JPH07501615A - Improved magnetic resonance analysis in real-time industrial use mode - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] References to related applications for improved magnetic resonance analysis in real-time industrial use mode Both applications are entitled “Magnetic resonance analysis in real-time industrial use mode”. U.S. Patent No. 5.015954 issued May 14, 1991 to Butien et al. No. 5.049.8 issued September 17, 1991 to Butien et al. It is closely related to No. 19. Both of these patents are of common assignment with this application. The disclosures of both of these patents are incorporated by reference as detailed herein. It is cited as.

field of invention The present invention provides continuous sampling of process material streams by pulsed nuclear magnetic resonance (NMR) techniques. Regarding equipment for measuring the type and amount of lattice bonds and free magnetically active liquid in the Details include moisture content, polymer content, crystallinity rate, fat rate, and other component analysis percentages and and the application of such measurements to industrial process control of and other parameters.

Background of the invention NMR technology is most notably used in medical applications where in-vivo examinations of various parts of the human body can be observed. It has been widely used in the field of instruments and in clinical laboratories for the past 40 years. In addition, the application of these technologies to industrial measurement and control tasks is increasing. There was a particular use and interest in The present invention utilizes existing optical and radiant energy Effective use of pulsed NMR techniques in the industrial field to replace or supplement metrology. enable use (technically and economically).

Pulsed NMR spectroscopy is described in our above-cited patents. This technology , the nucleus of the particular nuclide in the sample being measured (first, it is essentially precessed in a static magnetic field) A burst or pulse designed to excite the moving protons of such a sample. Use Luz. In other words, the precession is corrected by the pulses. Pulse suitability After this, a free induction decay (FID) of the magnetization associated with the excited nuclei occurs.

Traditional Fourier transform analysis can be used advantageously to investigate the core of the problem. A frequency domain spectrum is generated. Pulse duration, time between pulses, pulse The phase angle and sample composition are parameters that influence the sensitivity of this technique.

These frequency domain techniques can be easily used in industrial applications, especially online applications. It's not possible.

One object of the invention is to lead to (improved) accurate and fast measurements of the type and I of the nuclides in question. The measurement system is

A further object of the invention is to provide an industrial online method for measuring and calibrating the control process itself. This is its application to the in-problem.

Another object of the invention is to utilize time domain analysis in achieving such a system. It's there.

The raw variables in question are the moisture content, oil content, and polymer binding of polypropylene and polyethylene. crystallinity, density, stereoregularity, and molecular structure. However, hydrogen or hydrogen Other parameters can also be measured based on other sensitive species including fluoride, fluorine, etc. . The purpose of the invention is to allow a variety of such measurement tasks.

Other objectives include density, temperature, fill factor and size factors, triboelectricity and static electricity, vibration, and industrial measurements, including frequent measurements, repeated measurements, periodic measurements, and non-periodic measurements. It lies in allowing the dynamics of online applications.

A further object of the invention is to Accurate and fast measurement of the amount of Its application to on-line monitoring and control process problems, organic and inorganic substances measurements of free and bound water (based on hydrogen nucleus modified precession analysis) and other parameters. Meter measurements (based on hydrogen or other sensitive species including fluorine, sodium 23, etc.), Density, temperature, filling factor and size factors, triboelectricity and static electricity, vibration, and frequent measurements industrial online applications, including variations in repeated measurements, periodic measurements, and non-periodic measurements. It consists in integrating all the permissible features of mechanics for use.

A further object of the present invention is to analyze the NMR-active substances, more particularly moisture content in solids, crystals/ Food products and polypropylene where amorphous ratio, oil/fat or solvent ratio is measured online The stereoregularity and coexisting molecular structure of polypropylene mixtures are measured online. When applied to seeds (or mixed phases), these achievements can be further extended to industrial online processing. The aim is to expand the

Summary of the invention The present invention preferably provides magnetic resonance hardware, control (and associated data acquisition and data reduction means and steps). Ru. Using this system, continuous generation lines or similar repeat measurement systems can be It can capture data.

NMR systems are practical and economically viable in industrial online contexts. Reliably extract free decay data in a practical manner. This system is The basic magnetic field homogeneity should be arranged to a reasonable degree and cancel out the inhomogeneity effects, and the thermal drift effects. Each measurement of temperature stabilization and digital data reduction to a reasonable extent and offset Regularly use multiple runs (10-100) and statistics for industrially valid measurements. Characterized by the use of physical methods or other data manipulation. against polypropylene For discussion/analysis, these data are first transformed into a very fast Gaussian curve, Then the continuation (slower modification is the Gaussian curve with an independent variable K(t)'') (multiplied by the cosine of can be expressed as a free induction decay curve (FID). polypropylene For materials other than A modified Gaussian curve (a sine function often called an Abragam) is added to the Gaussian curve. multiplied by ), followed by a slower real Gaussian curve, and then an even slower exponential It can be expressed as an FID having a time component consisting of a functional domain. These differences The resulting component is the precession of protons in the sample being measured in a high static magnetic field. A pulse of energy near the transmitted and resonantly coupled radio frequency that induces a modification of motion. represents proton relaxation after initial excitation by gas. M-L technique (described below) ) is still applicable.

For polypropylene, the fast Gaussian curve FID section is measurement data of the magnetization decay of relatively passive structures present and picked up by the NMR receiver. Based on the data points (this component relates to the isotactic polypropylene structure) I can do it. The slower modified but Ouss curve component has a constrained structure and more mobility. It is a transition region between two different structures. The slow exponential FID part is usually based on moving structures. Ku. This component can be associated with an active polypropylene structure.

The faster is the Ouss curve, the slower is the modified Ouss curve and the exponential FID part and The FID as a whole typically has a decay origin set near the time center of the excitation pulse. It's hard to extrapolate to. The zero time intercepts of these curves give the FID intercept and/or or the intercept of one or more curve sections is used to provide ratio data, and the polymer -Stereoregularity or crystallinity is determined. Although density can also be determined through the present invention, This is because FID varies predictably as a function of stereoregularity/crystallinity and density. It's a good thing.

If the FID is analyzed as an Abragam, a slow but Ussian curve, and an exponential function, The ragam FID moiety is present in the sample and can be detected by the NMR system. Magnetization of immobile structures of highly bound protons (as in crystal structures) picked up by Based on the measured data points of the attenuation (; this part is usually the hydrogen or hydrocarbon needle of the compound ( or, in the case of other NMR-active species, chemically bound and highly constrained. It is based on the response characteristics of bound proton species such as fluorine and sodium23). Slower but better The line region is a somewhat more mobile structure between bound and loosely bound protons. This is a transition area. The slow exponential FID part is usually physically present in the sample. loose (bundled) NMR-active species such as moisture but not substantially chemically bound. Based on constrained or unconstrained structures (Abragam, Gaussian curves and exponential functions) The several FID sections as well as the FID as a whole are usually located near the time center of the excitation pulse. It is possible to extrapolate to the set attenuation origin. These curves (i.e. FID and the FID intercept and /or one or more intercepts of the curved section are used to provide ratio data, e.g. For example, wet substances (e.g. for upstream or downstream corrections or for pass/fail purposes) Processing of industrial chemicals, minerals and metals, agricultural products, and processed foods by determining moisture content for free water versus bound water in the case of temperature control) is determined. For example, moisture in foodstuffs Instead of measuring the relatively crystalline and amorphous components of the material, e.g. For example, it may be possible to measure the ratio of hard and soft components of a polyethylene material. This is achieved in a similar manner to moisture measurements. Polymer crystallinity and density are also inventive FID can be measured as a function of crystallinity and density. This is because it changes as expected.

Additionally, oils and solvents may be present in food products, and residual solvents may be present in polypropylene or other industrial in the material as an error factor to be resolved and/or as an objective parameter to be measured. can appear as data. The NMR response characteristics of such solvent or fat moieties can be determined by the present invention. To some extent, such solvent or oil subcomponents are not mentioned herein. An additional very slowly decaying exponential region in the calibration system being can appear.

The measurement system of the present invention is economical compared to widely used laboratory systems. It includes parts that have been reduced in scale and industrially strengthened. virtually fixed The magnetic field has a 4.000-8.000 Gauss field (nominal, approximately 4,700 Gauss) It contains closely spaced pole pieces. Allows for quick adjustment to precise and accurate magnetic fields. Helmholtz, which is highly adjustable and prefers slow adjustment to the thermal environment. file is installed. This is due to a constant (gamma) associated with the material at the resonant frequency. Guarantees that the product multiplied by a certain magnetic field strength matches the excitation frequency with a certain deviation This is to do so. Still further fine-tuning can be done in signal processing as described below. will be done.

The present invention provides an interoperable temperature control system to provide a measurement system that meets the above objectives. Streamline large flows of data in a practical way through the features mentioned below related to Allow. Building materials are also integrated into reliability considerations, as discussed below. . sample measurements (in contrast to the long-time measurements of many prior art systems) Often achieved in approximately 1 minute.

Based on the intercept delay time constant and/or the ratio of the integrated area under the curve and/or peak analysis. The measurements carried out according to the invention generally involve measuring the weight or volume of the sample in the measurement area. Precise weight measurement is independent of system gain, whereas precise weight measurement It is a necessary characteristic (and limiting factor) of a system of technology.

Advances in industrial on-line measurements of the same parameters treated in the present invention Technology efforts include non-NMR gravimetric systems, chemical systems, and radioactive systems. , acoustic systems, optical systems, and capacitive/capacitive systems; However, the purpose of the present invention is not fully satisfied, and the NM supporting continuous industrial processes The use of R forces the use of off-line laboratory equipment (although it is costly). However, the data sampling rate is insufficient) or industrially enhanced Some previous efforts in the 1980s in pulsed NMR instruments were one or two data points were used for FID analysis. The present invention solves all of these It breaks the deadlock.

A laboratory method for frequency-domain NMR analysis was introduced by Spies in “Deuteron NMR "Molecular mechanics of solid polymers made clear" (261 Colloids & Polymers) -Science, 193-209 (1983)) and “Protons” by Kaufman et al. Determination of transient temperature and crystal content of linear polymeric polyethylene by NMR spectroscopy Quantity” (27, Journal on Polymer Science, 2203-2209 (1989)) for crystal content determination. Detecting free radicals in coal time domain using pulsed and multipulse NMR free induction decay in coal to generate Regional analysis was conducted in ``The Utility of Pulsed Nuclear Magnetic Resonance in Studying Protons in Coal'' (81 , Journal on Physical Chemistry, 566-571 (1977 )) and “IH nuclear magnetic resonance study of domain structures in polymers” (52(9), JAPAN) On Applied Physics, 5517-5528 (1981)) Reported laboratory by Gerstein et al. (Ames Institute, University of Iowa) shown in the system. These devices or IBM Federal Systems Division models and Equipment in front of Bibratka GmbH model P2O1 and peer on February 7, 1984 US Pat. No. 4,430,719 issued to Son et al. This was an earlier attempt called commercial use. Pearson's work was recognized by Kaiser in 1985. Embodied in industrial plant control work of aluminum & chemical coporation . This did not work as a reliable quantitative device. auburn y International began selling Pearson/Kaiser products in 1987-1988. However, it did not meet the requirements of industrial online monitoring. these The failure was supplemented by the production of the patented invention by Butien et al. cited above. , all of which, when improved through the present invention, can be applied to industrial applications for various materials. Solve line problems.

Other features, features and advantages of the preferred embodiments may be seen in conjunction with the accompanying drawings. It will become clear from the detailed explanation below.

Brief description of the drawing 1 and 2 are cross-sectional views of a preferred embodiment of the invention, including electrical block diagram components. be.

FIG. 3A is a diagram illustrating the construction of the embodiment of FIGS. 1-2 in the process of analyzing polypropylene. 2 shows a one-hour voltage waveform of induced decay (FID).

Figure 3B shows Figure 1- in the process of analyzing other materials, especially polypropylene. 2 shows a voltage one-hour waveform of free induction decay (FID) in Example 2.

FIG. 4 shows the diagram of FIG. 2 is a flowchart of the measurement stage using -2.

Figure 5 shows nine samples of polypropylene with various isotactic indices ( Figure 1 is a voltage one-hour trace of an FID curve drawn from a hexane or hexane soluble material.

and, Figure 6 shows the difference in polypropylene measured by the invention (y-axis) versus the standard method (x-axis). This is a calibration curve for xylene solubles.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS 1-2 is a block diagram inserted cross-sectional view of a first embodiment of the apparatus and method of the present invention. Ru. Industrial process line IPL has logistics as indicated by arrow A. Ru. Some of this material is captured by probe P and passed through the incoming loline LI. and is supplied to the sample area S1. The above areas are themselves substantially interfering Substantially non-magnetic, non-conductive materials (glass, certain ceramics) that do not produce an FID signal usually about 30 cm long, consisting of It is defined by a tube 98 having a length. The sample area is near the entrance. It is defined between the valve v1 and the outlet valve V2. Gas jet J is also installed. ing. These are repeatedly pulsed on and off for periods of sample entry and ejection. Stir the fluid sample material into the solution. Region S2 is an important part of the sample.

It is surrounded by a sample coil 100 tuned to resonance, and a tuned circuit 1 02 and associated transmitter/receiver controller 104. grounded Loop 101 contains one or more excitation pulses that help form the magnetic field of coil 100. Lenses disposed above and below the coil 100 to surround the magnetic field formed by the coil 100. It is a legal shield. The controller 104 includes a built-in microprocessor and required power components. child, memory, program, to the hardware shown and the keyboard 10 I10 complex suitable for interconnection to an external microcomputer 106 by monitor (or other display device) 110, recording device 112 and/or process control device 114 (for controlling processes in IPL). Operate The controller is activated and controlled from the display keyboard 108 and the resulting data data and signals are shown on display 100 and at 110, 112 and/or 114. and used. Computer 106 also controls equipment operating conditions.

Region S2 of tube 98 and coil 100 are connected to yoke 118, pole piece 120, and surrounding Magnetic assembly 1 including a Lumholtz coil 124 and a coil current generator 117 placed in an electrostatic but adjustable intersecting magnetic field defined by 16 There is. The critical sample area S2 of tube 98 and the magnet are connected to high thermal conductivity face plate 128 and metal (but non-ferromagnetic) with the total mass relative to each other. ) contained in box 126, which causes the output from coil 100 to be emitted to the sample. harmonics and other interference with the signal returning from the sample and/or the coil 10. 0 and its tuning circuit 102 and pickup by the transmit/receive controller 104 be minimal for.

The magnetic assembly including the yoke 118 and other components therein shown on FIGS. 1-2. Assembly 116 includes optional inert gas fill and purge controls (not shown), internal gas The temperature of the motor M that drives the heater 134, the fan 136, and the yoke is the other reflecting the temperature in the sample area at and between the pole pieces 120. The environmental control chamber 132 includes a temperature sensor 138 applicable to the sensing area. It is. Thermal controller 140 processes the temperature signals from 138 to 134/136. coarsely adjust the heating/circulation at the pole piece 120 and the helm hole at the pole piece 120. The current passing through the coil 124 is adjusted as a sensitive and fast fine control, and at the same time Executes general control instructions for computer 106. to pump 144 and yoke 118 The attached coil 146 as well as the attached coil 146 to the plate 128 of the box 126. Additional thermal stability is provided by a closed loop heat exchanger 142 having a coil 148 can be given. Still further thermal stability can be achieved by supporting temperature controlled air (control equipment not shown). the annular area defined by the collar 134A around the sample. can be obtained.

The temperature controllers 134-138 and 142-146 described above control the sample temperature. Very coarse and less coarse thermal regulators are established to combat fluctuations. . In addition, a thermal guide collar 134A directs temperature-controlled air to the sample area. It can exist. As also described in another preferred embodiment, the heater 134 and/or the fan 136 may be deenergized when performing sensitive measurements to reduce interference. You can.

(Thus, the strength of the magnetic field between the poles 120 in the region S2 of the sample - penetration and impurity) The denaturation is controlled by a uniform base magnetic field throughout region S2. helmhol The coil 124 is energized by the coil current controller 117 and the sample is placed inside. Accurately adjust the final strength of the applied magnetic field. This magnetic field is connected to the magnetic pole 120 and the helm hole. This is the vector sum of the magnetic fields due to the Ruth coil 124. Control device 117 is a current generator is used to set the current through the Helmholtz coil 124. The coil 124 The magnetic field formed by the current flowing through the pole pieces 124 It is wrapped around the pole piece so that it is added to or pulled by the field. coil 124 The magnitude of the current through the pole piece (and associated yoke structure) is Or determine the strength of the magnetic field drawn from the magnetic field.

The actual quantification of the current through the Helmholtz coil is based on the magnetic energy described hereafter. and resonance techniques are performed in a preliminary run and the helm is held until maximum sensitive resonance is achieved. This is accomplished by adjusting the Holtz current. Next, the Helmholtz current is the system is adjusted to deviate from the resonance point by about 1-3 KHz.

The main elements of the electric control device are coils 100 and 101 and variable capacitor 10. 2-1 and 102-2, a resistor 102-3, and a diode 102-4. is configured to tune to a Q of 20 to 40 to achieve a resonance of 100. tuner 102, transmit/receive switch 104-1, and transmitter 104-2. and receiver 104-3, crystal oscillator 104-4, gated pulse generator (GP G) A control device 104 including a phase shifter 104-5 and a phase shifter 104-6.

The MOD and DEMOD elements of the crystal oscillator transmitter 104-2 and receiver 104-3 provides a nominal 20 MHz carrier wave that is phase modulated or demodulated. receiver includes variable gain amplifier elements 104-31 and 104-32 for operation. reception The resulting analog signal is processed by a high-speed, small (12-bit flash) 105-1 and, for example, the valves v1 and v2 via the valve control device 115. Keyboard 108, monitor 110, modem 109, recording element 112 for control and/or to an external computer 106 having a process control device element 114; Coil current controller 12, all via a digital-to-analog converter (not shown) Internal (for equipment) CPU element 105-2 that supplies data to .

The analog signal FID curve is prefiltered and anti-aliased. ing) is adjusted by a Bessel filter acting as a filter. Digital After conversion, this signal is smoothed in time by a fast Fourier transform filter program. be converted into The combination of these filters improves the accuracy of the system. Provides a considerable improvement in noise ratio.

Excitation of the coil 100 and excitation precession of protons in the sample and immediate relaxation/depletion After demodulation, controlled gain amplification, A/D conversion and point plotting, the attenuation is shown in Figures 3A and 3A. and a received signal having a free induction decay (FID) curve shape C shown in 3B. generate.

The one-hour voltage traces in Figures 3A and 3B represent a “cycle” of sample excitation and free induction decay. It shows the elements of The excitation pulse center is considered to. Transmitter 104 electronic The engineering component is not effectively received until saturation effects are overcome at tl. Next, The available curves are developed. This signal processing device is useful as mentioned below Continuous waveforms can be added or subtracted for precise adjustment.

The digitized FID curve data of FIG. 3A is stored in external computer 106. The program now finds the best curve that fits each stored FID curve. I'll start. This FID curve has three main components, shown as ASB and C in Figure 3A. It has component parts. The A curve, which occupies the first part of the FID curve, is a fast Gaussian curve. In contrast, the 8 curves occupying the middle part of the curves are slower modified Gaussian curves. E, which occupies the later part of the FID curve, is an exponential decay. high speed The Gaussian curve and the exponential part are relatively immobile parts of the polypropylene sample, respectively. Related to minutes and moving parts. The slower modified but Uss curve is between these two This is the transition region of Although it is important to determine the type of curves that make up this FID curve, , which means that once the curve is known, it will be close to the center of the transmitted burst signal. The time origin (tO, i.e., the excitation of cycle 1) is denoted as A, , B, and Eo. However, you can go back and enlarge it. This is due to the saturation of the electronics components of the equipment. The summation effect (occurring during and after the burst signal to tl) precludes direct measurement of the intercept. Important for accuracy. The part of the curve in question is from tQ, beyond which the curve is The line extends to t4, which is too small to be a problem, and this electronic circuit is (centered on tO #2 (starting from pulse) requires recovery time to prepare for Ru.

The digitized FID curve data of FIG. 3B is stored in external computer 106. The program now finds the best curve that fits each stored FID curve. I'll start. FIDID curve Three main component parts shown as A, B and E in Figure 3B have a certain amount. The curve that occupies the first part of the FID curve is the Abraham curve. , whereas the 8 curves occupying the middle part of the curves are slower Gaussian curves (i.e., a (slower than Bragham's Gaussian curve component), and E, which occupies the later part of the F curve, is a finger It is a mathematical decay. These abragam and exponential parts are sampled, respectively. The bound and unbound protons of the sample (for example, (1) water or hydrated molecules of the sample (2) highly polymerized and lightly polymerized materials; crystalline and non-crystalline components, including mixtures of substances in which they are both present; and (3) controlled by the mixed elastomeric polymer components). Although it is slower, the Uss curve is It is a transition region between the two and is partially coupled or uncoupled. controlled by protons in the state of Types of curves in this FIDID curve diagram is important, but once the curve is known, the transmitted A at the time origin (10, i.e., excitation of cycle 1), which is close to the center of the signal. Bo and E. (shown as ) and can be enlarged. This is t There are saturation effects in the equipment's electronics gears that occur during and after the burst signal to l. It is important for the existence of During this time, measurements cannot be made accurately and , the area in question under the curve is the degree of the number of nuclei in the sample, from tQ, it is Beyond that, the curve extends to t4, which is too small to be a problem, and this electronic circuit is recovery time to prepare for the cycle (starting with a pulse centered at tO #2) Requires.

Since this system is off-resonance, the received signal is offset from the non-resonant FID. is the product of the cosine terms at the start frequency. Many factors can cause a system to move away from resonance. It is difficult to operate at the resonance point because it easily comes off, but on the other hand, Operation near resonance is relatively easy to achieve and maintain. Near-resonance actuation of the system gives good results as before, but this is due to small side effects due to physical variations. This is because it only causes use.

3A and 3B, each (sub)cycle culminates in t5, recovery, i.e. sample proton fruit before starting a new transmit signal burst (to#2). Perform qualitatively complete mitigation. Typically, the excitation pulse interval is 5 to 10 microns. seconds, and the tO-t1 time is 5 to 15 microseconds (the shorter the better). ), tl-t2, dominated by effects due to relatively immobile structures, is 5 to 15 microseconds. t2-t3 is a period of 15 to 25 microseconds characterized by phase cancellation effects. and t3-t4 is the transition region between The predominant range is 50 to 500 microseconds. From several hundred microseconds to several seconds. the final t4-t5 region for recovery (re-equilibration) of the sample material, which may exists.

The overall FID curve, as mentioned, sometimes has fewer components, but polypropylene and similar It consists of several main components as shown in formula 1 when applied to a substance like . Additionally, in addition to or in place of one or more of these ingredients, high faster Gaussian curve or Abragam, slower modified Gaussian curve or sub-exponential For example, Ce-1kl+ (where X is between 0 and 1 or between 1 and 2). ) and these additional components may be present early, mid or late in the decay cycle. It can have a strong influence on the entire FID curve in the period part.

Formula I F (t) = [cos (Δf) t) + [High speed is a Gamko + [slower modified but Uss curve or slower but Uss curve Co + [exponential function] ]) This entire curve is automatically An iterative process based on the Mal-Luthor-Levenberg (M-L) approximation technique applied adjusted by the This technique converts the FID curve into a specific given error range. Sizes of all parameters, constants, frequencies, etc. included in Equation 1 to be optimally adjusted Determine. This is an iterative technique where the entire curve is determined at once. M-L technique Techniques are provided in the following references: by Dee, W., Marquardt. “Ind, App 1. Math,” Volume 11, pp. 431-441 , “Data Attrition and Errors for Physics” by Philip Earl Bebbington. "Difference Analysis" (N.Y. York, Manf-Grow-Hill), Chapter 11, and J.E.D. “Current status of numerical analysis” by Niss (London, Academic Press, David) Jacobs, 1977 edition), IIl, Chapter 2. Preferences of the present invention When applied to the measurement situation herein in a specific example, the selected parameters are Y-axis intercept ratio and time constant, temperature coefficient, frequency term and other parameters mentioned below It is.

M that limits these coefficients to a limited range of sample types described herein. It has been established that there are approximate limits placed on the -L technique. this technology does not converge after a given number of iterations, 30 in the preferred embodiment, then its The sample is discarded. In addition, this technique is difficult to avoid when the system is too close to the resonance point. Sometimes I fail.

Once the formula of the FID curve is known, each component establishes the intercept of each above component. Therefore, we can extrapolate back to to.

The resulting data utilized by computer 106 (FIGS. 1-2) is expressed as Equation 1 The formula for an FID curve consisting of three components (including a cosine term) shown in be. Each of these curves (and their intercepts) are experimentally related to the same kernel of the problem It is attached. In particular, once the FID curve formula is determined, the y-cross section of the indicated curve The following ratio of pieces is formed by a biexponential function/fast curve (or Abragam), ( R1), and modified Gaussian curve/Fast Gaussian curve (or Abragam), (R 2). Ratio (R1) and (R2), cross product (R12) and (R22), three curves Decay time, product temperature and cosine frequency of the modified Gaussian curve for each of the components The terms form a 10-dimensional model. Through regression analysis, these 10 terms were found to be polypropylene. The isotactic index of ren (xylene soluble material) or the density of polyethylene, Or some are tightly coupled, others loosely coupled, and some in between. is associated with the same core of matter for other substances. These results are a sample , and the gain of the system, thereby making it possible to measure these physical quantities There is no need to do so.

This M-L iterative process is performed using a chi-square function (measurement method), a technique well known in the art. curve tone by minimizing the sum of the squared differences of the data points from the constant data point and the derived equation Perform the adjustment.

Calibration of the system involves measuring a large number of known samples and using the M-L technique to calibrate each known sample. This is achieved by deriving the model equation constants associated with the sample. Next, regression An analysis is performed to derive coefficients relating various model equation constants to the desired measurand. Ru. A calibration formula is then created using these coefficients.

In operation, FID is obtained from the test sample and by M-L technique, the model equation A constant is found. These constants are then entered into the calibration equation and the desired parameter is Calculated.

These data can be printed as QC type measurements or back to the line IPL (Figure 1). On-line control that is activated to control a process or related equipment Can be used as a parameter.

The input operating parameters of the system may be in the form of FROM or ROM or disk or Parameters previously stored on a magnetic storage medium such as tape or telephone line and modem 1 This can range widely to include inputs sent via 09. Excitation/Kuru Generation of energy can be achieved using multiple techniques including coils or antenna devices. come. The stable magnetic field may be an electromagnet, a permanent magnet, an electromagnet with superconducting windings or other It can be generated using standard magnetic field generation technology.

Figure 4 is an expanded flowchart showing the measurement steps for establishing valid industrial measurements. be. First, in an abbreviated cycle attempt to establish curve C, the sample area is One free induction decay curve C is established to see if it is empty (quick FID ). If the sample area is not empty (N), (re)opening of valve V2 and The measurement is stopped so that the action of the de-J and gravity clears the area. new quick F Exclusion is established by the [) stage. Next, close valve ■2 and turn knob v1. and any other protocols that may be necessary (if any) to ensure sample acquisition. The sample enters by forming a configuration of line P and line LI. Jet J is Condition and stabilize this new sample.

The electronics signal processor baseline should be cycled for 3 to 4 cycles (each with a baseline output of (with addition of (C+) and (C-) to detect offset and compensate for it) +) and (-) subcycles). This baseline offset Avoiding the cut determination and simply taking this as an additional (i.e. Although it would be feasible to process with

Further adjustment is established by coil 124 to adjust HO, which ranges from 10 to This is made possible by FID curve generation of 20 magnetic field test cycles. (C-) F ID is subtracted from (C+)FID, i.e. the absolute C value is added to the resonance point. A processable digitized FID signal is obtained which has a maximum value at . Is HO like this? coil current generator 117 and coil 124 until a maximum value is achieved. HO is then varied to shift the system by a known amount from the resonance point. child These measurements are taken within a reliable range for such purposes, i.e. t3 -t4 index range [the above baseline measurements are also taken here] ]. Sufficient magnetic field conditioning typically occurs in less than 7 cycles.

Multiple measurement cycles are then typically performed to obtain usable data. Each service The cycle involves modulation transmit/receive/flash A-to-D conversion and data storage.

These curves are then averaged for M-L curve adjustment and the cuts listed above are applied. pieces and ratios are established. For rapid FID magnetic field testing and baseline calibration purposes It is possible to apply a similar cycle, but one that is omitted to some extent. come. Each of the subcycles [(+) and (=)] of each such cycle is a data reduction. It involves the acquisition and utilization of thousands of FID points in a small number of locations. Each subcycle exists on the order of 1 second and the number of such subcycles employed depends on the desired smoothing and signal-to-noise ratio (S/N). In general, the S/N ratio is a square root relation to the number of accumulated cycles. It will be improved in the future.

Higher accuracy and Reliability is required and measurements are distorted by sample tube composition Sometimes. glass is not used (in industrial use it is better to avoid glass) ), the substitution should not contain hydrogen. However, the signal from fluorine is the resonance point. Fluorine may be useful in some applications because it appears remotely. these beliefs The signal can be eliminated from the hydrogen at the desired level of sensitivity and, if desired, filtered (or can be deleted). Other cases of higher sensitivity measurements, e.g. When measuring the relative proportions of each crystal species in a mixture, the sample container is Should be glass or non-proton ceramic. However, in certain cases, fluoride Permissible use of hydrocarbons or reinforced fluorinated hydrocarbons for polymer measurements I can do it. In all such cases, the key is to combine the transmitted energy and mimic the sample Avoid using sample containers that contain species that will cause a FID decay curve to Ru.

Since regression analysis yields 10 variables residing in a 111-dimensional quantity, the result is a group It can't be rough. However, the measurements obtained from the model represented in this invention are The results can be compared advantageously with those obtained from accepted off-line measurement techniques.

Figure 5 was taken for polypropylene samples of variable isotactic index. It shows FID. These curves are well differentiated and for each curve (M- After regression, the three resulting component parts found by From the formula, the actual polypropylene isotactic index (xylene soluble matter; xyle yields the value used when ne5olubles) is calculated.

Figure 6 shows the effectiveness of the present invention when measured by the present invention (y-axis) versus the standard method. A calibration curve for xylene solubles of such polypropylene is shown.

Other embodiments, modifications, details, and uses may be made consistent with the letter and spirit of the above disclosure. and by the following claims construed in accordance with patent law including the doctrine of equivalents. It will be apparent to those skilled in the art that it is within the scope of this patent to be limited to:

FIG, 3A FIG, 3B FIG. 4 @&I=</1f: 10%81All Continuation of front page (81) Designated countries EP (AT, BE, CH, DE.

DK, ES, FR, GB, GR, IE, IT, LU, MC, NL, SE), CA , J.P. (72) Inventor: Smith, Thomas Bee Massachusetts, USA 0 1923-5008. Danvers, Eight Electrox Avenue (No address) (72) Inventor: Marino, Scott A. Massachusetts, USA 0 1923-5008. Danvers, Eight Electrox Avenue (No address) (72) Inventor Touch, Ronald Jay Massachusetts, USA 01923-5008. Danvers, Eight Electrox Avenue (No address) (72) Inventor: Roy, Ajoy Massachusetts, USA 01923-5008. Danvers, Eight ・ElectroX Avenue (no address)

Claims

[Claims] 1. A magnetic resonance apparatus for industrial process monitoring, comprising: (a) accessing and removing continuous samples of polypropylene from an industrial process, placing them in a sample measurement area, and discarding the continuous samples from said area; (b) means for applying a base magnetic field to said region to precess a sample nucleus therein and applying a local excitation pulse to said region to modify said precession; (c) transmitting/receiving antenna coil means and precession of the nucleus to be measured and relaxation detected as free induction damping in the coil means; (d) establishing a digitized form of said free decay and applying it to said decay curve corresponding to a true characteristic of said sample; A magnetic resonance apparatus comprising means for analyzing into components, said analysis comprising determining a time expression of said attenuation curve and establishing a zero intercept and time characteristic from said curve components, and said means (a)- (c) is constructed and configured to measure the number of relatively immobile structures, less immobile structures, and more mobile structures to perform such measurements and analyses. . 2. It is noted that the process monitoring is performed on polypropylene in which the immobile structure is determined from a fast Gaussian curve, the somewhat more mobile structure from a slower modified Gaussian curve, and the more mobile structure from an exponential function. The device according to claim 1, characterized in that: 3. Claim 1 characterized in that said process monitoring is carried out on polyethylene in which said immobile structure is determined from abra gum, the less mobile structure from a slower Gaussian curve and the more mobile structure from an exponential function. Equipment described in Section. 4. The above process monitoring and analysis of immobile structures determines the number of binding nuclei from abragam, while the above process monitoring and analysis of somewhat more mobile structures determines the number of intermediate binding nuclei from a slower Gaussian curve, and the more mobile structure determines the number of intermediate binding nuclei from a slower Gaussian curve. Apparatus according to claim 1, characterized in that said process monitoring and analysis of the structure determines the number of unbound nuclei from an exponential function. 5. The means for determining the curve components of the decay curve convert the digitized form of the free induction decay into a time expression of the decay curve and establish the zero intercept and time characteristics of the curve. 2. The apparatus according to claim 1, further comprising means for utilizing Marcart-Tollebenberg iteration for the purpose of the present invention. 6. Means for analyzing a free induction decay component corresponding to the amount of target nuclei in the sample is a means for establishing a calibration formula, wherein the zero intercept and time constant characteristics of the decay curve component of the standard sample are means associated with a known physical quantity of a target nucleus in said standard sample; and means for calculating said physical quantity of interest using said calibration equation, zero intercept, and time characteristics of an unknown sample. Apparatus according to claim 1, characterized in that: 7. 7. The apparatus of claim 6, wherein said means for comparing said calibration curve with an unknown sample decay curve comprises regression analysis. 8. 2. The apparatus of claim 1, further comprising feedback means for the sample measurement region providing a constant thermal environment in which thermal drift components are eliminated. 9. A method for monitoring an industrial process using magnetic resonance, comprising accessing and removing a continuous sample from the industrial process, placing the sample in a sample measurement area, and applying a base magnetic field to the area. applying a localized excitation pulse to said region to modify said precession; receiving and converting a free induction decay signal from relaxation of said sample nucleus; digitizing said free induction decay and analyzing it into components of said decay curve corresponding to the amount of nuclei of interest in said sample, said analysis determining a time expression of said decay curve; establishing zero intercept and time characteristics from said curve components, relating said zero intercept and time constant characteristics to quantities of relatively immobile structures, less immobile structures, and more mobile structures and discarding said samples from said range; A method comprising steps. 10. The above industrial process is performed on polypropylene and the immobile structure is determined from a fast Gaussian curve, the somewhat more mobile structure from a slower modified Gaussian curve, and the more mobile structure from an exponential function. 10. The method according to claim 9, characterized in that: 11. The above process monitoring and analysis of immobile structures determines the number of binding nuclei from abragam, while the above process monitoring and analysis of somewhat more mobile structures determines the number of intermediate binding nuclei from a slower Gaussian curve, and the more mobile structure determines the number of intermediate binding nuclei from a slower Gaussian curve. 10. The method of claim 9, wherein said process monitoring and analysis of the structure determines the number of unbound nuclei from an exponential function. 12. Determination of the time equation of said digitized decay curve is a step of applying Marcart-To-Levenberg iteration to said digitized decay curve, said digitized curve being divided into three time equation components. 10. A method according to claim 9, characterized in that it includes the step of being divided. 13. The step of relating the zero intercept and time characteristic to the amount of target nuclei in the sample includes the step of generating a calibration equation that relates the zero intercept and time characteristic of the decay curve component of the standard sample to the known amount of target nuclei in the standard sample. , and calculating the quantity using the zero intercept and time characteristic measured from the unknown sample in conjunction with the calibration equation, the quantity contained in the unknown sample being determined. 10. The method of claim 9. 14. 15. The method of claim 14, wherein said calculation using said calibration equation and said decay curve of unknown samples includes regression analysis. 15. 10. The method of claim 9, wherein the shape of the decay curve comprises at least two of Abragam, fast Gaussian, slow Gaussian, slow modified Gaussian, exponential, and modified exponential. .