Method for controlling a cleaning progress of coatings on a working member

p0001The present invention relates to the collection of deposits and in particular to detection of a change of a pad during a cleaning process.

p0002In the processing and production of liquid foods such as milk, beer or juices, it comes to production the formation of deposits. The occurrence of deposits in investments in the food processing industry, in particular in heat exchangers for pasteurization and sterilization processes an inevitable phenomenon. It occurs in both mineral, organic and mixed organic-mineral deposits which often thicknesses up to 0.5 mm or can achieve it. These deposits impair the heat transfer of heat exchanger, increasing the pressure losses and can negatively impact through rising counts and cooked flavor to the product quality.

p0003The cleaning of equipment to remove the deposits of product contact surfaces takes place in the prior art typically by cleaning-in-place techniques (CIP) techniques, which usually both alkaline and acidic cleaning chemicals and disinfectants are used in each case with an intermediate rinse. The application is carried out usually at elevated temperature. The CIP cleaning is carried out according to the prior art in technically specified intervals and sometimes with large safety margins during the cleaning period.

p0004The inventors have a total of 368 companies in the dairy industry and the brewing industry questioned in a market study. carried out according to the results of the survey in 72% of cases the cleaning cycles in a fixed time frame or depending on the circumstances of the production process such as shift changes, batch changes, product changes or production end.

p0005There was an average of ten cleanings per plant and per week and an average duration of the wash cycles of 110 minutes determined. This can be a cumulative average cleaning duration of about 1000 minutes (about 17 hours) per week estimate. In a three-shift operation with 5-day week this makes for 15% of the production time. This estimate is in good agreement with the statement of the company, estimate the average 14% of the plant operation time for cleaning.

p0006Purification is carried out almost exclusively with CIP processes (83%), in which basic, ie usually NaOH (1-4%), and acidic cleaning agents, ie mostly ENT₃ (1-3.5%), are used in successive cleaning cycles. The temperature is in basic cleaning cycle between 60 ° C and 140 ° C, the acid wash cycle between 50 ° C and 80 ° C. A review of the cleaning result does not take place at all companies surveyed (60% checking) and if so, usually by visual inspection system (55%).

p0007The surveyed companies expect from a pad sensor needs-based triggering of the cleaning and consequently reduce the number of cleaning cycles (65%). Furthermore, better information about cleaning completely done is by a sensor located and thus the opportunity to avoid unnecessarily long cleaning cycles (69%). Both requirements keep broadly balanced.

p0008When the number of cleanings the company expected savings of 0% -20% (mean 8.5%), in the cleaning time savings of 10% -50% (mean 22%). The expected reduction of plant downtimes are between 5% and 50% (mean 17%). In addition, savings are expected (14% average) in the consumption of cleaning agents in the amount of 10% -30%.

p0009Referring to the above-performed the market study is consequently a need for a sensor which allows detection of a cleaning stage of completion of a cleaning of surfaces in the equipment or working parts.

p0010This object is achieved by a method for controlling a cleaning operation Claim 1.

p0011The technical publication <nplcit id="ncit0001" npl-type="s"><text>Lonergan M., et. al .: "Array-Based Vapor Sensing Using Chemically Sensitive, Carbon Black-Polymer Resistor", Chem. Mater. 1996 8, page 2298-2312</text></nplcit> it is known to provide a sensor structure for detecting gases with a plurality of differently coated capacitive sensors, each responsive to certain gases by altering their surface coating. This change is read out in all the capacitive sensors at the same frequency to detect the change in capacitance for each of the sensors. Due to this read out at the respective same frequency change in capacitance values ​​a principal component analysis is performed.

p0012From the <patcit id="pcit0001" dnum="EP1111345A2"><text>EP 1111345 A2</text></patcit> is a position-measuring device known which comprises a Inkrementalsignaldetektoranordnung responsive to two incremental, which are provided with different graduation periods.

p0013The <patcit id="pcit0002" dnum="DE19755418A"><text>DE 197 55 418 A</text></patcit> 1 shows a sensor element for measuring complex impedances. The sensor element is arranged with a measuring device to measure, for example, liquids, ion concentrations, a water content in liquids. In order to measure a complex impedance, an alternating current is impressed, with an alternating voltage signal is detected, so that the real and imaginary part of the impedance can be determined.

p0014The <patcit id="pcit0003" dnum="US5789670A"><text>US 5,789,670</text></patcit> discloses a moisture sensor having interdigitated arranged parallel conductive traces. The conductive traces are connected at respective ends to an electrical terminal. When humidifying a defined on the conductive traces of the sensor array, the electrical resistance between the sensor elements varies, so that a turning of a windshield wiper is effected.

p0015The <patcit id="pcit0004" dnum="JP2002014067A"><text>JP 2002-014067</text></patcit> describes an apparatus for detecting a deterioration of a protective film, such as a corrosion or degradation due to moisture or ultraviolet rays. To the film to be detected, a voltage is applied so that an impedance of the film can be determined by detecting an electric current. The voltage is applied in two or more different frequencies in a range from 0.01 Hz to 1 kHz.

p0016The <patcit id="pcit0005" dnum="US6239601B1"><text>US 6,239,601 B1</text></patcit> shows determining a thickness of a layer of ice or a mixture of ice and water. For detecting an impedance of two impedance sensors are provided, which are operated at a first and second frequency. The first sensor includes a first electrode assembly comprising electrodes having a first electrode spacing, while the second sensor comprises an electrode arrangement with electrodes comprising a second electrode spacing. At two predetermined frequencies, an impedance is detected and determined by using measured values ​​that have been determined without the presence of the ice layer to be detected, a ratio of the capacitances of the two sensors which is proportional to the thickness of the ice layer.

p0017The <patcit id="pcit0006" dnum="EP1143240A1"><text>EP 1143240 A1</text></patcit> describes a method for determining characteristics of a sample liquid, which has a plurality of substances. To capture the characteristics of a Cyklovoltagramm is taken, are applied at the current flows, which are measured by a potentiostat as a function of applied voltage. From a maximum current to a information about the concentration of a substance can win by a principal component analysis of the maximum values ​​is carried out.

p0018The <patcit id="pcit0007" dnum="US6300123B1"><text>US 6,300,123 B1</text></patcit> discloses a sensor comprising a conductive electrode on which a conductive polymer layer is formed to have a bio-affinity with respect to certain substances. The polymer layer may bridge the two electrodes between which the impedance is measured. Sensing of the monitored substances, a measurement over a frequency range from 0.01 Hz to 100 kHz is performed using a bioactivity is determined from the detected impedance values.

p0019The <patcit id="pcit0008" dnum="US5859537A"><text>US 5,859,537</text></patcit> describes determining a material deterioration in a coating layer by detecting an impedance thereof. by an electrode test system The electrode system comprises an electrode which is applied to the surface of the coating layer to be detected. The electrode has a lower impedance compared to the impedance of the coating layer. The entire surface of the applied electrode may be used as a reference electrode, and an electrochemical impedance spectroscopy is used to detect corrosion.

p0020The <patcit id="pcit0009" dnum="US5221893A"><text>US 5,221,893</text></patcit> discloses a method for diagnosing deterioration of a paint film, in which an impedance of the film by means of electrodes which are mounted on a surface of the film is detected. Then is calculated using arithmetic units from the measured impedances deterioration of the film.

p0021The <patcit id="pcit0010" dnum="DE10008481A"><text>DE 100 08 481 A</text></patcit> 1 describes performing a measurement to determine a consistency of bone cement by means of detecting an imaginary part of the dielectric constant of the bone cement.

p0022The <patcit id="pcit0011" dnum="US6023170A"><text>US 6,023,170</text></patcit> discloses a method for determining a Aushärtungsgrads a material using a first dielectric constant, which is detected at a first time, and a second dielectric constant, which is detected at a later date.

p0023The <patcit id="pcit0012" dnum="DD248504A5"><text>DD 248 504 A5</text></patcit> shows a method for detecting mastitis in which a milk sample is taken from a respective udder quarter, is supplied with a current and a voltage drop is measured.

p0024In a preferred embodiment, the interdigital capacitor interdigitally arranged comb-shaped electrodes, which can have a spacing of a few microns to a few mm. This allows a signal applied to the interdigital capacitor electric field can penetrate in the deposited on the outer surface of the interdigital capacitor lining, whereby a reliable detection is possible of the same. Preferably, a chip on which the interdigital capacitor is disposed, an area in the range of 50-500 mm² , whereby a favorable compromise a manageable chip area and dimensions, which are necessary for a microscopic deposit formation, is achieved. In one embodiment, the chip has a size of 10 x 20 mm² on.

p0025Preferably, the interdigital capacitor comprises a layer whose outer surface forms the active sensor surface of a material on whose surface comparable behaves with respect to a deposit formation with the surfaces of the working part or plant, in which a lining change is to be detected. In a preferred embodiment, onto the layer of SiC, whereby both one representative plaque formation as well as a high chemical and mechanical protection of the interdigital capacitor is achieved.

p0026In one embodiment is an impedance of the interdigital capacitor on a continuous manner, preferably every 2 to 10 minutes is detected. In the detection can preferably in each case a range of impedances can be measured, for example over a frequency range of 0.5-500 kHz. The time intervals between the respective detection times is set preferably depending on the change of the pad, wherein, in a typical method of measurement approximately every 2 to 10 minutes, a measurement of the impedances or impedance spectra.

p0027Preferably, in a preferred embodiment for calculating the indicator variable, a principal component analysis of impedance spectra is used, for each completed recording a measurement vector is calculated in a main component space, which is also referred to as principal component vector. Using a calculated main component vector and one or several previously calculated main component vectors, the preceding observations are associated, in this embodiment, the indicator quantity can be calculated. As an indicator variable can be used for example in the principal component space, a distance of principal component vectors or a direction of a difference vector.

p0028The processing performed in this method, periodic recalculation of the main component transformation is preferably carried out without resorting to pre-determined measurement vectors, whereby a criterion for engagement in the processor is derived only from the trajectory of the measurement vectors in the principal component space. This allows controlling the process without an a piori knowledge of the process is required. Further, no calibrated main component system is required by the progressive implementation of the principal component analysis, so that problems can be caused by sensor drifts, unexpected cross influences, etc., are largely avoided.

p0029Preferred embodiments of the present invention will be explained in more detail with reference to the accompanying drawings. Show it:<dl id="dl0001"><dt>Fig. 1</dt><dd>a schematic block diagram of a sensor surface according to an embodiment of the present invention;</dd><dt>FIG. 2</dt><dd>a schematic representation of an interdigital capacitor sensor according to an embodiment of the present invention;</dd><dt>Fig. 3</dt><dd>a mapping of interdigital capacitor sensors according to an embodiment of the present invention, wherein an interdigital capacitor sensor is integrated into a system component; </dd><dt>Fig. 4</dt><dd>an illustration of a test track for the simulation of the formation of deposits and cleaning with heat bath;</dd><dt>Fig. 5</dt><dd>an illustration of an interdigital capacitor sensor according to an embodiment of the present invention with a dried coating which forms after one operation;</dd><dt>Fig. 6</dt><dd>a picture of the interdigital capacitor sensor according <figref idrefs="f0005">Fig. 5</figref> after alkaline and acid cleaning;</dd><dt>Fig. 7</dt><dd>a picture of the interdigital capacitor sensor according <figref idrefs="f0005">Fig. 5</figref> in which an alkaline cleaning of a deposit has been performed; and</dd><dt>Fig. 8</dt><dd>a graph wherein the course of a size indicator is shown as a function of time for a cleaning process.</dd></dl>

p0030<figref idrefs="f0001">Fig. 1</figref> shows a schematic block diagram of a device for detecting a surface modification according to an embodiment of the present invention. According to<figref idrefs="f0001">Fig. 1</figref> 10 is interconnected with an impedance sensing device for sensing impedance values ​​of the interdigital capacitor an interdigital capacitor. The impedance detection device is further connected to an evaluation device which is configured to calculate using the detected impedance values ​​an indicator variable that indicates a temporal change of the pad.

p0031An embodiment of the capacitor 10, which is preferably a planar capacitor in order to facilitate a mounting or incorporation into the walls, for example, from pipes or tanks, is in <figref idrefs="f0002">FIG. 2</figref> shown. The capacitor 10 can also be arranged directly on a printed circuit board. In the embodiment according to<figref idrefs="f0002">FIG. 2</figref> the capacitor 10 is formed as an interdigital capacitor, wherein on a dielectric substrate 20 on a surface 22 of the dielectric substrate 20, a first comb-shaped electrode 24 and a second comb-shaped electrode are arranged 26th The first and second comb-shaped electrode are preferably formed in thin film technology with a screen width of some 10 microns. As in<figref idrefs="f0002">FIG. 2</figref> can be seen, the first and second comb-shaped electrode 24 and 26 are interdigitated with one another, so that between the electrodes, a capacitance is formed. The electrodes 24 and 26 are coated with an insulating and resistant to chemicals protective layer 30 comprising in this embodiment, silicon carbide (SiC). In addition, on the type of sensor, a resistance meander (not shown) can be integrated for temperature measurement.

p0032In operation, an outer surface 30a of the protection layer 30 is used as a sensitive surface to which a coating can be deposited. The operation of the capacitor 10 based on the fact that the electrical impedance Z of the same is determined by the dielectric material of the substrate 22, the liquid is contained on the sensor and the layer 30 deposited on the surface. Changes as the thickness of the lining or its composition, this can be detected by measuring the impedance Z. The penetration depth of the electric field in the liquid depends substantially on the electrode spacing and located as a first approximation on the order of the electrode spacing. The electrode spacing can be varied by means of known technologies in the micrometer range, the choice of the electrode distance is made in consideration of the liquid, and a typical composition and thickness of a deposited from the liquid layer.

p0033In a selection of the size of the planar capacitor, that is, more specifically of the chip on which the capacitor is applied, it should be noted that both a handling of the chip area as well as a corresponding dimension of a microscopic deposit formation are required. In one embodiment, the size of the chip 10 x 20 mm comprising², Which surface dimension provides a suitable compromise between a manageable chip area and the dimensions which are necessary for a microscopic deposit formation. In a typical embodiment, the capacitor has a capacitance of about 60 pF in air environment, this value significantly increases in an aqueous environment.

p0034The capacitor 10 can be used as a sensor of a device for detecting a surface change in the capacitor 10 in the working part or plant, in which a lining change is to be detected, are arranged or mounted on a wall. Typically, the sensor chip for mounting is adhered on a circuit board contacted and encapsulated with a liquid-tight plastic so that only the active sensor surface remains free.

p0035<figref idrefs="f0003">Fig. 3</figref> shows a possible attachment, wherein the finished sensor module with the active surface almost flush in a known flange, such as the Varivent ™ - Flange (system component of the company GEA ink Hagen) is installed. The cable required for measuring the impedance in this embodiment back to an external measuring device, which can comprise for example a known impedance, led out.

p0036Referring to experiments conducted by the inventors to study the sensor behavior in deposit formation and cleaning, in the following, an operation of the apparatus is explained for detecting a lining change.

p0037In the tests, a test stand was built, should the conditions as they are to be expected in a tubular heat exchanger with respect to fluid temperature, wall temperature and flow conditions simulate. <figref idrefs="f0004">Fig. 4</figref> shows an illustration of the experimental setup.

p0038The liquid test medium from which a covering should be deposited, a whey powder solution was used. In a subsequent cleaning an alkaline and an acidic industrial cleaners was elected. In order to keep the consumption of liquid test medium in the laboratory low and in the interest of accelerated deposit formation but operating with low flow velocities in the test section. The trainer was designed in such a way to reproduce long controlled manually a complete operating cycle of a food processing plant. The operating cycle here includes pasteurization (with deposit formation), flushing, an alkaline cleaning, a renewed rinsing, an acid cleaning and rinsing again.

p0039Due to the above considerations, a DN25 test tube was chosen as the working part, which was operated by a media throughput of about 200 l / h. The DN25 pipe provides sufficient width sure to accommodate the sensor. Further, the average flow rate of 0.1 m / s, which is established at a flow rate of 200 l / h is sufficient, in order to keep the flow turbulent, what to for the heat transfer from the tube wall into the test medium and for realistic thermal conditions the pipe wall is important. For temperature, the test track runs in a heat bath, which can take much higher temperatures than the experimental medium. Thus, the characteristic overheating of the heat exchanger tube wall can be adjusted.

p0040every 2 to 10 minutes impedance spectra of the sensor have been taken during a test run for the detection of a change of a covering. The measurement of the impedance Z was carried out over a frequency range of 0.5 to 500 kHz respectively. This frequency range has the advantage that it not expensive circuits may be used. However, the Fequenzbereich may also comprise a conventional for electrochemical impedance spectroscopy frequency range which includes, for example, 10 Hz to 1 MHz. As will be explained in more detail below, is determined from the recorded impedance spectra on a computing method of purification success. The total number of spectra are recorded during a test, is typically between 50 and 100. This ensures that a sufficient amount of data is available for the computing process. To carry out a simultaneous visual control of deposit formation and lining clean the sensor has been developed in some tests at regular intervals.

p0041Although a clear and unambiguous relationship between the type of flooring, its thickness and the experimental conditions could not be established, it was observed that the formation of deposits on the sensor surface does not visually noticeable differences to form deposits on exposed metal surfaces, for example on the face of the mounting flange, having , This shows that the active sensor surface comparable to the stainless steel surfaces of the piping system behaves with respect to a deposit formation and cleaning, so that the same can be used as a representative test area for critical plant components. This realization is what makes it possible to use the capacitor 10 as a sensor for detecting a change pad to close by the sensor and its measurable surface finishes and cleaning state on the system status.

p0042<figref idrefs="f0005">Fig. 5</figref> shows a state of the sensor, after the same has been exposed at 85 ° C over a period of 120 minutes a 8% whey powder solution. As can be seen, is formed on the outer surface of the sensor, which is the active surface, a coating which is not significantly different from a coating on the surfaces of the flange.

p0043The cleaning tests, a clean, shiny surface sensor has been reached in the test facility usually at no more impurities were observed visually.

p0044<figref idrefs="f0006">Fig. 6</figref> shows a representation of the arrangement of <figref idrefs="f0005">Fig. 5</figref>After an alkaline and acid cleaning process is performed. As can be seen, both the metal surfaces of the mounting flange and the outer surface of the sensor according to the alkaline and acidic cleaning visually completely blank.

p0045As for a cleaning course it is noteworthy that the alkaline cleaning step for removing organic coating components is very time consuming and critical. In preliminary experiments it was observed that a treatment time of at least 10 minutes to 20 minutes was necessary in order to achieve a good overall cleaning of the surfaces. This was confirmed in the experiments, in which the time had to be partially extended to alkaline Reinigungssgang to about 30 minutes.

p0046After the alkaline cleaning are generally not observed abnormal mineral deposits, as <figref idrefs="f0007">Fig. 7</figref> Instructions. <figref idrefs="f0007">Fig. 7</figref> shows the sensor arrangement of the experimental setup according to an alkaline cleaning for 60 minutes in 2% NaOH at 60 ° C. As can be seen, both the metallic flange and the active surface of the sensor element to a mineral residual lining.

p0047The experiments described above that the capacitor 10, in particular due to the nature of the protective layer 30 comprising in this embodiment, SiC, is adapted to detect a surface change in a working portion or a plant, since the formation of deposits on the outer surface, ie, the active surface of the capacitor 10 is representative of the deposit formation on the metal walls of a working part.

p0048In the following, a preferred embodiment of detecting a surface modification will be described. In this embodiment, an impedance spectrum is recorded by the impedance detecting device 12 in successive time intervals, which can extend for example over a frequency range of 0.5 to 500 kHz. In each case, evaluated a change between the last two measurements in order to obtain an indicator variable that indicates a change of the pad. This difference method allows an end point of cleaning by means of the sensor is easily detectable, since the end of the cleaning is characterized by a steady state in which the chemical cleaning operations have come to a standstill at the sensor surface and the contact surfaces. In an initial cleaning the supplied by the sensor readings continuously change in large steps, while the measured values ​​only vary randomly at a final cleaning around an average. The difference method provides thereby a simple method for detection of the endpoint, in which a determination is made as to whether a predetermined number of preferably sequential indicator quantities lies within a predetermined range.

p0049The calculation of the size indicator may be performed by a principal component analysis, which is calculated in a one-step or multi-step process from the sensor data. In this method each obtained upon detection of impedance values ​​of a main component transformation are subjected, to obtain from the detected impedance spectra a measurement vector in a main component space. Further explanations of principal component analysis are for example in the<patcit id="pcit0013" dnum="EP1143240A1"><text>EP 1143240 A1</text></patcit> to find. The main components calculation is repeated at predetermined time intervals, to calculate the impedance spectra are used, which have been detected by the last calculation time point to the current time of calculation. Further, also a mean value or a selection of the impedance spectra, for example, by only every second spectrum is used, performed. In carrying out the principal component analysis and a predetermined number of measurement vectors, for example, the last n measurement vectors, all available measurement vectors or averaged measurement vectors can be used.

p0050By repeatedly calculating creates a trajectory of the measurement vectors in the principal component space, wherein the end of the purification process is characterized in that the trajectory terminates at a saturation point of the main component compartment or only varies randomly by the same. However, this saturation point may be exposed to certain influences which hinder the application of a decision criterion for recognizing a cleaning end of the trajectory in the principal component space. For example, between two successive cleanings lying with the same operating parameters can change the sensitivity of the sensor due to sensor drift before the start of cleaning, so that arise for two cleanings each different saturation points.

p0051To obtain a process that is stable and largely independent of such influences, therefore, an indicator variable of the measurement vectors are calculated in the principal component space, which allows to detect and make a decision with respect to an engagement end of the cleaning process. These filters can be used, for example, a distance criterion or a direction criterion. Preferably the current measurement vector in principal component space and the previous measurement vector used in the principal component space. The indicator variable is typically also recalculated with each newly conducted principal components calculation, so that a time profile of indicator quantities yields.

p0052In order to decide whether an end of the cleaning process is present and an engagement in the cleaning process is to be performed, the indicator variable is used, for example, by a determination is made whether the indicator quantities permanently, ie above a certain number, in a predetermined area. Thereupon, a cleaning end signal is generated, indicative of an end of the cleaning process. Further, for example, using other sensed parameters, such as a period of time until it reaches the end or a sensed temperature at the end point, a decision can be made whether to intervene in the cleaning process or another process is to be performed.

p0053<figref idrefs="f0008">Fig. 8</figref> shows a sequence of indicator quantities as a function of time was determined at a caustic cleaning a pipeline. The measured values ​​obtained were thereby subjected periodically a main component transformation. The indicator variable was in each case calculated from a difference between two successive measurement vectors or sensor records the principal component space, the Mahalanobis distance was used as a distance criterion.

p0054As in <figref idrefs="f0008">Fig. 8</figref> can be seen, shows that the cleaning progress, as in many reactions, an exponential law follows. In order to decide whether an intervention in the cleaning process is to be performed, a determination is made whether an indicator variable, that is durable over a certain number, in a predetermined area. For example, you can specify a limit to a first approximation, the in<figref idrefs="f0008">Fig. 8</figref> is designated by the reference numeral 40, to the indicator signal must fall below permanently. Characterized the end point of the cleaning is shown, in accordance with the graph<figref idrefs="f0008">Fig. 8</figref> is reached after about 25 minutes. This date coincides with the visually determined endpoint of the cleaning liquor.

p0055A significant advantage of the above-explained automatic recognition of the cleaning end is to use the difference method, whereby a calibration is not necessary. The system is self-calibrating and may interference that may occur, for example, by different detergent or variations of liquid food products, offset to a certain extent.

p0056Thanks to modern numerical methods necessary for the evaluation of calculations can be performed on smaller microcontrollers, since the executable program code under MS-DOS just having a size of 54 kB. When cleaning processes in the range of 20-60 minutes cycle times of one minute for a measured value and calculation are quite sufficient. For the measured value even simpler circuits are available, as neither the accuracy nor the scope of laboratory measuring instruments for the process measurement is necessary.