acid products is achieved by making immediate response adjustments, whether directly or indirectly with respect to component concentrations at appropriate locations in the reaction system. It can be carried out by infrared measurement of the component concentrations of samples collected continuously from locations downstream of the reaction system. The reaction system includes a number of columns and streams, each of which can provide concentration information that informs the system operator where adjustment is needed if appropriate monitoring equipment is in place. This system operator may be a human or computerized control system. In other words, process control can be manual or automated. Since the measurement in the process control of the present invention is continuously taken at intervals of, for example, 30 seconds to 3 minutes prior to nearly instantaneous adjustments to any process variable in the reaction system, the system operator can take data from the infrared analyzer into the process variable. It is useful to have a computerized control unit that analyzes by comparing with known control limits set for and automatically executes the appropriate adjustments to maximize pure acetic acid production. The data can also be supplied to the display unit for interpretation by a person who manually adjusts the reaction system components or process variables. As mentioned above, this adjustment can directly or indirectly change the concentration of one or more components at one or more locations in the reaction system. Direct adjustment can be made by adding or extracting the component at any place in the reaction system. Indirect adjustment of the component concentration can be done in any number of ways. For example, the solution temperature or temperature profile in the column is adjusted to affect component concentration. Reducing or increasing the flow rate of the stream from one vessel to another can affect the concentration of components in these vessels as well as the concentration of other vessels throughout the reaction system. As will be appreciated by those skilled in the art, there are many relationships between different components, including solutions, at various places in the reaction system, and adjusting the concentration of one component at one place in the reaction system is more than one component at one or more places in the reaction system. Can affect concentration. Thus, real-time infrared measurement, analysis and adjustment are used in the present invention to control the acetic acid manufacturing process to maximize the efficiency and output of the reaction system.

While a description of embodiments of the invention has been illustrated and described in detail, these embodiments are not intended to limit in any way the scope of the appended claims. Other advantages and modifications will be readily apparent to those skilled in the art. Therefore, in a broader sense, the invention is not limited to the specific embodiments described and representative apparatuses and methods, and the illustrative embodiments shown and described. Accordingly, the invention may be derived from this detailed description without departing from the applicant's comprehensive scope or principles of the invention.

Claims (73)

A process control method for component concentration in a reaction system for preparing acetic acid from carbonylation of methanol, Collecting a reaction system sample comprising at least one component selected from the group consisting of water, acetic acid, methyl acetate, methyl iodide, acetaldehyde, hydrocarbons, propionic acid, hydrogen iodide and mixtures thereof from downstream of the reactor vessel in the reaction system Steps, Measuring the concentration of at least one of the components with an infrared analyzer, In response to the measured concentration, adjusting the component concentration in the reaction system in at least one of a reactor vessel, a downstream column or a downstream delivery line in the reaction system. The method of claim 1, Wherein said adjusting step is performed directly by adding or removing components from at least one of a reactor vessel, a downstream column, or a downstream delivery line in said reaction system. The method of claim 1, The adjusting step comprises at least one of the temperature of the reaction system solution, the temperature profile in the column, the flow rate of the reaction system solution or the discharge gas velocity in at least one of the reactor vessel, downstream column or downstream delivery line in the reaction system. Process control method made indirectly by adjustment. The method of claim 1, Collecting a sample from the middle phase of the diesel oil recovery decanter column to measure the concentration and density of the hydrocarbons in the sample by an infrared analyzer and in the hydrocarbon removal column and the alkane removal column in response to the measurement of the concentration and density of the hydrocarbon. Adjusting the flow rate of the medium phase to at least one. The method of claim 1, Collecting a sample from the middle phase of the gas oil recovery decanter column, measuring at least one of the concentrations of methyl iodide and methyl acetate in the medium phase by an infrared analyzer and in response to the measured methyl iodide and methyl acetate concentration, Adjusting at least one of the flow rates from the decanter to the medium and light phase reactor vessels. The method of claim 1, Collecting a sample from at least one of the medium phase and the minor phase of the gas oil recovery decanter column to measure the concentration of acetaldehyde in the sample with an infrared analyzer and in response to the measured acetaldehyde concentration, the rate of water supply into the reactor vessel, Adjusting at least one of the exhaust gas rate out of the reactor vessel and the normal flow rate to the acetaldehyde removal system. The method of claim 1, Collecting a sample from the phase of the gas oil recovery decanter column and measuring at least one of the concentration of acetic acid and the concentration of water in the sample with an infrared analyzer and returning to the gas oil column in response to the measured concentration of water or acetic acid Adjusting at least one of the flow rate of the gas, the flow rate of the light phase into the reactor vessel, and the temperature of the solution in the light oil column. The method of claim 1, Collecting a sample from the bottom of the gas oil column to measure at least one of the concentration of hydrogen iodide and water in the sample with an infrared analyzer and into the gas oil column in response to the measured concentration of water or hydrogen iodide, Adjusting the flow rate of the stream comprising water. The method of claim 1, Collecting the sample from the feed stream to the drying column to measure the concentration of water in the sample with an infrared analyzer and the retardant feed rate from the light oil recovery decanter column to the light oil column and the reactor in response to the measured water concentration Adjusting at least one of the rate of supply of water to the vessel. The method of claim 1, Collecting a sample from a heavy oil column to measure the concentration of propionic acid in the sample with an infrared analyzer and in response to the measured concentration of pyrionic acid, the rate of water supply into the reactor vessel and the exhaust gas out of the reactor vessel Adjusting at least one of the speeds. The method of claim 1, Collecting a sample from a heavy oil column to measure the concentration of water in the sample with an infrared analyzer and adjusting a temperature profile in a drying column in response to the measured concentration of water. The method of claim 1, Collecting a sample from at least one of the feed stream, the product output stream, the bottom waste stream, and the overhead output stream of the heavy oil column to measure the concentration of propionic acid in the sample with an infrared analyzer and in response to the measured concentration of propionic acid , Temperature profile in the heavy oil column, flow rate from the heavy oil column overhead drum to the heavy oil column, flow rate from the overhead drum to the drying column, flow rate of the bottom waste stream and product output stream Adjusting at least one of the flow rates. The method of claim 1, The infrared analyzer is a Fourier transform infrared spectrometer. The method of claim 1, Wherein said infrared analyzer operates in one of a mid-infrared region and a near-infrared region. The method of claim 1, Delivering the measured concentration to a control unit for real time analysis. The method of claim 15, Wherein said adjusting step is performed substantially immediately after measurement and analysis. The method of claim 1, Wherein said measuring and adjusting step is performed every about 30 seconds to three minutes. A process control method for component concentration in a reaction system for preparing acetic acid from carbonylation of methanol, Collecting a reaction system sample comprising components of water, acetic acid, methyl acetate, methyl iodide, acetaldehyde and hydrocarbons from the middle phase of the light oil recovery decanter vessel downstream of the reactor vessel in the reaction system; Measuring the density of the sample and the concentration of at least one of the components with an infrared analyzer; In response to the measured concentration or density, (a) the flow rate of the recycle stream to the reaction section, (b) the rate of flow of the medium phase into the acetaldehyde removal system, (c) the rate of flow of the medium phase into the hydrocarbon removal column, ( d) adjusting at least one of the rate of supply of water to the reactor vessel, and (e) the rate of exhaust gas out of the reactor vessel. The method of claim 18, and wherein said recycle stream in (a) flows from the light phase of said decanter vessel to said reactor vessel. The method of claim 18, and wherein said recycle stream in (a) flows from the middle phase of said decanter vessel to said reactor vessel. The method of claim 18, and wherein said recycle stream in (a) flows from dry column overhead into said reactor vessel. The method of claim 18, and wherein said recycle stream in (a) flows from the gas oil column to the flash tank. The method of claim 18, and wherein said recycle stream in (a) flows from a hydrocarbon removal column into a flash tank. The method of claim 18, The infrared analyzer is a Fourier transform infrared spectrometer. The method of claim 18, Wherein said infrared analyzer operates in one of a mid-infrared region and a near-infrared region. The method of claim 18, Delivering the measured concentration to a control unit for real time analysis. The method of claim 26, Wherein said adjusting step is performed substantially immediately after measurement and analysis. The method of claim 18, Wherein said measuring and adjusting step is performed every about 30 seconds to three minutes. The method of claim 18, Adjusting only (a) in response to the measured concentration or density. The method of claim 18, Adjusting only (b) in response to the measured concentration or density. The method of claim 18, Adjusting only (c) in response to the measured concentration or density. The method of claim 18, Adjusting only (d) in response to the measured concentration or density. The method of claim 18, Adjusting only (e) in response to the measured concentration or density. A process control method for component concentration in a reaction system for preparing acetic acid from carbonylation of methanol, Collecting a reaction system sample comprising components of water, acetic acid, methyl acetate, methyl iodide and acetaldehyde from the light phase of the light oil recovery decanter vessel downstream of the reactor vessel in the reaction system; Measuring the density of the sample and the concentration of at least one of the components with an infrared analyzer; In response to the measured concentration or density, (a) the flow rate of the light stream into the reactor vessel, (b) the flow rate of the light stream into the light oil column, (c) the temperature of the solution in the light oil column, (d Adjusting the rate of supply of water to the reactor vessel, (e) the rate of outgas out of the reactor vessel, and (f) the rate of normal flow to the acetaldehyde removal system. The method of claim 34, wherein The infrared analyzer is a Fourier transform infrared spectrometer. The method of claim 34, wherein Wherein said infrared analyzer operates in one of a mid-infrared region and a near-infrared region. The method of claim 34, wherein Delivering the measured concentration to a control unit for real time analysis. The method of claim 37, Wherein said adjusting step is performed substantially immediately after measurement and analysis. The method of claim 34, wherein Wherein said measuring and adjusting step is performed every about 30 seconds to three minutes. The method of claim 34, wherein Adjusting only (a) in response to the measured concentration or density. The method of claim 34, wherein Adjusting only (b) in response to the measured concentration or density. The method of claim 34, wherein Adjusting only (c) in response to the measured concentration or density. The method of claim 34, wherein Adjusting only d in response to the measured concentration or density. The method of claim 34, wherein Adjusting only (e) in response to the measured concentration or density. The method of claim 34, wherein Adjusting only f in response to the measured concentration or density. A process control method for component concentration in a reaction system for preparing acetic acid from carbonylation of methanol, Collecting a sample comprising components of water, acetic acid and hydrogen iodide from the bottom of the gas oil column downstream of the reactor vessel in the reaction system; Measuring the concentration of at least one of the components with an infrared analyzer, Adjusting the temperature of the solution in the light oil column in response to the measured concentration. The method of claim 46, The infrared analyzer is a Fourier transform infrared spectrometer. The method of claim 46, Wherein said infrared analyzer operates in one of a mid-infrared region and a near-infrared region. The method of claim 46, Delivering the measured concentration to a control unit for real time analysis. The method of claim 49, Wherein said adjusting step is performed substantially immediately after measurement and analysis. The method of claim 46, Wherein said measuring and adjusting step is performed every about 30 seconds to three minutes. A process control method for component concentration in a reaction system for preparing acetic acid from carbonylation of methanol, Collecting a reaction system sample comprising water and acetic acid components from a feed stream to a dry column downstream of a reactor vessel in the reaction system; Measuring the concentration of at least one of the components with an infrared analyzer, In response to the measured concentration, adjusting at least one of (a) the flow rate of the light flow from the light oil recovery decanter column to the light oil column and (b) the rate of supply of water to the reactor vessel. . The method of claim 52, wherein The infrared analyzer is a Fourier transform infrared spectrometer. The method of claim 52, wherein Wherein said infrared analyzer operates in one of a mid-infrared region and a near-infrared region. The method of claim 52, wherein Delivering the measured concentration to a control unit for real time analysis. The method of claim 55, Wherein said adjusting step is performed substantially immediately after measurement and analysis. The method of claim 52, wherein Wherein said measuring and adjusting step is performed every about 30 seconds to three minutes. The method of claim 52, wherein Adjusting only (a) in response to the measured concentration or density. The method of claim 52, wherein Adjusting only (b) in response to the measured concentration or density. A process control method for component concentration in a reaction system for preparing acetic acid from carbonylation of methanol, Heavy oil column downstream of the reactor vessel in the reaction system, feed stream to the heavy oil column, product output stream out of the heavy oil column, bottom waste stream out of the heavy oil column and overhead outside the heavy oil column Collecting a reaction system sample comprising components of water, acetic acid and propionic acid from at least one of the output streams, Measuring the concentration of at least one of the components with an infrared analyzer, In response to the measured concentration, (a) the flow rate from the overhead output stream back to the heavy oil column, (b) the flow rate from the overhead output stream to a dry column, and (c) the bottom waste stream. Flow rate of (d) flow rate of the product output stream, (e) temperature profile in a drying column, (f) temperature profile in the heavy oil column, (g) feed rate of water to the reactor vessel, ( h) adjusting at least one of the exhaust gas velocities out of the reactor vessel. The method of claim 60, The infrared analyzer is a Fourier transform infrared spectrometer. The method of claim 60, Wherein said infrared analyzer operates in one of a mid-infrared region and a near-infrared region. The method of claim 60, Delivering the measured concentration to a control unit for real time analysis. The method of claim 63, wherein Wherein said adjusting step is performed substantially immediately after measurement and analysis. The method of claim 60, Wherein said measuring and adjusting step is performed every about 30 seconds to three minutes. The method of claim 60, Adjusting only (a) in response to the measured concentration or density. The method of claim 60, Adjusting only (b) in response to the measured concentration or density. The method of claim 60, Adjusting only (c) in response to the measured concentration or density. The method of claim 60, Adjusting only (d) in response to the measured concentration or density. The method of claim 60, Adjusting only (e) in response to the measured concentration or density. The method of claim 60, Adjusting only (f) in response to the measured concentration or density. The method of claim 60, Adjusting only (g) in response to the measured concentration or density. The method of claim 60, Adjusting only (h) in response to the measured concentration or density.